SEC Info℠

Home

Search

My Interests

Help

Sign In

Please Sign In

Oracle Systems, et al. – ‘SC TO-T/A’ on 7/13/04 re: Peoplesoft Inc – EX-99.(A)(5)(CXXXV)

On:  Tuesday, 7/13/04, at 6:27am ET   ·   Accession #:  1193125-4-117066   ·   File #:  5-42583

Previous ‘SC TO-T’:  ‘SC TO-T/A’ on 6/10/04   ·   Next:  ‘SC TO-T/A’ on 7/15/04   ·   Latest:  ‘SC TO-T/A’ on 1/13/05

in

this entire Filing. an “Entity” Search.

Show

Documents searched

and

each “hit”. the 1st “hit”.

↓↑Hints

 Text Wildcards:  ? (any letter/digit),  * (many).  Search Logic:  for Documents:  & (and),  | (or);  for Text:  | (anywhere),  "(&)" (near).  Details... 

  As Of                Filer                Filing    For·On·As Docs:Size              Issuer               Agent

 7/13/04  Oracle Systems                    SC TO-T/A              4:1.1M Peoplesoft Inc                    RR Donnelley/FA
          Pepper Acquisition Corp.

Document/Exhibit                   Description                      Pages   Size 

 1: SC TO-T/A   Amendment No. 61 to SC TO-T                         HTML    165K 
 2: EX-99.(A)(5)(CXXXIII  Text of Press Release by Parent Dated     HTML      9K 
                          July 12, 2004                                          
 3: EX-99.(A)(5)(CXXXIV)  Conclusion of Law Filed by Parent on      HTML    244K 
                          July 8, 2004                                           
 4: EX-99.(A)(5)(CXXXV)  Finding of Fact (Redacted) Filed by        HTML    531K 
                          Parent on July 8, 2004                                 

Exhibit (a)(5)(cxxxv)

LATHAM & WATKINS LLP

J. Thomas Rosch (Bar No. 37668)

Gregory P. Lindstrom (Bar No. 82334)

Daniel M. Wall (Bar No. 102580)

505 Montgomery Street, Suite 1900

San Francisco, California 94111-2562

Telephone: (415) 391-0600

Facsimile: (415) 395-8095

ORACLE CORPORATION

Dorian Daley (Bar No. 129049)

Jeffrey S. Ross (Bar No. 138172)

500 Oracle Parkway, 7th Floor

Redwood Shores, California 94065

Telephone: (650) 506-5200

Facsimile: (650)506-7114

Attorneys for Defendant

Oracle Corporation

UNITED STATES DISTRICT COURT

NORTHERN DISTRICT OF CALIFORNIA

SAN FRANCISCO DIVISION

TABLE OF CONTENTS

Pursuant to the Court’s June 25, 2004 Minute Order in this matter, Defendant Oracle Corporation hereby submits its Proposed Findings of Fact. Based on the record of the case, including the testimony of witnesses called to trial, exhibits, designated deposition testimony, written discovery, and all other such materials that are before the Court, Oracle submits that the Plaintiffs did not, to a preponderance of the evidence:

(a) define a cognizable relevant product market and specifically did not define a sufficiently concrete and definite product market in “high function” financial management or human resources software;

(b) prove that products other than newly licensed Oracle, PeopleSoft, and SAP offerings properly are excluded from Plaintiffs’ alleged relevant market;

(c) define a cognizable relevant geographic market and specifically did not demonstrate a United States only market is proper;

(d) prove that there exists deleterious economic price discrimination against the interests of “high function” or “large complex enterprise” customers that would support price discrimination markets;

(e) prove that the proposed transaction would lead to such a concentration of market shares in cognizable relevant markets as to warrant a presumption of adverse competitive effect,

(f) prove that concentrated market shares, even if proven, would warrant such a presumption under the circumstances of this case and in light of the nature of the procurement process and buyer power in the subject industry;

(g) prove that, as a result of the transaction, Oracle unilaterally could impose a profitable, non-transitory price increase;

(h) present any evidence that, as a result of the transaction, Oracle could impose a profitable, non-transitory price increase as a result of coordinating with other competitors; or

(i) provide any factual support for Professor McAfee’s auction-based model of expected competitive effect from the proposed transaction.

1. Defendant Oracle Corporation (“Oracle”) is a Delaware corporation with its principal place of business in Redwood City, California. (Joint Stipulations of Fact (6/1/04) at 1.) Oracle provides organizations with database technology, enterprise software applications, consulting services, and related products in the United States and abroad. (Joint Stipulations of Fact (6/1/04) at 1.)

2. PeopleSoft, Inc. (“PeopleSoft”) is a Delaware corporation with its principal executive office in Pleasanton, California. (Joint Stipulations of Fact (6/1/04) at 2.) PeopleSoft provides organizations with enterprise software applications and consulting services in the United States and abroad. (Joint Stipulations of Fact (6/1/04) at 2.)

3. Human Resources Management software (“HRM”) and Financial Management Systems software (“FMS”) are types of enterprise applications software (“EAS”). (Joint Stipulations of Fact (6/1/04) at 2.) Many organizations (including corporations, federal, state, and local government agencies, and non-profit organizations) use HRM and FMS software in some form to automate their financial management and human resources functions. (Joint Stipulations of Fact (6/1/04) at 2.) Employee records and benefits administration are some of the more common business functions addressed by HRM programs sold by software vendors. (Joint Stipulations of Fact (6/1/04) at 2.) General ledger, accounts payable, and accounts receivable are some of the more common business functions addressed by FMS programs sold by software vendors. (Joint Stipulations of Fact (6/1/04) at 2.)

4. Oracle, PeopleSoft, and SAP sell HRM and FMS software in the United States and throughout the world. (6/1/04 Joint Stipulated Facts) Lawson Software, AMS, Microsoft, and other software vendors also sell HRM and FMS products in the United States and throughout the world, (see infra.)

5. “Sales” of EAS software typically involve payment by customers of a license basis by which customers pay a license fee and an annual maintenance fee. EAS customers are

sensitive to initial license fees, in part because the license fee directly impacts the amount of the maintenance fee, which typically is 15-25% of the license fee. (Tr. 133:12-15 (S. Hatfield); Tr. 655:2-4 (P. Maxwell); Tr. 1385:6-11 (M. Gorriz).) The total cost of ownership for EAS includes the license fee, maintenance, support, implementation, hardware, network, change management, and training costs. (Tr. 1658:13-1660:14 (C. Bass).) Implementation costs typically are substantial, running up to five times or more the amount of the initial license fee. (Tr. at 2475:10-18 (P. McAfee); 1215:8-10, 1284:20-1285:5 (C. Peters); 1565:14-20 (C. Wolfe); 4364:3-21 (D. Teece); 1660:15-1665:21 (C. Bass).) Customers frequently buy HRM or FMS modules as part f a larger cluster of software and technology products.

6. Oracle initiated a tender offer for the shares of PeopleSoft on June 6, 2003. (Joint Stipulations of Fact (6/1/04) at 2.)

7. Complaint Filed. Plaintiffs sued Oracle Corporation on February 26, 2004, to enjoin Oracle’s tender offer for PeopleSoft.

8. Alleged violation of Clayton Act Section 7. The government has alleged that Oracle’s acquisition of PeopleSoft will violate Section 7 of the Clayton Act and that it would lessen competition substantially in interstate trade and commerce. Complaint 39.

9. Allegedly, large enterprise customers are most vulnerable. The Complaint alleges that “a large corporation may require a multi-million dollar software solution that is configured to the organization’s needs and can perform these important [financial management and human resource] functions seamlessly and simultaneously across multiple divisions or subsidiaries, multiple lines of business, and multiple legal jurisdictions,” that the “software used to accomplish these tasks varies greatly depending on the needs of the customer,” and that “customers with the most demanding requirements typically find that the set of vendors that can meet their requirements is limited to Oracle, PeopleSoft and one other firm, Germany’s SAP AG.” Complaint 9. “High-function HRM and FMS software,” according to the Complaint, is required by “customers with high-level functional needs (“enterprise customers”). Complaint 14.

10. Alleged product market for “high-function” HRM and FMS software. Based on this premise, the Complaint alleges that the relevant product markets in which competition will be lessened are (a) Human Resource Management (“HRM”) software and accompanying services that can be integrated into some of associated functions for a single vendor with performance characteristics that meet the demands of multi-faceted organizations with high-level functional needs (“high-function HRM software”) and (b) Financial Management (“FMS”) software and accompanying services that can be integrated with suites of associated functions from a single vendor with performance characteristics that meet the demands of multifaceted organizations with higher level functional needs (“high-function FMS software”). Complaint 23. The government acknowledges it coined the term “high-function software” as “shorthand... to describe the phenomenon that everybody in the industry seems to recognize, which is, different customers have different requirements.” (Tr. 30:25-31:3 (C. Scott Opening Statement))

11. Alleged geographic market limited to the U.S. The Complaint alleges that the relevant geographic market is limited to the United States. Complaint 26. The government acknowledges it selected a United States market in part because a worldwide market showed that SAP is the clear market leader and the relative shares of Oracle and PeopleSoft were correspondingly diminished in the broader market view. (Tr. 30:2-10 (C. Scott Opening Statement)) The government further alleges that customers in the U.S. demand different functionality than do customers in Europe (Tr. 9:14-10:8 (C. Scott Opening Statement)) and that the complexity of the software requires vendors maintain a “presence” in the U.S. to sell to U.S. customers. (Tr. 10:6-14, 11:20-22 (C. Scott Opening Statement).)

12. Alleged unilateral competitive effects. The government argues the proposed merger will reduce competition under a “unilateral effects” theory. (See, Plaintiffs’ Trial Brief, filed June 1, 2004, pages 18-25.) The Complaint alleges that the “elimination of one of only three vendors of high-function software will likely result in higher prices.” In addition, Oracle and PeopleSoft allegedly are the two best alternatives for a significant number of customers that

do not view SAP to be a viable substitute. Complaint 31. The government states PeopleSoft is Oracle’s “closest rival” and “many” customers face an Oracle/PeopleSoft “duopoly” today because they are already running Oracle and PeopleSoft in their companies. (Tr. 3330:20-3331:7 (C. Scott Argument).) The government also alleges that currently Oracle and PeopleSoft compete to take sales away from each other (Tr. 3330:16-20 (C. Scott Argument)) and because Oracle and PeopleSoft compete head to head today, the loss of that competition post-merger will result in a price increase. (Tr. 3337:4-13 (C. Scott Argument)) Additionally, the Complaint alleges that “Oracle will no longer have the incentive to innovate in order to differentiate itself from PeopleSoft” and that “enterprise customers within the current installed bases of Oracle and PeopleSoft will likely suffer harm if the merger is permitted.” Complaint 32.

13. Alleged relevant market for “each transaction.” In order to clarify the allegations regarding market shape and size in the government’s pleaded claim, Oracle propounded an interrogatory that asked “Do you contend that each transaction by which high-function HRM and FMS software is procured constitutes a relevant antitrust market?” (Interrogatory No. 26.) The government responded that “the vendors in those markets sell diversified products through a long procurement process that permits price discrimination, and no two customers or sales are identical. Therefore, economically speaking, it would be valid to treat every sale of high-end HRM or FMS software as a distinct antitrust market. However, for ease of analysis, Plaintiff has aggregated similarly situated customers into product markets where price discrimination occurs as provided by the Horizontal Merger Guidelines and as reflected in the Complaint.” (Plaintiff’s Response To Interrogatory No. 26.) The government’s principal economist, Dr. Kenneth Elzinga, confirmed that this was “one way to think about” the government’s claim. (Tr. at 2229:1-14 (K.Elzinga).)

14. Enterprise applications software (“EAS”). EAS is “the [broad] universe of software applications that are used by an enterprise to automate its overall business processes.” (Parties’ Joint Submission of Definitions dated July 1, 2004, at 6.) EAS refers collectively to

software modules from the following categories (“pillars”) of business process applications: Human Resources Management (HRM), Financial Management Systems (FMS), Supply Chain Management (SCM), Customer Relationship Management (CRM), Product Life Cycle Management, Student Information Systems (SIS), and Business Intelligence (BI) pillars. (Parties’ Joint Submission of Definitions dated July 1, 2004, at 6.)

15. Enterprise Resource Planning Software (“ERP”). ERP has no generally accepted meaning and “refers, at a minimum, to the set of business applications software that is designed to automate an organization’s back-office functions, such as payroll, general ledger, accounts payable and receivable, benefits management, and employee recordkeeping. It traditionally encompasses HRM and FMS. Some also include one or more other pillars in the definition of ERP, such as Supply Chain Management (SCM) or Customer Relationship Management (CRM).” (Parties’ Joint Submission of Definitions dated July 1, 2004, at 6.) For instance, Christy Bass of Accenture explained that Accenture uses the term ERP today to refer to EAS modules in the HRM, FMS, SCM, CRM, and business intelligence pillars. (Tr. 1675:16-21, 1676:1-3 (C. Bass).) Richard Knowles of SAP, who testified that SAP “invented” the term ERP, indicated that SAP uses ERP to refer exclusively to EAS modules in the HRM and FMS pillars. (Tr. 2810:1-3, 2810:8-17 (R. Knowles).) Curtis Wolfe, the CIO of the State of North Dakota, described North Dakota’s ERP system to include EAS modules in the FMS, HRM, and Student Information Systems (SIS) pillars. (Tr. 1534:10-18 (C. Wolfe).)

16. Core HRM. Core HRM means those specific EAS modules that individually or collectively automate Payroll, Employee Tracking and Benefits Administration business processes. (Parties’ Joint Submission of Definitions dated July 1, 2004, at 6.)

17. Core FMS. Core FMS means those specific EAS modules that individually or collectively automate General Ledger, Accounts Receivable, Accounts Payable, and Cash and Asset Management business processes. (Parties’ Joint Submission of Definitions dated July 1, 2004, at 5-6.)

18. Core HRM and FMS are a small subset of EAS. Core HRM represents only three of tens of modules in the HRM pillar. (Plaintiffs’ Exhibit 3010 (PeopleSoft’s Human

Capital Management Solutions) (displayed); Tr. 268:8-269:11 (R. Bergquist). Similarly, core FMS represents only four of tens of modules in the FMS pillar. (Plaintiffs’ Exhibit 3011 (PeopleSoft’s Financial Management Solutions) (displayed); Tr. 270:5-271:12 (R. Bergquist).) HRM and FMS, in turn, represents only two of the at least six pillars that comprise EAS. (Parties’ Joint Submission of Definitions dated July 1, 2004, at 6-7.) As a result, core HRM and FMS comprise only a small fraction of the revenues of the different software vendors. For the first nine months of 2003, for example, Oracle had only nine deals in which a Fortune 500 or Global 2000 customer bought core FMS, and seven where they purchased core HRM. (Tr. 3455:12:3456:10 (S. Catz).)

19. “High-function software” is not an industry-recognized term. High- function software” has no recognized meaning in the industry, (see, e.g., Tr. 349:7-10 (R. Bergquist), 2298:6-20 (K. Elzinga).) Plaintiffs’ witnesses, including its expert Dr. Elzinga, agree “high-function software” is not a term with any standard industry meaning. (Tr. at 349:7-10 (R. Bergquist), Tr. at 2298:6-20 (K. Elzinga).) Professor Elzinga acknowledged that the phrase “high function” was coined for this case, and is not used by anyone in the industry. (Tr. at 2151:24-2152:3, 2181:9-16 (K. Elzinga).)

20. Plaintiffs’ proposed definition of high-function software refers to product and vendor features, customer uses, and preferences that occur in various, unpredictable combinations. The government’s definition of “high-function,” which drives its alleged product market definition in this case, is as follows:

A “category of software products” (also called “enterprise,” “upmarket,” or “Tier One” software) “capable of executing a wide array of business processes at a superior level of performance,” and which has “at least the following advanced performance capabilities, not all of which each user may demand:” it a) is

scalable, b) is highly configurable, c) performs a variety of related transactions seamlessly and d) handles international aspects of a business.

(Parties’ Joint Submission of Definitions dated July 1, 2004, at 2-4.) Even witnesses who accepted the government’s concept of high-function software – including Professor Elzinga – could not identify any objective measure that can predictably identify “high function software” in contrast to other types of software. (Tr. 2181:9-16, 2311:3-17 (K. Elzinga).) For instance, Richard Bergquist of PeopleSoft stated that there was no clear cut answer as to the number of concurrent users that would meet the scalability requirement of “high-function software.” (Tr. 353:15-18 (R. Bergquist).)

21. No vendor fulfills all “high-function” needs. Plaintiffs’ expert Professor Iansiti acknowledged that no vendor was capable of meeting all of the high-function needs of customers all of the time. (Tr. 2085:3-8 (M. Iansiti).) Moreover, Emerson Electric CEO Charles Peters testified that none of the ERP vendors could meet the truly high function needs of Emerson because “the software vendors that you’re talking about in the financial area have kind of basic functionality. We — we have demands for higher-level functionality than they offer.” (Tr. 1220:2-6 (C. Peters).) Emerson has 45 ERP platform “brands” running within the organization today, and about a hundred different instances. (Tr. 1199:13-17 (C. Peters); Ex. D5168 (Charles A. Peters PowerPoint presentation, “Positioning for the Future”) at 002155).) This includes solutions from ERP vendors, best-of-breed, legacy and home-grown (or custom) applications. (Ex. D5170 (Emerson FMS Footprint).) Christy Bass of Accenture similarly testified that no single vendor offers a comprehensive solution to meet the EAS needs of customers all the time. (Tr. 1657:4-10 (C. Bass); Ex. P3198R (C. Bass 4/21/04 Dep.) at 112:17- 114:5, 138:14-139:2, 176:11-22.)

22. Government’s definitions are based on customer preferences or usage. According to Professor Iansiti, “high function” FMS and HRM software matches the needs that flow from “certain characteristics” of large, complex enterprises. (Tr. 2088:21-2089:4 (M. Iansiti).) When asked whether software must meet each and every one of those needs to be

considered “high function” software, Professor Iansiti responded that “[t]he more needs that [the software] matches, the more successful it is.” (Tr. 2089:5-8).) Professor Iansiti could not identify a “clear line” distinguishing “high function” software needs from lower-function or mid-market software needs. (Tr. 2090:14-21 (M. Iansiti).) He stated that whether an organization is using EAS software to meet “high function” requirements “depends on how [the organization] [is] using the software....” (Tr. 2107:19-22 (M. Iansiti).) He also testified that whether the software has been implemented and “whether the software is actually working” at an organization is important to understanding if the software matches a “certain type of need” because “[p]eople can make mistakes.” (Tr. 2102:17-18, 2103:10-19, 2104:3-4 (M. Iansiti); see also Tr. 2105:4-10 (M. Iansiti) (“The question, again, is how much the management values those needs, and they would have other ways of running the business. As I explained before, just because – given piece of software can have value to you in certain ways does not mean that you necessarily going to buy it [sic]. You have to make an economic decision, and – trade off values and costs.”); 2108:7-9 (“[t]hey can select to go through more traditional management processes, for example, that don’t require software”).) While he acknowledged that a number of large enterprises were running Lawson software applications, Professor Iansiti did not interview any of these customers to determine if any are using Lawson to meet “high function” needs. (Tr. 2096:19-2097:8) (M. Iansiti).)

23. “Large and complex enterprise” is a term created by the government. “Large and complex enterprise” has no independent or industry-standard meaning and the evidence offers no clear guidance regarding the size or level of complexity that would qualify a customer as a “large and complex enterprise.” (Tr. 2087:3-4 (M. Iansiti) (there is no “precise line” in defining enterprises); 2087:16-19 (no precise number of employees required for “large”); 2090:14-21 (describing whether an enterprise is “large and complex” as “a complicated process to assess”).)

24. Segmentation of customers by size is for convenience, and does not offer a meaningful way of defining the market. Software vendors may use the terms “large,” “up-

market,” “enterprise,” and/or “mid-market” to focus sales and marketing initiatives. (Tr. 2816:6-2818:8 (R. Knowles) (testifying that SAP divided mid-market and large enterprise at $1.5 billion based on SAP’s sales resources and estimated amount of IT “spend” available from those customers).) While many in the software industry differentiate between large customers and mid-market customers, there is no “bright line” test for what is a “large” or “up-market” customer. (Tr. 348:23-349:3 (R. Bergquist) (acknowledging “different parties tend to define it differently”); 2033:1-12 (M. Iansiti); Ex. P3032 (J. Henley 5/4/04 Dep.) at 98:20-25.) Likewise, there is no “bright line” test for what is a “mid-market” customer. (Tr. 2820:9-19 (R. Knowles) (SAP executive noting that the separation between mid-market and large enterprise customers is “not an exact science”); Ex. D7174 (Pollie 5/26/04 Dep.) at 54:14-55:3 (testifying that the meaning of the term mid-market “varies from, from everyone you talk to.”); Ex. P3191 (K. Block 12/16/03 Dep.) at 88:12-21; 94:19-95:3 (noting the term mid-market is used in many different ways by many different people).) EAS vendors, analysts, systems integrators, and others in the industry define the mid-market variously. (See, e.g., Tr. 864:19-865:2 (P. Keating) (noting variability of definitions and that Bearing Point generally refers to mid-market as customers in its General Business Group, which is synonymous with companies having less than $2 billion in revenue), 1846:17-1847:15 (P. Wilmington) (PeopleSoft used to define mid-market as less than $500 million revenue, but after the JDEdwards acquisition raised mid-market to include companies with less than $1 billion revenue).

25. Different vendors use different metrics for identifying a large enterprise. There are many different ways to evaluate a customer’s relative size: number of employees, revenues and IT budget are common metrics. Microsoft, for example, distinguishes its corporate and enterprise account segments from mid-market and small business segments on the basis of number of employees, number of PCs, and amount of IT budget. (Tr. 2978:9-16 (D. Burgum).) BearingPoint identifies its Tier 1 customers based on revenues of greater than $100 million and inclusion on the Fortune 2000 and Global 500 lists. (Tr. 865:13-21 (Keating).) Prior to Oracle’s tender offer, PeopleSoft used a proxy of $500 million in revenue to distinguish mid-market customers from large customers. (Tr. 348:5-18 (R. Bergquist).) SAP defines its “large

enterprise” market as companies with more than $1.5 billion in revenues. (Tr. 2819:12-20 (R. Knowles).) Oracle segments the market based on the customers’ revenue level or number of employees. (Ex. P3070 (F. Prestipino 5/18/04 Dep.) at 21:5-23:11.)

26. Customers of any size may have complex business process automation requirements. Complexity is in the eye of the beholder. Customers need not be large to be “complex.” (Tr. 2908:20-25 (R. Knowles);

In an interview before the tender offer, Jeffrey Read, PeopleSoft’s Vice President for Mid-Market market positioning acknowledged, “Few have found that just because the companies we’re dealing with are smaller, their demands are any less complicated.” (Ex. D6074 (PeopleSoft article) at 1.) In a separate interview, Mr. Read acknowledged, “The big challenge for small to mid-sized businesses is that they’re just as complex as large organizations, but they’re trying to do it all faster, with fewer people, and with less money.” (Ex. D6075 (PeopleSoft article) at 2.) Christy Bass confirmed that all of Accenture’s clients have complex functional needs. (Tr. 1613:12-23 (C. Bass).)

27. Core FMS and HRM address customers’ common basic needs. Core FMS and HRM modules are offered by all the EAS vendors that have HRM and FMS offerings. (P3179 (Ciandrini 1/16/04 Dep.) at 256:2-257:10 (“And I would say in most of these evaluations, in particular for financial accounting applications, for HR and payroll, the functionality is generally all there in all of the competitors.”).) With respect to this basic functionality, there is little differentiation among the products of the EAS vendors. (Tr. 1505:14-18 (M. Glover) (basic HR functionality). Indeed, PeopleSoft recognized the “current dilemma” of seeking to differentiate FMS solutions that offer essentially the same functionality. (D5889 (PeopleSoft PowerPoint slides re Selling Financial Solutions) at 3 (“ERP products, and some niche players

tend to have almost the same core functionality. We can all post to the GL, manage assets, perform tax accounting, manage travel and expenses, etc.”).)

28. Core FMS and HRM may not be central to customer needs and preferences. Core HRM and FMS modules typically address the same basic needs in all customers, such as (for FMS) general ledger, accounts payable, accounts receivable, and asset management and (for HRM) employee management and benefits. What EAS modules are considered “core” or important to a specific customer, however, will depend on that customer’s needs and preferences. For instance, the State of North Dakota considers functionality for the administering of grants to be important functionality for an FMS system. (Tr. 1578:5-6 (C. Wolfe); see also Tr. 1230:12-13 (C. Peters) (more critical HR pieces for Emerson are opinion surveys and time and labor reporting).)

29. Core FMS and HRM do not address strategic customer needs. Core FMS and HRM automate basic business processes, and are not of high strategic value. (Tr. 1228:9-13 (C. Peters) (HRM is a “simplistic” application, not mission critical); 3208:17-22 (K. Harris) (FMS is not strategic-value EAS).); Tr. 4238:14-4239:3 (D. Kutnick).) As customers evaluate business process solutions, FMS and HRM are necessary but not “strategic” since they do not drive value to the bottom line in the way that CRM or SCM products are more likely to do. (Tr. 4239:1-3 (D. Kutnick).) Government witness Cox Communications, for example, projected $0 in “tangible, assignable benefits” from implementing a new suite of core FMS products, while at the same time Cox projected “tangible, assignable benefits targeted at $70 million over 4 years” from the simultaneous implementation of a new SCM suite. (Ex. P0075R (Cox Project Catalyst) at COX002972;

30. The same software is sold to large and mid-market customers. The FMS and HRM software sold to large customers is the same as that sold to mid-market customers.

(Tr. 819:8-11 (R. Allen), 1787:25-1788:2 (P. Wilmington), 3436:24-3437:11 (S. Catz); Ex. D7166 (D. Morea 5/17/04 Dep.) at 18:15-19:2 (AMS); Ex. P3179 (Ciandrini 1/16/04 Dep.) at 235:15-22 (Oracle).) All the vendors – including Oracle, SAP, and PeopleSoft – have a single product “and that one product is sold up and down the line” to customers of all sizes. (P3171 (Ellison 1/20/04 Dep.) at 148:10-151:15.) Similarly, ADP offers the same product, Enterprise HR, to its clients with 1,500 employees as it does to United Technologies, which has approximately 90,000 employees. (Tr. 4089:10-11,4100:5-9 (J. Rising).) While some EAS vendors have introduced special licensing packages of FMS and HRM that are marketed to smaller customers, the actual software code in the FMS and HRM products sold to both large and mid-market customers is not any different. (Ex. P3070 (Prestipino 5/18/04 Dep.) at 35:19-36:10 (Oracle); Tr. 3437:5-9 (S. Catz).)

31. Oracle’s small customer initiative is based upon the technology also sold to larger customers. Oracle has recently launched its E-Business Suite Special Edition to appeal to its smallest customers – those who can use only 50 seats or less. It contains the same code as the software sold to the largest and middle-sized customers, but which arrives pre- configured by the consulting organization. (Tr. 3436:24-3438:5 (S. Catz).) It contains a subset of several of the same modules found in Oracle’s E-Business Suite, including FMS but excluding HRM. (Tr. 3437:5-11 (S. Catz); Ex. P3070 (Prestipino 5/18/04 Dep.) at 25:5-22, 32:19-33:19.) The package is aimed at small manufacturing customers, based on its heavy emphasis on those modules.

32. SAP offers the same application suite regardless of customer size. SAP sells a product called MySAP ERP Suite, which includes HRM, FMS, corporate controlling, and corporate services. (Tr. 2811:7-13 (R. Knowles).) MySAP ERP is the same whether sold to a mid-market or large enterprise customer. (Tr. 2819:21-25 (R. Knowles).) SAP offers a product called All-in-One, which is “essentially a scaled-down version of mySAP ERP with a lot of functionality turned off.” (Tr. 2813:20-2814:2 (R. Knowles).) All-in-One is marketed both through an indirect channel of resellers to the $200 million and below customer revenue segment and by SAP’s direct sales force. (Tr. 2813:20-2814:2 (R. Knowles).) SAP also offers a product

called Business One, which is a “packaged software offering” targeting the $200 million and below customer revenue segment and sold through an indirect channel of resellers. (Tr. 2813:10-17 (R. Knowles).)

33. Vendors may market differently to customers of different sizes for reasons unrelated to the complexity of their functional needs. The biggest companies typically have the biggest IT budgets and expect to mix-and-match their software products to optimize performance. (Tr. 1199:13-17 (C. Peters) (testifying Emerson has 45 ERP platform “brands” running within the organization today, and approximately a hundred different instances); Ex. D5169 (Emerson HR Footprint); Ex. 5170 (Emerson FMS Footprint).) Christy Bass confirmed this point, stating that no ERP vendor serves all of the functional needs of Accenture’s Global 2000 clients, and therefore these clients must utilize point solution providers to complete their systems. (Tr. 1657:4-10 (C. Bass); Ex. P3198R (C. Bass 4/21/04 Dep.) at 112:17-114:5, 138:14-139:2,176:11-22; see also Tr. 1219:21-23,1220:5-6 (C. Peters ) (stating no single enterprise-wide ERP solution can meet Emerson Electric’s application needs because Emerson “has demands for higher-level functionality than they offer,” and thus it is “inevitable [Emerson will] have to run a system that accommodates a broad range of software platforms.”).) Smaller customers have more limited budgets and tend to only be able to afford more pre- integrated and even pre-configured software sets that are easier and less expensive to implement. (Ex. 3070 (Prestipino 5/18/04 Dep.) at 32:19-33:19, 37:24-38:12.) This is not to say that small enterprises’ needs necessarily are less complex. PeopleSoft has publicly stated that small and mid-sized businesses are “just as complex” as large organizations. (Ex. D6065 (PeopleSoft article by Jeffrey Read, “How Smaller Companies Can Leverage Big Enterprise Strategies”).)

34. There is strong competition for “non-core” and/or “add-on” HRM and FMS functionality. ““Non-core” functionality is sold by lots of vendors, and bought by many customers. “Almost every company has software from multiple vendors,” even those running Oracle’s e-Business suite for their core needs. (Tr. at 3365:5-11 (R.Wohl).) “It’s very, very common for [Oracle’s] customers not only to have existing best-of-breed players, but to be looking at the best possible choices in the industry, which always involve consideration of best of breed players.” (Tr. at 3373:10-14 (R.Wohl).)

35. Pricing does not set “high-function software” apart. The government was unable to show any differences between the pricing of “high function software” and that of other software. There is nothing in the pricing of EAS that delineates a high-function software market.

36. Pre-integration into suites does not set “high-function software” apart. Enterprise and mid-market customers can purchase EAS in either modules, groups of modules, or suites. Pre-integration into suites does not set “high-function software” apart. In fact, “the full package offering, the full suite, is most attractive to companies other than the largest companies” because of the added inconvenience experienced by larger companies installing new suites (Tr. at 3371:10-3 (R.Wohl).) “Smaller companies, mid-size companies, can consider realistically putting the entire suite in. So, it’s a much more credible prospect for them.” (Tr. at 3371:4-7 (R.Wohl).)

37. Configurability does not set “high-function software” apart. Customers seek solutions, and like pre-integration, configurability is just one choice. All vendors’ software is configurable to some degree, and knowing that a system is configurable tells you nothing about whether it is “high function.”

38. PeopleSoft concedes that the competition is much broader than the government contends. An internal PeopleSoft PowerPoint Presentation entitled “HRMS The Next Generation” states in part: “The key to thinking about the competitive landscape is to expand our scope much wider than what we typically think of as our competition. We typically think of SAP, Oracle, and maybe a few niche players as competitors. But in reality, the scope of companies that deal with human capital is much broader than that. There are all kinds [sic] of consultancies, business process outsourcers, and outsourcing technology vendors that really need to be looked at if we want to examine where is the human capital market going.” (Ex. D7008 (PeopleSoft HRMS PowerPoint) at 8).) PeopleSoft’s HRMS installation competitors include

outsource providers such as ADP, Ceridian and Cyborg (Tr. 1934:15-25 (P. Wilmington)), Tesseract and GEAC as legacy systems (Tr. 1935:1-4 (P. Wilmington)), and homegrown systems (1935:5-11 (P. Wilmington). Mr. Wilmington acknowledged that ADP is the number three worldwide vendor of HR payroll and Lawson is fifth. (Tr. 1927:10-19 (P. Wilmington).) Additionally, Mr. Wilmington acknowledged that SAP is its number one competitor, including for HRM software in the United States. (Tr. 1941:10-22 (P. Wilmington); D6011 (Campaign Definition for the Crush SAP Campaign).) Moreover, PeopleSoft competed with Lawson on 27 “enterprise deals” within one year (as opposed to SAP 33 times and Oracle 38) (Tr. 1859:7-11 (P. Wilmington)), and, contrary to plaintiffs’ contentions, Mr. Wilmington acknowledged that customers consider “do nothing” as an option for delaying the purchase of software. (Tr. 1792:17-24 (P. Wilmington).) The breadth of alternatives was underscored in PeopleSoft’s 2003 Operations Review, where legacy and homegrown systems, outsourcing, and Lawson, were all shown as competitors of PeopleSoft for global HRM installations. (Ex. D6320R (PeopleSoft 2003 Operations Review) at 42.)

39. Evidence from systems integrators also shows the wide range of competitive alternatives customers considered. IBM’s Nancy Thomas explained that, in the ERP selection process for large and complex commercial enterprises, IBM advises its clients that they have a “continuum” of “several options,” including best of breed, hybrid, and single vendor solutions integrated by custom code or specialized software solutions, such as Hyperion. The continuum of options presented by IBM to its clients also includes the possibility that clients may elect the status quo and remain with their legacy system. (Ex. D5231R (IBM Financial Systems Evaluation Slides) at IBM 500757-758; Tr. 532:14-534:9 (N.Thomas).) Similarly, in a presentation to the State of North Carolina, Deloitte evaluated several approaches (an ERP system, outsourcing, stand alone packages, custom development, enhancements to current system, and best of breed), before noting that the State had three options: an ERP system, enhancements to current system, or a phased best of breed approach including outsourcing. (Ex. D5137 (Deloitte North Carolina Evaluation) at 7-8.)

Sixteen percent of all IBM ERP implementations since 2001 have been for Lawson, Microsoft Great Plains, Axapta, and Hyperion – i.e., application vendors other than SAP, Oracle, PeopleSoft, and J.D. Edwards. (Ex. P0900 (IBM Deal Summaries) at 1-4.)

BearingPoint has a Lawson alliance and has recently augmented its Lawson practice by hiring all of the Lawson specialists of the former Arthur Andersen & Co. (Tr. 969:11-23 (P. Keating); Ex. D5050 (BearingPoint Lawson End-to-End Solutions) at BP-LIT-E(1)(000246-251).) As a result, BearingPoint consultants “have delivered more than 50 successful Lawson implementations or upgrades across many industries.” (Ex. D5050 (BearingPoint Lawson End-to-End Solutions) at BP-LIT-E(1)(000247).)

40. State plaintiffs have competitive alternatives other than Oracle, PeopleSoft, and SAP. Among state customers for example, plaintiff Michigan considered a wide range of alternatives, and in fact selected Lawson HRM after it outscored Oracle’s HRM software. (Ex. D8103 (D. Johnson 5/28/04 Dep.) at 40:15-23.) Michigan uses FMS provided by KPMG and remains “quite satisfied with it.” (Ex. D8103 (D. Johnson 5/28/04Dep.) at 15:13-17, 19:20 – 20:6.) Plaintiffs Massachusetts and Minnesota are among the many states that has selected AMS; other states which run AMS include Illinois, Missouri, Kentucky, Louisiana and Nevada. (Ex. D5137 (Deloitte North Carolina Evaluation) at 14-15).) Moreover, Arizona’s selection of Lawson HRM software (over PeopleSoft and SAP) further shows that public sector customers pick from a wide range of competitive alternatives. (D8101 (T. Boncoskey 5/21/04 Dep.) at 49:1-18.) Finally, outsourcing is also an option for states. In a presentation made by Deloitte to the State of North Carolina, outsourcing the State’s HR function was presented as a viable option. Indeed, Deloitte noted that the State of Florida had outsourced its HR/payroll

function. (Ex. D5137 (Deloitte North Carolina Evaluation) at 14; see also Tr. 3150:8-14 (M. Sternklar).)

41. Federal customers have competitive alternatives to Oracle, PeopleSoft, and SAP. There are other federal agencies (e.g., NFC and DOI) which provide “cross-servicing” – in effect outsourcing their systems and expertise to serve other agencies’ ERP needs. (Ex. P3059 (M. Johnson 1/13/04 Dep.) at 254:19-257:4.) Also, over 60 federal agencies run AMS for their FM software needs

42. Private-sector customers have competitive alternatives to Oracle, PeopleSoft, and SAP. Private-sector customers consider many options as well. FleetBoston considered a wide range of options outside of PeopleSoft and Oracle, including outsourcing their HRMS. (Tr. 3290:12-3291:9 (B. Mearns).) Kerr-McGee initially considered PeopleSoft, Oracle, Lawson, SAP, J.D. Edwards and Baan for its human resources software needs. (Ex. P3062 (S. Elliot 05/20/04 Dep.) at 50:9-51:7; 51:12-24.) The Lutheran Church’s Worker Benefits Plan (“WBP”) considered SAP, Oracle, PeopleSoft, Lawson, and JD Edwards for its EAS software needs. (Ex. D8105 (R. Cushman 05/18/04 Dep.) at 42:21-43:2.) Similarly, Nextel acknowledged that is considered at least five different options for an overall IT approach. (Tr. 1095:6-1096:5 (R. Cichanowicz); see also id. at 1094:20-23 (R. Cichanowicz) (“Yes, there are choices out there. We had choices.”); P4080R at 18).)

43. Large enterprises already have core HRM and FMS. However defined, most if not all large customers already have software installed and running today to perform their core HRM and FMS business processes. (Tr. 958:18-22 (P. Keating) (testifying that virtually all BearingPoint Tier 1 clients have core HRM and FMS software installed); 4239:18-21 (D. Kutnick) (“[o]n the high end, it’s a mature market, “ and most such customers have “already bought and they already have it.”); Ex. D7173R (D. Burgum 5/13/04 Dep.) at 174:7-9

(“Certainly I would expect that all Fortune 500 companies have got a financial system in place.”); Ex. P3059 (M. Johnson 1/13/04 Dep.) at 134:6-135:5 (“Today, I’m not seeing a lot of core HR procurements, I will say. A lot of those procurements have been done many years ago. A lot of what I see today, from an HR perspective, are the surrounding modules...”); Ex. D7143 (S. Mills 5/27/04 Dep.) at 70:5-23.) According to Steve Mills, the head of IBM’s software business, most large enterprises already have a core FMS and HRM solution – either a home-grown solution or package software. (Ex. D7143 (S. Mills 5/27/04 Dep.) at 70:5-23.).) Microsoft, too, recognizes that the companies that the government is concerned about already have HRM and FMS software. Mr. Burgum of Microsoft testified that the large enterprise market is saturated and that competition is for replacement sales. (Tr. 3110:13-3111:10 (D. Burgum).)¹ Moreover, one of the risks of acquiring SAP that was identified to the Microsoft Board of Directors was: “Core Business of ERP is a mature market with Fortune 1000 already largely penetrated. Continued market share gains may prove difficult.” (D5332R (Project Constellation, board of Directors Appendix, January 5, 2004), at 19304; Tr. 3111:23-3113:17 (D. Burgum).)

44. Customers regularly evaluate their competitive options, including their option to do nothing. Customers regularly evaluate their core HRM and FMS alternatives, including staying with an incumbent system and doing nothing. (Tr. 580:3-9 (L. Bradley).) Customers often begin the software selection process recognizing that they may not make a purchase. (Tr. 1503:19-22 (M. Glover) (Greyhound considered doing nothing, procuring an HRM suite, or doing something in between).) A significant percentage of customers decide to “do nothing,” or not to purchase a new core HRM or FMS product, even after going through a vendor selection process. (Tr. 2710:13-23 (T. Campbell); 1467:16-1468:3 (M. Glover) (Greyhound evaluated an HR system in 1997 but decided to do nothing); 1490:7-22 (M. Glover) (Greyhound evaluated an HR system from 2001 through 2003 and again elected to do nothing).)

Accenture estimates that as many as 20-30 percent of its clients evaluate software but elect to do nothing. (Tr. 1673:14-21 (C. Bass).) BearingPoint estimates that as many as 10 percent of its clients elect to do nothing after an evaluation. (Tr. 930:2-12 (P. Keating).) Mr. Ayala of Microsoft recognizes that one of the difficulties in selling to large enterprises is that they have the option of continuing to run what they already have. (Ex. P3255R (O. Ayala 5/18/04 Dep.) at 52:24-54:9.) Vendors recognize that doing nothing is an option for customers, and will discount accordingly. A PeopleSoft Executive Briefing Document regarding a sales opportunity at United Parcel Service, for example, states that the “Competition” is “Stay with current applications or building in-house.” (Ex. D6225 (PeopleSoft UPS briefing).) A similar document for Fannie Mae dated 03-11-03 States: “We seem to be competing with homegrown and niche solutions.” (see also D6023 (PeopleSoft Executive Briefing Document for Fannie Mae).) (See, e.g., Ex. D7017 (PeopleSoft Laidlaw Briefing) at PS-T2843836 (competition listed as “do nothing” for HRM procurement).) Customers use the threat of doing nothing strategically. Additionally, Emerson, for instance, told Oracle during pricing negotiations that if Oracle did not offer a certain price, there would not be a deal. (Tr. 1245:14-19 (Charles Peters) (testifying that during the Oracle software negotiations, Emerson told Oracle “to make this a deal, you need to meet this price. So it’s either this price or no deal. . . [a]nd no deal, in my mind, is synonymous with in-house integration of best-of-breed products and, as we’ve discussed subsequently, continual maintenance of legacy systems.”) ( Tr. 3214:1-6, 3215:17-23 (K. Harris) (“I’ve had a number of negotiations where my option was do nothing or do something with technology. I mean, those are two viable options, and the important thing . . .is that you make sure that the provider of the technology that you’re negotiating with just understands that, by the way, your other option is do nothing.”)

45. Customers can easily extend the life of their current EAS Systems. Customers expect their EAS systems to last. Curtis Wolfe of the State of North Dakota expects the State’s newly implemented PeopleSoft system to last 20-30 years; indeed, the State has been operating its legacy systems for approximately that period of time. (Tr. 1588:15-1591:7, 1572:14-24, 1591:1-7 (C.Wolfe).) Following Neiman’s recent purchase, Phillip Maxwell

testified that he had no intention of buying another FMS system during his lifetime. (Tr. 722:24-723:6 (P. Maxwell).) Similarly, North Dakota expects to use its ERP system for “30 to 40 years.” (Tr. 1572:16-24(C. Wolfe).) Nextel expects to be out of the market for new ERP systems for ten years (Tr. 1107:16-19 (R. Cichanowicz), while Cox Communications expects no new purchases for FMS and HR for five to ten years. (Tr. 189:12-18 (S. Hatfield).) This is not an unreasonable expectation; core HRM and FMS are mature, stable products, and a customer can keep its system “fresh” by staying on maintenance. There also is a growing availability of low-cost programming resources available that enable a company to extend the life of an incumbent system. (Tr. 1216:8-10 (C. Peters).) Vendors in Asia offer support and maintenance services for enterprise application software products. (Tr. 1214:17-24 (C. Peters).) This offshore support is significantly cheaper. (Tr. 1212:25-1213:2 (C. Peters) (the difference in cost is six-to-one in favor of the Philippines).) Emerson currently has 400 people deployed in the Philippines, helping 30 different Emerson divisions with their back office functions. (Tr. 1213:10-19 (C. Peters).) Some of these employees are “very, very high-end CPAs.” (Tr. 1213:22-25 (C. Peters).) Peters is currently working on providing support for the home-grown EAS systems of Copeland, a $2 billion division of Emerson, in the Philippines/India. (Tr. 1212:2-5 (C. Peters).) Copeland is seeking six people, at $15,000/year, “[s]o for a hundred thousand dollars a year, they feel they can maintain their-their current system and operate it.” (Tr. 1212:8-12 (C. Peters); see also 1467:16-18; 1490:7-15; 1490:23-1491:6; 1466:17-25 (M. Glover) (After evaluations in 1997 and 2001, Greyhound continues to use a legacy HR system that Ms. Glover described as “very old.”), 707:8-10; 722:8-12; 722:13-20 (P. Maxwell) (Neiman Marcus still relies on a 20-year-old legacy system for merchandising accounts payable, a homegrown payroll summary software that was developed almost 30 years ago, and in-house applications for HR workbench functions that were developed eight years ago.) The availability of such low-cost, off-shore programming resources makes the customers “do nothing” threat very credible.

46. Lawson has core HRM and FMS functionality. Lawson sells FMS and

HRM products, (Tr. 3591:5-8 (J. Coughlan)), and offers core FMS and HRM functionality. (Tr. 3591:13-17, 3592:3-4 (J. Coughlan); Ex. P3062 (S. Elliot 05/20/04 Dep.) at 33:9-25.) Professor Iansiti agreed that Lawson has core HRM and FMS functionality comparable to the functionality of Oracle, PeopleSoft, and SAP. (Tr. 2095:13-21 (M. Iansiti).) Industry analyst Dale Kutnick testified that if PeopleSoft were acquired or otherwise somehow did not survive, his largest and most complex customers could turn to Lawson as an alternative for HR and FMS software. (Tr. 4238:17-4239:11 (D. Kutnick).)

47. Many large companies use Lawson. Many large customers use Lawson for core HRM and/or FMS functionality. (See Ex. D7065 (Rule 1006 Summary of PeopleSoft Customer Data) at 1 (PeopleSoft data identifies a number of customers with more than $1 billion in revenue or more than 10,000 employees for whom Lawson was identified as the competition for “core” HRMS or FMS sales). For instance, Lawson’s April 2004 Strategic Account Forum was attended by a variety of large Lawson clients including: Catholic Healthcare; Commerce Bancorp; Federal Reserve Bank; HCA; Johnson & Johnson; Mayo Foundation; McGraw Hill; Northrop Grumman IT; Pathmark; Safeway; Sara Lee; Schlumberger; the State of Arizona; the State of Michigan; Sutter Health; TIAA-CREF; and the University of Wisconsin. (Ex. D7002R (Lawson Strategic Account Forum; Tr. 3268:19-3629:24 (J.Coughlan).) (See also Ex. D7038 (Internal PeopleSoft Profile of Lawson) at 6 (Lawson customers include “nine Fortune 500 commercial banks and nine Fortune 500 insurance companies” including reference customers “TCF Bank, Manufacturer’s Life, Commerce Bancorp, Federal Reserve Bank, Northern Trust,” “eight of the top 15 integrated delivery networks” in healthcare, “five of the top 10 U.S.-based retailers, eight of the top 20 apparel retailers, seven of the top 25 grocery chains, 23 of the top 100 restaurant chains and 20 of the top 100 specialty chains,” public sector reference customers like State of Arizona, Atlanta Public Schools, State of Michigan, State of South Dakota, City of Greensboro North Carolina, and Denver Public Schools and professional services firms include Lucent Technologies, Cisco Systems, Schlumberger, Watson Wyatt, and Northrop Grumman”); D5931 (Executive Briefing Document/McGraw Hill Companies) at 1 (McGraw Hill uses Lawson HR and Payroll); Ex. D6204 (PeopleSoft Executive Briefing on Kinkos) at 1 (Kinkos uses Lawson financials).)

48. Lawson competes for business from other large companies. The record also contains many examples of Lawson’s competing for large customers against Oracle, PeopleSoft, and others. (See, e.g., Ex. P2356 (PeopleSoft How to Beat Lawson), at 21-31 (analyzing PeopleSoft wins versus Lawson at Amerigroup, Beverly Enterprises, Roundy’s, Albertsons, Dixon’s, South Financial Group, Carlson Companies, Occidental Petroleum, and Ocean Energy; Ex. D5907 (PeopleSoft discount approval form for San Diego USD) (PeopleSoft discount approval form for San Diego Unified School District seeking a 56.8% discount based on a “VERY competitive procurement with SAP and Lawson”); see also Ex. D6236 (PeopleSoft USA Product Competition.xls); Ex. D6236A (modified version to recreate spreadsheet format) (shows deals, by competitor, broken out between “Enter.” and “Mid-Mkt” deals, and reflects 27 “Enter.” deals competing against Lawson).) For instance, Lawson competed with SAP, PeopleSoft, Microsoft, Exact, Ultimate, and Oracle for sale of human resources, payroll, and financials software to ExpressJet, a $1.5 billion regional carrier that spun off from Continental Airlines. (Tr. 2839:19-2841:22 (R. Knowles); Ex. D5641R (SAP Discount Approval for ExpressJet) at 1.) Only Lawson and Oracle made ExpressJet’s short-list. (Tr. 2841:23-2842:6 (R. Knowles); Ex. D5641R (SAP Discount Approval for ExpressJet) at 1-2.) Indeed, PeopleSoft’s internal CRM database (“Customer One”) identifies Lawson as its competitor in over 65 deals, all for large customers (more than $1 billion in revenue or more than 10,000 employees). (Ex. D6236 (USA Product Competition) at 77332).

49. AMS has core [HRM and] FMS functionality. AMS’s HRM software includes, but is not limited to, the following functionality: personnel management, payroll, benefits and compensations, competency and management, performance management, and statutory reporting. (Ex. D7166 (D. Morea 5/17/04 Dep.) at 19:3-18.) AMS’s FMS software includes, but is not limited to, the following functionality: accounts payable, accounts receivable, fixed asset management, cash management, and financial reporting. (Ex. D7166 (D. Morea 5/17/04 Dep.) at 19:19-20:9.)

50. AMS is particularly effective in the public sector. AMS is a frequent and successful competitor for public sector procurements against Oracle, PeopleSoft and SAP. (See Tr. 2883:20-2884:1 (R. Knowles) (SAP competes with AMS frequently in the federal space); Tr. 972:6-15 (P. Keating) (AMS is a viable competitor for large and complex federal procurements, and BearingPoint has lost large procurements to AMS); Tr. 363:14-364:2 (R. Bergquist) (PeopleSoft competes with AMS for FMS deals in the public sector); D5699 (SAP State System List) at SAP 17899-17904 (list of software vendors for each state shows 21 states using AMS for HRM, FMS, or both).)

Among AMS’s other customers are

(Ex. D7166 (D. Morea 5/17/04 Dep.) at 76:24 - 79:22, 81:22 - 83:12). In reality, AMS has had more successful implementations in public sector customers over the past five years than all of its competitors combined. (Ex. D7166 (D. Morea 5/17/04 Dep.) at 102:7 - 12.)

51. DOJ itself chose AMS. Indeed, the government’s competitive effects theory does not even account for its own purchasing choices. Plaintiffs trivialize AMS even though some plaintiffs - including                      and                      - are themselves AMS licensees for HR and/or FMS.

Most notably, the Department of Justice itself recently chose AMS FMS over Oracle, PeopleSoft, and SAP. (Ex. D7166 (D. Morea 5/17/04 Dep.) at 21:22-23:21). Finally, AMS has more desirable federal, state, and local public entities as its customers than any other ERP provider. (Ex. P3059 (M. Johnson 1/13/04 Dep.) at 254:19-257:4.)

52. Microsoft has core HRM and FMS functionality. The evidence is overwhelming that Microsoft competes in the applications space, including for large customers. Microsoft has a division called Microsoft Business Solutions (“MBS”), which sells EAS. (Tr. 2972:14-18 (D. Burgum.) Microsoft has a few different EAS offerings, including Great Plains, Microsoft CRM, Navision, Solomon, and Axapta. (Tr. 2976:2-2977:12 (D. Burgum).) Microsoft’s EAS solutions offer core HRM and FMS functionality. (Ex. D5334 (MBS-Axapta) at 4; D5338R (MBS Axapata 3.0 Technical Positioning) at p. 3069.)

53. Microsoft positions its Axapta product for large enterprises. Axapta is positioned for “multinationals, corporations and spokes.” (Tr. 3095:2-7 (D.Burgum).) Axapta has the flexibility and scalabilty needed to meet enterprise-scale ERP demands. (D5334 (Microsoft Business Solutions – Axapta brochure) at 1.) A March 2004 internal Microsoft email explains the scalability in the context of proposed answers to customer concerns about Microsoft’s statements about this case:

Our recent sworn statements to the US Justice Department was intended to define our plans with respect to which customer segments we will proactively market our ERP solutions. It is not intended to say our products are not suitable for larger enterprises. To the contrary, Axapta has the flexibility and scalability needed to address enterprise-scale ERP demands such as those of [CUSTOMER] . We are confident our existing customer references will provide the evidence you need to verify this. Because of our desire to establish a solid footing in the mid-market space, and because of our existing strong relationship with ERP vendors focusing on the large enterprise space (organizations of greater than $5 billion in revenue), we are not proactively marketing our solutions to the largest enterprises.

(Ex. D5359 (3/15/04 email re “MBS into Enterprise – Oracle Testimony”) at 2407.)

54. Large companies use Microsoft. The record contains several examples of large customers that use Microsoft to meet core needs.

55. Microsoft markets to large customers. MBS currently markets its products (primarily Axapta) to large organizations in what it calls the Corporate Account Space. As Mr. Burgum acknowledged, those companies have annual IT spends of up to $45 million and are not “mom and pop” companies. (Tr. 3092:8-3092:23 (D. Burgum).) Indeed, at its customer conference, Microsoft executives met with prospects such as a global leading textile manufacturer with 15,000 employees, a $2 billion per year poultry producer, and an international athletic wear manufacturer with global operations and over $1 billion in annual sales. see also Ex. D7173R (D. Burgum 5/13/04 Dep.) at 168:5-169:6.) In other words, by 2006, Microsoft will target the “hubs” of all but the largest 1,600 companies in the world. (Tr. 3091:18-3092:7 (D. Burgum).) And Microsoft will be poised to serve divisions or spokes of the largest 1,600 corporations in the world. (Tr. 3097:1-6 (D. Burgum).) As an example, earlier this year, Microsoft competed with SAP, PeopleSoft, Lawson, Exact, Ultimate, and Oracle for sale of human resources, payroll, and financials software to ExpressJet, a $1.5 billion regional carrier that was spun off from Continental Airlines. (Tr. 2839:19-2841:22 (R. Knowles); Ex. D4641R (SAP Discount Approval for ExpressJet) at 1.) If there were any doubt about its intentions, they are answered by the recent alliance between Microsoft and one of the leading systems integrators, BearingPoint.

In short, Microsoft has dramatically accelerated its efforts in the business applications and business services space in the last two years, and “we’re at our peril if we’re ignoring Microsoft in ...[a]ny space to do with applications and business — business applications and business services.” (Tr. 4241:13-4242:6 (D. Kutnick).)

56. Microsoft’s Project Green is Microsoft’s global EAS project. Microsoft is working on its next generation EAS, code-named “Project Green,” that will include a broad range of functionality, including HRM, FMS, SCM, CRM and analytics. (Tr. 3104:13-19 (D. Burgum).) Microsoft intends to target what it terms the “Corporate Account Space” with “Project Green.” (Ex. D5343 (7/30/03 Green Update) at 19441). Microsoft is expending billions of dollars on this project and its internal documents show that “[u]ltimately, ‘Green’ will scale to serve business needs of small businesses to large corporations.” (Ex. D5344R (Green Strategy Team Feedback Session, July 2003) at 10772.)

57. SAP has prioritized “containing” MBS, its ultimate EAS competitor. Numerous SAP internal documents demonstrate increasing concern over Microsoft. In a January 23, 2004 e-mail the CEO of SAP America wrote in reference to Microsoft, “These guys are here!” (Ex. D5636 (SAP E-mail re: Microsoft Q2 2004) at 1; Tr. 2892:2-10 (R. Knowles).) SAP determined by October 2003 that “the Microsoft warning signs should be taken very seriously, as Microsoft is venturing out into SAP territory” and that “the immediate containment of Microsoft in the Business Solution Space must be a priority.” (Ex. D5637 (SAP Email re Findings from the Global Win/Loss Program) at 6; Tr. 2897:9-2898:21 (R. Knowles).) SAP added Microsoft to the list of six most closely watched competitors earlier this year. (Tr. 2862:16-2863:8 (R. Knowles)), and of those, only Microsoft had positive market share gains and revenue gains during 2003 in the EAS market. (Tr. 2863:3-12 (R. Knowles); Tr. 2895:13-2897:8 (R. Knowles); Ex. D5637 (SAP Email re Findings from the Global Win/Loss Program) at 6.) Finally, Leo Apotheker, President, Global Field Operations and Member of the Executive Board at SAP, predicted in an April 20, 2003 email: “We believe that there is going to be two end to end enterprise apps. vendors in the future, SAP and Microsoft.” (Ex. D6018 (PeopleSoft and SAP E-mails Regarding Microsoft).)

58. Oracle believes Microsoft will be a very significant and successful competitor in the EAS market. Larry Ellison sees Microsoft becoming Oracle’s number one U.S. competitor “within 24 months.” (Ex. P3171(Ellison 1/20/04 Dep.) at 181:22-182:2.) Keith Block testified that Microsoft is coming on strong and should be considered a tier one vendor. Ex. P0908 (Oracle Q3 Recap) at 2 (noting that Microsoft was mentioned in license approvals in Q3 2004 50% more than in Q2, which was up 30% more from Ql); Ex. P3033 (J. Henley 5/4/04 Dep.) at 197:1-4; Ex. P3192 (K. Block 1/5/04 Dep.) at 387:10-21.)

59. PeopleSoft believes Microsoft will be a very significant and successful competitor the EAS market. PeopleSoft believes Microsoft is a threat. When Microsoft announced it was buying Great Plains, Renee Lorton, General Manager of PeopleSoft’s FMS division, emailed Ram Gupta, PeopleSoft’s Executive Vice President of Products and Technology, with the subject: “Breaking News: Microsoft Buys Great Plains YIKES.” (Ex. D6316 (PeopleSoft e-mail re Microsoft’s Purchase of Great Plains).) In the e-mail, Ms. Lorton says: “This is huge news and should have all of us in the Financials world (and others) shaking in our boots .... And I don’t buy the claims that they are only targeting the mid-market space. As we all know, if he [Gates] can’t own the market, he doesn’t bother. He’s entering with a solid, high quality mid-market product that has good analytical tools and a segue into SCM and CRM. I think he has a gun at our and Oracle’s back.” (Ex. D6316 (PeopleSoft e-mail re Microsoft’s Purchase of Great Plains).) Doug Merritt, then-PeopleSoft General Manager of HCM, said regarding PeopleSoft’s competition: “Of everyone I’m going to mention, I think the biggest long-term threat is Microsoft.” (Ex. D7072 (PeopleSoft Daily Planet Merritt interview).) PeopleSoft’s internal “Competitive Selling Guide for Microsoft Business Solutions” states: “Microsoft executives have publicly stated that, because of its close relationships with a variety of enterprise application vendors (i.e. PeopleSoft and SAP), it does not plan to move up market and cause conflict with its strategy to address large enterprise customers. Important Note: Realistically, however, we have seen MSGP compete in many up-market opportunities. MSGP will state that they are looking at organizations that are less than $400M or so (even smaller as a ‘sweet-spot’). Steve Ballmer has even told our executive team that MBS intends to remain in the

smaller market, however, it is clear that this is not the case. We have seen Microsoft in deals up to $800M and even above.” (Ex. D6057 (PeopleSoft Competitive Selling Guide for Microsoft Business Solutions) (emphasis in original).)

60. Systems integrators believe Microsoft will be a very significant and successful competitor the EAS market. Perry Keating believes that Microsoft will be a strong player in the EAS market: “Microsoft’s not a company that plays for second place.” (Tr. 926:24 (P. Keating).) BearingPoint identified Microsoft (along with IBM and Siebel) as a potential entrant into the Tier One market in response to a questionnaire issued by the European Union. The only barrier to entry identified by BearingPoint was “self-choice.” (Ex. P0203; Tr. 937:14-938:13 (P. Keating).) “[O]ur logic was that given the size, balance sheet, cash position, marketing force of Microsoft, IBM, and Siebel... [they] have the capability, the money, the marketing, the client base, to move into those – to move into that [Tier 1 ERP] market.” (Tr. 921:17-25 (P. Keating).) Keating’s belief is that Microsoft “absolutely” plans to ultimately target the Tier-One client base, the Global 2000 and 500. (Tr. 942:7-13 (P. Keating).)

Ms. Bass of Accenture indicated that “[i]t’s just a matter of time” before Microsoft is competing with SAP. (Tr. 1640:8-10 (C.Bass).)

61. The government’s experts acknowledged Microsoft’s threat. Professor Elzinga agreed that “Microsoft has a lot of money, and you can only speculate... what would happen if they threw a lot of it into this sector . . .” (Tr. 2244:16-2245:3 (K. Elzinga).) Marco Iansiti stated that “Microsoft is a successful company and that they [sic] have been very innovative in a variety of areas.” (Tr. 2114:13-15 (M. Iansiti).) As an expert for Microsoft in prior litigation, Marco Iansiti stated that Microsoft had an “innovative development” organization, and is “staffed with motivated, knowledgeable individuals with a keen passion for developing software products.” (Tr. 2114:7-9, 16-20 (M. Iansiti).)

62. Outsourcing is a competitive alternative. “Outsourcing” refers to the third-party delivery of a service necessary for a client to run its operation. (Tr. 3128:6-12 (M. Sternklar).) Outsourcing of core HRM business processes is a viable alternative to using EAS software. (Ex. D7134R (Fleet’s HR/Payroll Outsourcing Analysis) at A.4.a – A.4.b. Moreover, Verizon admitted that it could outsource its entire financial management system, an option that is “constantly under evaluation at some level.” (Tr. 604:15-21 (L. Bradley).) Laurette Bradley of Verizon, for example, testified that it would be possible for Verizon to outsource both its HR and its financial management functions, and that Verizon “constantly” evaluates whether it should do so. (Tr. 604:15-21 (L. Bradley).)

. (Ex. P3200 (L. Anderson 5/13/04 Dep.) at 133:15-134:10 (under seal).) Accenture offers financial services outsourcing. (Ex. D7156R (Z. Brand 5/17/04 Dep.) at 49:16-19.) The EAS vendors recognize outsourcing as an important competitor in the large enterprise market. (Tr. 4255:22-4256:11 (L. Ellison); Ex. D7008 (PeopleSoft HRMS PowerPoint) at 8.) There is “a market trend for large companies to consider outsourcing, as well as software solutions, in order to meet their global HR payroll administration requirements.” (Ex. D7156R (Z. Brand 5/17/04 Dep.) at 21:23-22:5, 15-25.); see also Tr. 3148:10-20 (M. Sternklar).) The outsourcing field has become more complex in recent years, with different business models, and, more competitors –or as Mr. Rising of ADP put it, “there are more people on the dance floor.” (Tr. 4096:11-14 (J. Rising); see also Tr. 3148:10-20 (M. Sternklar).) Outsourcing has long-term financial benefits because clients can avoid the recurring and significant expense of upgrading and maintaining the technology. (Tr. 3133:2-10, 17-24 (M. Sternklar).)

63. Many large companies use BPO as a viable alternative to using EAS software. Business process outsourcing (“BPO”) refers to providing services that are necessary to support a client’s entire business operation, such as running the client’s HR delivery organization. (Tr. 3128:13-20 (M. Sternklar).) Hewitt, Exult, Fidelity, Ceridian, Convergys, Paychecks, and Affiliated Computer Systems offer BPO. (Tr. 3128:24-3129:20, 3149:7-9, 3150:8-10 (M. Sternklar), 4095:20-4096:4 (J. Rising); (Ex. D7061 (Kerr-McGee HR

Outsourcing Assessment) at 30.).) Some systems integrators offer BPO, including IBM Global Services, Deloitte and Accenture. (Tr. 3150:20-23, 3151:2-13 (M. Sternklar)

Mellon Financial and Aon (the insurance company), also have HR outsourcing businesses. (Tr. 3150:15-3151:1 (M. Sternklar). There are many large companies currently using BPO. (Tr. 3149:1-3150:2 (M. Sternklar) (Exult, which was recently acquired by Hewitt Associates, provides BPO services to Prudential, Bank of Montreal, and International Paper, and had Bank of America as a client before losing it to Fidelity), 3150:3-5 (M. Sternklar) (Hewitt Associates’ clients include Goldman Sachs, Sony, and Medco), 3150:8-11 (M. Sternklar) (Convergys’s clients include the State of Florida, Sodexho Marriott, and RR Donnelly), 3151:2-10 (M. Sternklar) (IBM Global Services provides outsourcing to Proctor and Gamble); Ex. D7156R (Z. Brand 5/17/04 Dep.) at 66:3-13 (IBM services Proctor and Gamble).)

64. Some outsourcers run their own software. Some outsourcers run their own software. (Tr. 3158:15-19 (M. Sternklar) (Hewitt now runs the Cyborg technology it acquired). Fidelity, for example, uses a combination of commercially available products and proprietary technology that has been built or acquired by Fidelity. (Tr. 3129:21-3130:1 (M. Sternklar; Ex. 7020 (HRO: The Wave of the Future?) at 21.) As part of its “Project Voyager,” Fidelity will migrate its clients to its proprietary HR Access within the next 18 to 24 months. (Tr. 3154:3-15 (M. Sternklar).) Similarly, since acquiring its own software code from PeopleSoft to form the base of its proprietary “Enterprise HR” system, ADP has released eight new versions of the software. (Tr. 4085:16-4086:11, 4086:22-25 (J. Rising).) In all of these cases, the customer negotiates a price for “services” with the outsourcer and does not deal directly with the software vendor.

65. ADP offers outsourcing for large companies. ADP’s Employer Services Business Division is responsible for providing HR, payroll, and benefit services to corporate clients (Tr. 4081:23-4082:1 (J. Rising).) ADP’s annual revenues are approximately $7 billion, of which over $4 billion comes from the Employer Services Division. (Tr. 4083:25-4084:7 (J.

Rising).) ADP’s HRMS product, named Enterprise HR, is an HR application designed to meet clients’ HR needs while interfacing with ADP’s payroll processing service. (Tr. 4085:25-4086:11 (J. Rising).) ADP’s Enterprise HR is “a very robust HR solution.” (Ex. D7160 (G. Secord 2/19/04 Dep.) at 35:23-36:13.) Companies with multiple unions can use ADP’s Enterprise HR, and such companies have not had any trouble using Enterprise HR to meet their HR needs. (Tr. 4090:4-10 (J. Rising).)

66. ADP competes against EAS vendors. In selling Enterprise HR, ADP competes against Oracle, PeopleSoft, SAP, Lawson, IBM, and incumbent/legacy systems. (Tr. 4091:9-4092:15 (J. Rising).) ADP competes with software vendors such as PeopleSoft “at every. . . stage of the sales cycle, from the original qualification to the final decision,” i.e., the “bake off.” (Tr. 4095:10-19 (J. Rising).) Compared to ERP vendors, ADP offers “HR functionality that is fairly comparable to ERP providers,” plus additional outsourcing services that the ERP vendors cannot offer. (Ex. D7160 (G. Secord 2/19/04 Dep.) at 68:23-69:22.) Outsourcing with a company such as ADP lowers risk because the outsourcer takes care of all upgrades and maintains the core platforms, allowing the client to avoid the “upgrade treadmill.” (Tr. 4092:16- 4093:9, Tr. 4094:3-20 (J. Rising).) Outsourcing costs less than maintaining an in-house software system due to the many “hidden costs” of such a system. (Tr. 4093:10-16 (J. Rising).) PeopleSoft, for example, identified 38 examples of ADP as a competitor in large customer opportunities for core HRMS and FMS products. (Ex. D7064 (Oracle presentation).)

67. Large companies use ADP. ADP has approximately 30-35 clients with more than 10,000 employees who use Enterprise HR. (Tr. 4098:6-10 (J. Rising).) Those clients include United Technologies, Tyco, Aramark, Comcast, Cintas, Sysco, Xerox, Sempra Energy, Hormel, and ING. (Tr. 4098:11-4099:8, 4100:10-24, 4103:11-13 (J. Rising); Ex. D5010 (ADP Clients with Over 10,000 Pays).) United Technologies services approximately 90,000 employees with Enterprise HR. (Tr. 4089:7-13 (J. Rising).) ADP services approximately 30,000 employees for Xerox; 60,000 employees for Comcast; and 30,000 employees for Cintas. (Tr. 4100:25-4101:8 (J. Rising).); (Tr. 4104:13-25 (J. Rising); Ex. D5010 (ADP Clients with Over 10,000 Pays).)

68. Fidelity offers outsourcing for large companies. Fidelity Employer Services Company recently reorganized to form a single business for delivering more integrated HR outsourcing services to the marketplace. (Tr. 3125:2-23 (M. Sternklar).) Fidelity provides full HR administrative services to its clients, including payroll payment, tax withholding, 401 (k), management tools, learning tools to facilitate training, and customer service relating to benefits and HR. (Tr. 3128:21-3129:20, 3133:25-3134:18, Tr. 3131:24-25 (M. Sternklar).) Fidelity’s outsourcing contracts usually last for a minimum of five years. (Tr. 3141:6-11 (M. Sternklar).) Fidelity’s outsourcing clients always evaluate the HR outsourcing offering on a total cost of ownership basis. (Tr. 3140:13-16 (M.Sternklar).) For its HR outsourcing offering, Fidelity is targeting “larger companies who are pretty complex.” (Tr. 3135:12-3136:1, 3147:5-14 (M. Sternklar).)

69. Many large companies use Fidelity. Fidelity’s clients with more than 10,000 employees include Asea Brown Boveri (“ABB”), Agilent Technologies, Bank of America, First American Corporation, Fidelity Investments, Fleet Boston Financial, IBM, Knight Ridder, and Visteon Corporation. (Tr. 3136:10-14, 3137:7-18, 3141:23-3142:6, 3143:12-21, 3144:14-17, 3145:6-20; 3168:23-25 (M. Sternklar); Ex. D5188 (Fidelity Large Market Client List).) Fidelity is servicing approximately 20,000 to 30,000 employees for First American Corporation. (Ex. D5188 (Fidelity Large Market Client List); Tr. 3143:12-18 (M. Sternklar).) Fidelity is servicing approximately 20,000 employees for Knight Ridder. (Ex. D5188 (Fidelity Large Market Client List); Tr. 3144:18-20 (M. Sternklar).) Fidelity is servicing approximately 10,000 employees for Visteon Corporation. (Ex. D5188 (Fidelity Large Market Client List); Tr. 3145:6-11 (M. Sternklar).) Fidelity’s agreement with ABB, which operates in multiple states, covers approximately 12,000 to 14,000 ABB employees. (Ex. D5188 (Fidelity Large Market Client List); Tr. 3137:7-12, 3138:1-5 (M. Sternklar).) ABB had been considering upgrading its technology with a PeopleSoft or SAP solution, but instead chose to outsource with Fidelity. (Tr. 3138:14-3139:13, 3139:19-3140:10 (M. Sternklar).)

70. Bank of America uses Fidelity. After considering upgrading its PeopleSoft HRM system, Fleet Boston looked to outsourcing because they were concerned with cost of upgrading to PeopleSoft 8.0. (Tr. 3990:12-25 (B. Mearns).) Before contracting with Fidelity, Fleet Boston was using a PeopleSoft platform. (Ex. D7020 (HRO: The Wave of the Future?) at 25; Tr. 3282:13-3283:13, 3296:5-24; Ex. D5188 (Fidelity Large Market Client List); Tr. 3144:2-12 (M. Sternklar).) Outsourcing with Fidelity provided all the functionality and more than the PeopleSoft product. (Tr. 3296:19-24 (B. Mearns).) Bank of America, which recently acquired Fleet Boston, selected Fidelity to outsource all its HR/Payroll functions. (Tr. 3300:19-22 (B. Mearns).) Fidelity’s seven-year agreement with Bank of America will cover approximately 170,000 to 180,000 active employees and 70,000 to 80,000 retirees and will entail a wide range of HR and benefits services, such as payroll, timekeeping, personnel administration, employee records, benefits, and stock plans. (Tr. 3145:12-25, 3146:14-20 (M. Sternklar); Ex. D5189 (Bank of America and Fidelity Press Release) at FMR 001847.)

71. Fidelity purchased HR Access to use as its base software system. Fidelity purchased HR Access, which is an HR and payroll software program, from IBM in the fall of 2003. (Tr. 3152:3-6, 23-25 (M. Sternklar).) Fidelity purchased HR Access because Fidelity “believe[s] strongly in the need to own [its] own technology that is core to [its] business.” (Tr. 3153:7-23 (M. Sternklar).) Fidelity intends to use HR Access as the platform for Fidelity’s HR outsourcing offering; this is part of Fidelity’s focus on serving the broader outsourcing market, instead of continuing to license or acquire software designed for an individual corporation. (Tr. 3153:12-23 (M. Sternklar).) Fidelity purchased HR Access to replace Oracle as Fidelity’s HR processing engine and data repository. (Tr. 3153:24-3154:2 (M. Sternklar).) Fidelity intends to migrate all of its clients to HR Access because it will allow Fidelity to deliver its services more economically and to have better control over the delivery of its services. (Tr. 3154:16-23 (M. Sternklar).) Fidelity will begin servicing clients with HR Access in the next 18-24 months and expects to migrate all of its clients within approximately five years. (Tr. 3154:11-15, 3154:24-3155:3 (M. Sternklar).) HR Access currently has approximately 500 to 600 clients in Europe. (Tr. 3153:1-6 (M. Sternklar).) Fidelity also pursued HR Access to establish a presence outside the United States and to offer global HR and payroll processing solutions to U.S. clients with international operations. (Ex. D5185 (Fidelity Workplace Services Update) at FMR 000980.)

72. The Oracle license is a minimal part of Fidelity’s cost of providing its offering. Fidelity currently has a license to use Oracle HRM software. Under the agreement, Fidelity has a paid up license for several million users and an option to acquire more at normal costs; as a result, the license fee represents less than                 of the cost of providing the outsourcing offering. (Tr. at 3135:3-11 (M. Sternklar); Tr. at 3135:3-11 (M. Sternklar);

73. Customers have other competitive alternatives for addressing their core HRM and FMS needs. As government expert Professor Iansiti acknowledged, an enterprise customer that does not have to use Oracle, PeopleSoft, or SAP software to support their HRM or FMS needs. (Tr. at 2092:19-22 (M. Iansiti) (“a number of different organizations can use more traditional management processes to manage their complexity. There’s a lot of value to these software products, but they also cost a lot of money....”).)Customers can use a variety of other solutions, including other software application suites, home-grown and legacy solutions, point solutions connected over a middleware bus, consolidation software such as Hyperion, system integrators. Tr. 1217:5-1218:9; 1227:71-21; Ex. D5169 (Emerson HRM footprint), Ex D5170 (Emerson FMS footprint).

74. Homegrown and/or custom software and extensions to commercially purchased software are competitive alternatives. Many customers use homegrown and/or custom software and extensions of software. (Tr. 1223:4-6 (C. Peters) (Emerson’s “main Java vendor has done, I think, 60 projects for [Emerson] so far in the last two or three years. So, writing our own is very much an alternative.”), 1210:19-1211 (C. Peters) (Emerson’s Copeland division uses a homegrown FMS system that was made in the 1960s); Ex. D7008 (PeopleSoft HRMS powerpoint) at 8 (Apple uses a homegrown system).) Internal PeopleSoft documents identify custom and homegrown software as competition, (see, e.g., D6202 (PeopleSoft Executive Briefing Document for Sprint Corporation); Ex. D6225 (PeopleSoft UPS briefing);

D6353 (Executive Briefing Document for Wal Mart) (identifies “Competition” as “Wal-Mart IT”).) The evolution of middleware makes homegrown and custom software an even more attractive option. (Tr. 1222:17-21 (C. Peters).)

75. The customer witnesses do not include the government’s “high function” features among their “needs.” Many of the customer witnesses offered by the government as examples of “large and complex” customers that demand “high function software,” in fact, do not have core HRM and FMS business process requirements that distinguish their product demands from those of the so-called mid-market customer demands. (Tr. 1498:4-21 (M. Glover)(Greyhound still uses JD Edwards, which they acknowledged is a “mid-market” product); 1127:21-22 (S. Wesson)(AIMCO has only 6500 employees, which would place it in the mid-market); 1577:12-14 (North Dakota employs only 8430 permanent employees, which would place it in the mid-market).

76. Customer needs are unique and not defined by a customer’s size. Each EAS customer has unique needs and preferences which will define its preferred solution. ; Tr. 1579:13-21 (C. Wolfe) (North Dakota’s procurement and software needs were unique); D8101 (T. Boncoskey Dep.) at 62:11-16 (Arizona’s procurement and needs were unique); Tr. 1485:11-20 (M. Glover)(Greyhound has “a lot of data that... is pretty unique to us,” which produces unique needs in terms of software.) Whether or not the customer is large and complex or mid-market is an artificial distinction and does not predict the customer’s preferred solution or functional needs for a single software platform. (See, e.g.,Tr. 2820:9-19 (R. Knowles) (establishing separation between the mid-market and large enterprise market is “not an exact science”).

77. If anything, large enterprises are more likely than mid-market companies to integrate different solutions into a diverse IT “footprint.” Large customers are more likely

than mid-market customers to use a variety of different EAS and other solutions and to integrate them, including a variety of solutions to meet their core HRM and FMS needs. (Tr. 963:25-964:23 (P. Keating) (estimating 25 percent of BearingPoint Tier 1 clients use best of breed applications connected through a middleware bus instead of an integrated suite of ERP products).) (Ex. D6078 (PeopleSoft Spring CRM Summit: Questions for PeopleSoft and CEO Craig Conway) at 135747 (“I don’t know of many companies that have committed entirely to one vendor end-to-end. Rather, most businesses of any complexity tend to evolve their administrative enterprises through best-of-breed selections over time.”); see also Tr. 755:2-16, 757:21-758:9, and 792:7-21 (R. Allen).) Directly contrary to plaintiffs’ premise that large, complex organizations are the most vulnerable to price increases for ERP suites, very few of IBM clients implement a single ERP vendor. (Ex. D5231R (IBM Financial Systems Evaluation Slides) at IBM 500751; Tr. 531:8-532:16 (N.Thomas).) In the ERP selection process, IBM has informed its clients that it is more difficult to implement a single vendor across financials and HR than it is to use a best-of-breed or hybrid approach; standardization to a single ERP vendor is “disruptive, costly, and culturally difficult,” and thus it must be justified by a “business driver / opportunity.” (Ex. D5231R (IBM Financial Systems Evaluation Slides) at IBM 500757, 762; Tr. 536:7-537:22 (N.Thomas).)

78. If anything, large and complex customers have more options than other customers, not fewer. Large or complex customers are not limited to Oracle, PeopleSoft, or SAP for core HRM and FMS software solutions. As they evaluate their current systems and possible future IT investments, they can consider other solutions including Lawson, AMS, Microsoft, outsourcing, best-of-breed or point solutions, decentralizing business processes, middleware bus integration, consolidation solutions, or changing the way they do business in order to ensure they get the best “value proposition” for meeting their needs. Lawson runs HRM or FMS software throughout many large enterprises, including HCA, Safeway, Liberty Mutual, Manulife, and Schlumberger. (Tr. 3600:1-3642:12 (J. Coughlan).) Emerson Electric uses a variety of legacy systems, custom/homegrown applications, best-of-breed vendors, ERP suite solutions and outsourcing to meet its enterprise application needs. (Ex. D5169 (Emerson HR

Footprint); D5170 (Emerson FMS Footprint); Tr. 1198:4-19 (C. Peters).) Some of IBM’s clients use a hybrid ERP approach; Zurich Insurance Group, for example, uses both SAP and Oracle financials to run the same functionality – financials – in different parts of its business. (Tr. 537:24-539:19 (N. Thomas). Best of breed approaches are used by Royal Bank of Canada, Wachovia, and Morgan Stanley Dean Witter. (Ex. D5231R (IBM Financial Systems Evaluation Slides) at IBM 500760; Tr. 535:3-19 (N. Thomas).) Some of IBM’s clients, such as J.C. Penney’s, elect to maintain a fragmented ERP footprint and use Hyperion to tie together these disparate parts of their enterprise. (Tr. 547:7-21 (N. Thomas).) Hyperion supports several functionalities, including consolidation, planning, budgeting, forecasting, and analysis reporting. One of the things that Hyperion allows a large and complex enterprise to do is to have its subsidiaries or divisions running on different financial management systems, and then it extracts the data and rolls it into a consolidated report for the parent. (Tr. 546:6-547:21 (N.Thomas).)

79. The government’s witnesses do not all have global needs. Many of the government’s customer witnesses do not have global core HRM and FMS demands. (Tr. 154:13-15 (S. Hatfield) (Cox Communications operates in the US only); 1089:5-7 (R. Cichanowicz) (Nextel operates in the US only); 1576:21-23 (C. Wolfe) (North Dakota operates in the US only); 1746:15-20 (K. Johnsen) (PepsiAmericas operates in only one currency, and only in English); 1153:4-9 (S. Wesson) (AIMCO operates only in one currency, and only in English).); 695:10-23 (P. Maxwell)(Neiman Marcus operates stores only within the U.S., operates only in one currency, and only in one language).)

80. Alternatives to Oracle, PeopleSoft, and SAP also offer global support. Lawson Software supports global core HRM and FMS demands. Lawson supports all major currencies, presently provides language support for French, French Canadian, Spanish, and

English, and has been implemented by customers such as Schlumberger in more than 100 countries. (Tr. 3645:16-3646:6; 3744:1-3 (J. Coughlan).) Likewise, AMS supports global core FMS needs for public sector customers, and has been installed in 168 countries for the Department of State and handles multi-currency transactions. (Ex. D7166 (D. Morea 5/7/04 Dep.) at 18:6-14; 84:13-86:11.) Microsoft also supports global core HRM and FMS demands. (Tr. 3085:25-3086:12 (D. Burgum); Ex. D5334 (“A Global Solution”) at p.2.) Esselte, for example, chose Microsoft Axapta to support over 6,000 employees worldwide because of its multi-language support. (Ex. D7159 (L. Spund Dep.) at 242:15-25; Ex. D6548 (Esselte presentation) at 9, 12.) Esselte described Axapta as “ideal for maintaining multi-site, multi-country companies” like Esselte. (Ex. D7159 (L. Spund Dep.) at 243:1-17.) Additionally, Outsourcers support global core HRM and FMS demands. ADP for example currently has payroll engines in 26 different countries. (Tr. 4089:24-4090:3 (J. Rising).) ADP also is expanding its current alliance with SAP though “Project Gloria” to provide a global payroll and other HR functions for multinational clients. (Tr. 4106:1-4107:3 (J. Rising); Ex. D7156R (Z. Brand 5/17/04 Dep.) at 15:23-16:6.) Project Gloria is designed to allow ADP to offer a consistent solution across the world, as well as consolidated reporting. (Ex. D7156R (Z. Brand 5/17/04 Dep.) at 58:12-23.) Fidelity also is expanding its offerings to support Europe, having acquired the HR Access platform to establish a presence outside the United States and to offer global HR and payroll processing solutions to U.S. clients with international operations. (Ex. D5185 (Fidelity Workplace Services Update) at FMR 000980.)

81. Many of the government’s witnesses do not require unlimited scalability. The government’s customer witnesses lack scalability demands. Neiman Marcus, for example, licensed Oracle HRM software, but has not implemented it, and is running its HR on an incumbent architecture. (Tr. 708:6-19 (P. Maxwell).) Similarly, [AIMCO] has licensed PeopleSoft [FMS] software, but still is running on Lawson. (Tr. 1161:15-22 (S. Wesson).) The state of Michigan remains “very satisfied” with the financial management system that it purchased from KPMG in 1993. (Ex. D8103 (D. Johnson 5/28/04 Dep.) at 19:20-20:6; 15:25- 17:3.)

82. Modern software engineering has solved scalability issues. Modern software programming and distributed computer architectures reduce significantly any concerns about software scalability. (Tr. 3353:16-3357:3 (R. Wohl).) Any vendor writing an application designed to scale up to 5,000 users today could just as easily design it to scale up to 25,000 users, as well. (Tr. 3357:4-3363:7 (R. Wohl).)

83. Lawson’s products are scalable. Lawson historically has faced a perception in the marketplace that its product is not scalable. (Tr. 3597:12-15 (J. Coughlan).) Contrary to those perceptions, however, Lawson HRM is running or being implemented throughout the enterprise in several large accounts, including HCA (250,000 employees), Mayo Clinic (40,000 employees); Catholic Care Initiatives (70,000 employees); Polo Ralph Lauren (10,000 employees); Williams-Sonoma (36,000 employees); McDonalds (100,000 employees); Hillsborough County Schools (30,000 employees); State of Arizona (60,000 employees); State of Michigan (65,000 employees); Johnson & Johnson (100,000 employees); Northrop Grumman IT (23,000 employees); Sara Lee (30,000 employees); and the Government of Tanzania (200,000 employees). (Tr. 3599:24-3644:21; 3650:23-3651:6 (J. Coughlan); Ex. D7136 (Lawson Large K-12 Customers).) Similarly, Lawson FMS is operating in highly scaled environments such as Safeway (500,000 postings per day to general ledger and 750,000 per day to Accounts Receivable); Sara Lee (20,000 postings to accounts payable per day, 5,000 invoices per day) and Schlumberger (6,000 users; 1,200 concurrent users). (Tr. 3641:9-20; 3642:3-3643:15; 3645:1-4 (J. Coughlan); Ex. D7002R (Strategic Account Forum) at MI0017914.) The State of Arizona in fact selected Lawson in part because of the product’s scalability. (Ex. D8101 (T. Boncoskey Dep.) at 15:3-10, 20:21-24; 23:13-24; 24:20-25:4.)

84. Other alternatives to Oracle, PeopleSoft, and SAP also are scalable. Outsourcers’ HRM software is scalable. FleetBoston, for example, evaluated Fidelity’s HRM offering and concluded it was sufficiently scalable to handle HR for FleetBoston’s 50,000 employees. (Tr. 3280:14-18 (B. Mearns); Ex. D7020 (HRO: The Wave of the Future?) at 21.) Moreover, Bank of America has selected Fidelity to serve the HR needs of its over 170,000

employees and also its retirees. (Tr. 3145:12-3146:20 (M. Sternklar); Ex. D5189 (Fidelity Press Release Regarding Bank of America.) at FMR 001847.) ADP’s offering also is scalable; ADP has not experienced any limit in terms of the number of employees that its Enterprise HR software can serve. (Tr. 4089:3-6 (J. Rising).) In fact, over 90,000 United Technologies employees currently are served by ADP’s Enterprise HR software solution, and a similar number of Aramark and Tyco employees will be in the near future. (Tr. 4098:16-4100:4 (J. Rising).) Microsoft’s Axapta software also is scalable. According to Microsoft’s internal documents and its marketing brochures, Axapta is highly scalable. For example, the Axapta 3.0 Technical Briefing document indicates that Axapta’s target market includes companies with up to 10,000 employees. (Ex. D5338R, (MBS—Axapta 3.0—Technical Positioning) at MS-OPSUB 000000003068; Ex. D5334 (MBS—Axapta) at 1.) Customers in fact cite scalability as a reason to buy EAS applications from Microsoft. One of the main reasons Esselte chose Axapta, for example, is the product’s scalability, and in particular the ability to deploy it in multiple countries and run it on a small server in each country connected to a large data center. (Ex. D7159 (L. Spund 05/25/04 Dep.) at 86:4-14.) In addition, suppliers that the government characterizes as “mid-market vendors” like JD Edwards also offer scalable software. (D6162 (McKinsey document) at 19; Tr. 804:16-24 (R. Allen); Tr. 826:25-827:23 (R. Allen).)

85. The government witnesses do not all need support for complex pay scales, unions, or mixed labor forces. No vendor offers software that anticipates every collective bargaining agreement, and customers do not view this type of support as a “core” software function. Oracle also does not have a pensions module for its HRM offering. (Tr. 4311:6-12 (L. Ellison).) Customers typically use best of breed or custom products to address collective bargaining business process requirements. (Tr. 1465:17-1466:6 (M. Glover) explaining that Greyhound uses Kronos time keeping software to manage the different pay rates out of the union contracts the company has.)

86. Outsourcers handle union pay. Outsourcers like ADP are able to leverage their payroll expertise to deal with most collective bargaining business process requirements without having to custom-write software code from scratch. (Tr. 4122:15-23 (J. Rising).)

87. Lawson handles complex benefits programs, including union support. Lawson Software’s HRM products include modules for supporting union pay and other complex benefit programs that other vendors do not support. (Tr. 3612:4-20 (J. Coughlan); Ex. D7136 (Lawson Large K-12 Customers).) Customers find Lawson’s ability to handle complex pay or benefits packages attractive; Worker’s Benefit Plans for the Lutheran Church, Missouri Sinod, for example, chose Lawson over PeopleSoft because Lawson was better able to support its 30,000 employees across 6,000 legal entities. (Ex. D8105 (R. Cushman 05/18/04 Dep.) at 9:15-10:24; 12:4-9; 12:16-20; 58:9-59:17; 60:23-61:8.)

88. The government’s customer witnesses actually solve their business process needs many different ways. All of the government’s customer witnesses use software to automate their core HRM and FMS business functions, but all of them use different solutions – that is, different combinations of software packages, outsourcing support, custom programs, business processes, and manual procedures – for achieving those functions. Greyhound, for example, uses Oracle and JD Edwards for FMS, consolidates its financials with Hyperion, and uses HR 1 from Ceridian for its HRM needs (Tr. 1496:25-1499:2 (M. Glover).) PepsiAmericas uses EAS software in combination with point solutions that address specific business process needs. (Tr. 1730:2-15 (K. Johnsen).) Nextel uses about a dozen point solutions today and would continue to do so after installing an ERP suite. (Tr. 1091:2-16 (R. Cichanowicz).) AIMCO uses PeopleSoft Financials and Lawson HR. (Tr. 1129:7-21 (S. Wesson).) Pepsi Americas has 2-3 point solutions on HR side and 2-3 point solutions on FMS side. (Tr. at 1730:8-9 (K. Johnsen).) For its HRM needs, Neiman Marcus currently uses Genesys for core HR and payroll, a home-grown system for commissioned salesperson functions, a 30-year old legacy system for their payroll summary function, an in-house system for HR Workbench functions, and a combination of in-house and outsourcing for benefits. (Tr. at 679:7-11; 721:14-22; 722:8-23 (P. Maxwell).) For its FMS needs, Neiman Marcus relies on a combination of Oracle, Chesapeake Software for

cash management and check reconciliation, and a 20-year old legacy system for merchandising accounts payable. (Tr. at 651:14-16; 707:5-7; 707:11-13 (P. Maxwell).) DaimlerChrysler’s German HR department requested that their ADP PAISY software be retained and not migrated when PeopleSoft was implemented. (Tr. at 1403:22-1404:19 (M. Gorriz).) In evaluating solutions, the government’s customer witnesses did consider more alternatives than Oracle, PeopleSoft and SAP as solutions. Neiman Marcus licensed Oracle’s HR system, and then decided to look beyond Oracle for its benefits and payroll needs, sending an RFP to Ceridian, ADP, Hewitt, Fidelity, and Mellon. (Tr. at 708:17-19; 715:4-14 (P. Maxwell).)

89. Indeed, in the market at large, no large enterprise uses one single software solution. The government witnesses were unable to identify a single example of a large complex enterprise who uses only one software vendor or platform for EAS or ERP. (See Tr. 966:12-21 (P. Keating) (“[E]very time I hear the story [of a single solution on a single instance running on a single location] I go look for it but, I can’t find it.”). As PeopleSoft CEO Craig Conway has noted, “I don’t know of many companies that have committed entirely to one vendor end-to-end. Rather, most businesses of any complexity tend to evolve their administrative enterprises through best-of-breed selections over time.” (Ex. D6078 (PeopleSoft Q&A) at PS-C135747.) Indeed, an internal PeopleSoft Operations Review from 2003 reflects the results of a Meta Group study of HRMS “installations” at 455 “global” companies with over 10,000 employees, and the study reflected that many companies had more than one vendor’s HRMS software installed. (Ex. D6320 (PeopleSoft 2003 HCM Operations Reviews) at PS-C378663. PeopleSoft was installed at 45% of the studied companies, while the next most common installation was “Home Grown/Custom” at 23% of the studied companies. (Ex. D6320 (PeopleSoft 2003 HCM Operations Reviews) at PS-C378663. Oracle was installed at 19% of the companies, while SAP was installed at 16% and ADP was installed at 8%. (Ex. D6320 (PeopleSoft 2003 HCM Operations Reviews) at PS-C378663. The study noted that the total exceeded 100% because some companies responded that more than one vendor was installed. (Ex. D6320 (PeopleSoft 2003 HCM Operations Reviews) at PS-C378663. Indeed, the first five vendors alone — out of 13 different options — amounted to 111%. (Ex. D6320 (PeopleSoft 2003

HCM Operations Reviews) at PS-C378663. Christy Bass of Accenture testified that none of Accenture’s clients are using a single vendor for all of their ERP needs, and that is “unrealistic” for a Global 2000 customer to use a single ERP solution. (Tr. at 1668:5-10 (C. Bass).) These large customers use point-solution software for ERP because the ERP software vendors do not provide the necessary functionality. (Tr. at 1665:5-1669:15) (C. Bass); Tr. at 1730:10-18 (K. Johnsen); Tr. at 1209:7-11 (C. Peters) (Emerson Electric’s plan for FY ‘07 is to grow the Oracle footprint to 57%, maintain the JD Edwards footprint, and still use other ERP systems for the remainder).

90. Business processes and needs other than core HRM and FMS (such as supply chain, inventory management, and customer relationship management) drive guarded buying decisions. Customers’ buying decisions often are driven by a need to automate business processes that are outside the scope of core HRM or FMS functions. This may include non-core HRM or FMS solutions, or more likely other EAS software such as CRM or SCM products. (Tr. 3204:6-11 (K. Harris); Ex. P3185 (D. Dortenzo 5/18/04 Dep.) at 337:7-338:7 (under seal).)

Likewise, the State of North Dakota testified that its purchasing decision was driven in large part by higher education applications, not FMS or HRM. (Tr. 1578:15-1579:7 (C. Wolfe).) CH2M Hill purchased Oracle FMS in 1993 because of Oracle’s superior project management module. (Tr. 205:20-206:2 (R. Bullock).) Vendors also recognize that more strategic products such as “customer facing” applications drive buyer decisions, and allocate discounts and investment dollars accordingly. (Tr. 4239:18-4240:16 (D. Kutnick).) Professor Elzinga acknowledged, for example, that in the Teradyne deal highlighted in his testimony, Oracle offered a 100% discount on HRM licenses in order to sell the SCM applications that Teradyne was interested in. (Tr. 2398:13-2399:20). These areas of the EAS industry are highly competitive. (Tr. 3340:8-18 (S. Catz).) As Ken Harris testified, Nike’s

decision to replace its Oracle FMS system with SAP was driven entirely by Nike’s earlier decision to select SAP for its SCM needs. (Tr. 3201:15-3202:10, 3257:1-8 (K. Harris).) Harris also testified that while Gap was negotiating for a renewal of its database system with Oracle, it received a license to Oracle’s financial system thrown in for free after Harris told Oracle that he probably had the budget to afford a certain amount but that Oracle had to “sweeten this pot” to get Gap to sign the contract (Tr. 3216:18-3217:8 (K. Harris).)

91. Every customer values HRM and FMS software differently. The value placed on a core HRM and FMS solution differs by customer and is relative to the incumbent system, capabilities of other software products outside core HRM and FMS, competing demands from other IT projects or other business priorities, individual risk tolerances, individual feature preferences, emphasis placed on performance versus cost, and many other internal factors particular to each customer. (See Tr. 3201:10-14; 3202:24-3203:9 (K. Harris).) Every customer values the benefits of automating different aspects of their business processes differently, and their priorities may change over time.

92. Customers typically purchase bundles of modules including both core and non-core HRM or FMS, business applications from other pillars and/or one or more stack. Large enterprise customers rarely if ever buy core HRM or FMS modules in isolation. (Tr. at 3461:14-23 (S. Catz).) Professor Hausman testified that FM and HRM software are purchased in bundles with other products, and that it is very rare for a customer to buy core FMS or HRM alone. (Tr. at 3807:21-3808:1 (J. Hausman), see also Tr. at 3813:12-13 (J. Hausman) (“I’ve looked at the data, and the vast majority of customers are buying more than just FMS, more than just HRM.”).) In the PeopleSoft data reviewed by Professor Hausman, the number of different software products purchased in individual deals ranged from 1 to 95, and 85% of the bundles contained at least one product outside of core HRMS and FMS. (Tr. at 3817:16-3818:8 (J. Hausman).) Professor Hausman explained that the product market cannot be limited to just FM and HRM software, because the other products in the bundle have price constraining effects. (Tr. at 3814:3-3816:10(J. Hausman).) Instead, customers almost always purchase a cluster of products such as Oracle’s eBusiness Suite that provide the customer with a “stack” of software

and technology, which may include core HRM or FMS applications, add-on modules, “customer-facing” business applications such as CRM software, and the infrastructure components (application servers and database) on which the applications run. (Tr. at 3461:14-3462:5 (S. Catz); Tr. at 3807:21-3808:1 (J. Hausman).) The particular mix of products almost necessarily will differ from bundle to bundle, and customer to customer. (Id.) Selling software and technology in a cluster is quite common – indeed, the sample of Oracle discount request forms used by the government’s experts for their analyses showed that broader clusters were evident more than 90% of the time. (Tr. at 3813:9-3814:2 (J. Hausman); Tr. at 3815:7-12 (J. Hausman).)

93. Prices are computed on a bundled basis. The evidence indicates the entire bundle of software represented by any deal, whatever its composition, is generally priced as a whole with a so-called “blended discount” and vendors don’t price the individual products separately. (Tr. at 3813:23-3814:2 (J. Hausman).) Oracle’s experts likened this to “salad bar” pricing, in which whatever combination of items a customer selects is weighed and priced as a single product. (Tr. at 4494:22-4495:9 (D. Teece).) This practice makes it hard to compute the price for any particular module, including service, etc. (Tr. at 3814:3-22 (J. Hausman).) Lawson’s John Coughlan testified that Lawson will discount prices where customers purchase several software components at once. (Tr. at 3772:10-3774:23 (J. Coughlan).)

94. Customers’ choices are affected by overall IT and corporate budgets. Customers always have more things to do than the budget or dollars to do them, and so they are constantly engaged in a prioritization process. (Tr. 3204:1-3 (K. Harris); Ex. D6447 (PeopleSoft Executive Briefing on Ford) (describing “Competition” as “Do Nothing.” Ford currently has capped discretionary spending across the company. “If it is not building cars or improving car quality, there is a moratorium on spending.”) Neiman Marcus, for example, evaluated an HRM project, procured an Oracle license, and then determined that they could not afford the implement the software, and so shifted their IT budgets to other alternatives. (Tr. 714:9-24 (P. Maxwell).) To this day, Neiman Marcus is running a combination of custom and incumbent applications, and is exploring outsourcing its HR needs instead of allocating its IT budget to

implementing the Oracle HRM license Neiman Marcus recently acquired. (Tr. 713:11-20 (P. Maxwell).) Greyhound similarly evaluated whether to procure a HRM system, selected PeopleSoft, but did not have the money to acquire and implement the software and instead invested in other aspects of Greyhound’s IT environment. (Tr. 1490:7-22 (M. Glover) FleetBoston decided due to of limited capital not to invest in HR/payroll, but instead to allocate its capital funds to “its core business systems, the revenue generating systems for the bank.” (Tr. 3289:17-3290:11 (B.Mearns).)

95. Customers assess TCO as part of their evaluation of alternatives. Total cost, not license prices, often is the number one criteria for purchasing an enterprise application software suite. (Ex. P0203R (BearingPoint E.U. Questionnaire Response) at 9.) And the purchasing criteria used by BearingPoint’s clients for enterprise application software suites – from most important to least important – regarding: scalability, vendor with proven track record, price, ease of integration into existing EAS, and broad features and deep functionality. (Ex. P0203R (BearingPoint E.U. Questionnaire Response) at 9.)

Similarly, Fidelity’s outsourcing clients always evaluate the HR outsourcing offering on a total cost of ownership basis. (Tr. 3140:13-16 (M. Sternklar).)

96. Customers balance price, features, and many other factors in the evaluation process. Customers do not always choose the best-functioning HRM or FMS product; they trade off between convenience, cost, performance, and many other factors. (Tr. 357:8-17 (R. Bergquist) (acknowledging that customers even in the Fortune 50 can meet their business process automation needs a variety of ways, with high-end HRM or FMS products being just one approach to achieving the same goal); Ex. D7052 (Manulife Case Study); Ex. D7050 (TIAA-CREF Case Study); Tr. at 3619:9-3628:2 (J. Coughlan); see also Ex. P0203R (BearingPoint E.U. Questionnaire Response) at 9 (indicating price is not the most important

criteria for purchasing enterprise application software).) Professor Iansiti explains that “just because — given piece of software can have value to you in certain ways does not mean that you necessarily going to buy it (sic). You have to make an economic decision, and — trade off values and costs.” (Tr. 2105:4-10 (M. Iansiti).); see also Tr. 2108:7-9 (M. Iansiti) (“They can select to go through more traditional management processes, for example, that don’t require software.”)

97. The government’s market definition is not supported by the evidence. The evidence shows customers purchase unique bundles of software, hardware and database products in unique combinations, for their unique purposes and reasons. There is no consistent definition of the “high-function needs” of enterprise customers, or of “high function software.”

98. In contrast, the vendors and analysts do agree that the “technology stack” represents the competitive landscape.

The array of technology customers use to operate their business, including hardware, database, application services and applications, accounts for the variety of services and support that are sold in bundles to customers now and depicts IBM, Microsoft, SAP, and Oracle as being the principal competitors across the stack. (Tr. 4172:4-4174:9 (D. Kutnick); (Tr. 4406:23-4407:15 (D. Teece).) PeopleSoft does not offer multiple stack elements. (See Ex. D6026 (PeopleSoft Email) (Email in which PeopleSoft acknowledges that it is not “on the map,” and “need[s] better integration technology and cannot afford to build it . . . [and that] [t]he problem has not gone away”).)

99. Limiting the alleged market to the United States is arbitrary. The government’s alleged market is limited to the United States. (Complaint 26). The government asserts that it “picked” a market limited to the United States because SAP is the largest vendor and its shares are even larger in a worldwide market, which the government claims would “distort” the competitive picture of how Oracle, PeopleSoft and SAP compete in the United States (Tr. 30:2-10 (C. Scott Opening Statement).) In fact, the government’s limited focus on the United States “distorts” the realities of competition. A reliable unilateral effects analysis requires an accurate assessment of market shares; Professor Elzinga admitted “if you get the shares wrong, you are going to get a misleading inference” that will bias the unilateral effects analysis. (Tr. 2269:2-8 (K. Elzinga), see also Tr. 2270:25-2271:3 (K. Elzinga); Tr. 3798:14-21, 3792:21-24, 3793:3-11 (J. Hausman).) Limiting the alleged market to the United States is arbitrary.

100. SAP has global operations. SAP is headquartered in Walldorf, Germany. SAP has global operations, including major business operations in more than a dozen countries and customers in more than 120 countries around the globe. (Tr. 2805:20-2806:2 (R. Knowles).) SAP has six sales regions worldwide. (Tr. 2808:8-15 (R. Knowles).) SAP America, Inc. is responsible for sales in the United States and Canada. Tr. 2808:16-19 (R. Knowles).) SAP America is in business to sell software solutions created by SAP. (Tr. 2808:8-15, 2806:16-17 (R. Knowles).) In addition to selling software solutions created by SAP, the largest price discounts offered by SAP America – the discounts that the government asserts will be impacted by the proposed acquisition – must be approved by SAP. (Tr. 2836:22-24 (R. Knowles).)

101. Oracle has global operations. Oracle is headquartered in Redwood Shores, California, but it too has extensive global operations. Oracle has over 4,000 General & Administrative employees around the world, with three shared service centers operating 24/7. (Tr. 3485:10-12 (S. Catz).) In addition, Oracle has offices in over 80 countries and sells its products in over 120 countries. (Tr. 3486:16-18 (S. Catz).)

102. PeopleSoft has global operations. PeopleSoft is headquartered in Pleasanton, California. Although its global operations are not as extensive as either SAP or Oracle, it sells its software worldwide. (Ex. D7149 (PeopleSoft Form 10-K for the period ending December 31, 2002) at 7.) “We market and sell our software products in most major world markets through a direct sales organization. The domestic direct sales organization is based in over 25 field sales offices located in major metropolitan areas throughout the United States. International sales activities are conducted out of our offices located near major concentrations of users in Europe, Japan, Asia-Pacific, Latin America and other parts of the world.”)

103. The System Integrators have global operations and focus on providing services to the largest companies in the world. Each of the five largest systems integrators has global operations, including practices which implement the software solutions sold by SAP, Oracle and PeopleSoft. For example, Accenture has operations in 48 countries. (Tr. 1610:17-20 (C. Bass).) IBM Global Services also has global operations. (Tr. 465:14-21; 467:5-11 (N. Thomas).) Accenture, BearingPoint, IBM Global Services and the other systems integrators focus much of their practice on assisting the largest companies in the world – multinational companies in the Global 2000 or Fortune 500. (Tr. 1611:16-14 (C. Bass).); (Tr. 866:14-21; 951:22-952:2 (P. Keating); (Tr. 467:21-468:16 (N. Thomas).); Ex. P0825 (IBM Financial Management Solutions) at IBM 500831; Ex. 5139 (Deloitte Lawson City of Dallas Implementation) at DS000003).)

104. The government itself alleges global support is critical to large enterprise customers. As part of its effort to ignore Lawson, Fidelity, ADP and other companies as potential restraints on post-acquisition pricing, the government asserted that the solutions offered by those companies do not have sufficient “multi-language,” “multi-currency,” or “multi-jurisdictional” capabilities. (E.g., Tr. 3736:3-3738:10 (J. Coughlan).) Indeed, Professor Elzinga claimed that one of the distinguishing characteristics of “high function” software, in his mind, is that it is “capable of being implemented around the globe.” (Tr. 2281:16-2282:13 (K. Elzinga).) As Professor Hausman explained, the explicitly international characteristics of the “high function” software market defined by plaintiffs require a worldwide geographic market. (Tr. 3796:3-3797:1 (J. Hausman).) In short, the government’s product market definition strongly infers that the relevant geographic market is global.

105. The government nonetheless excludes the majority of the Global 2000 from the analysis. Only 751 of the members of the Global 2000 are U.S.-based companies. (Tr. 3797:8-15 (J. Hausman).) Professor Elzinga acknowledged that he attempted to exclude from his calculations all sales to Global 2000 companies that are not based in the United States. (Tr. 2260:16-23 (K. Elzinga).) That means that the majority of the “complex” enterprises to whom SAP, Oracle and PeopleSoft attempt to sell products and services are not included in the relevant market or Professor Elzinga’s market share analysis. (Tr. 2261:4-18 (K. Elzinga).)

106. DaimlerChrysler illustrates that the market is worldwide. For example, one of the government’s emphasized customer witnesses, DaimlerChrysler AG, is headquartered in Stuttgart, Germany.² Under Professor Elzinga’s approach, business applications software purchase decisions made by DaimlerChrysler in the United States would be included in the market, but purchase decisions made in Stuttgart would be excluded. (Tr. 2262:20-2263:3 (K. Elzinga).) This is particularly confounding as the same individuals within DaimlerChrysler are responsible for purchasing decisions and concede they could have made their purchase inside or outside the United States (Tr. 1421:16-1422:2 (M. Gorriz).)³ Similarly, Prof. Elzinga’s methodology would include a sale to Toyota if the purchase decision was made in the United States, but exclude a sale to Toyota if the decision was made in Japan. (Tr. 2261:21-2262:7 (K. Elzinga).) Since there is no justifiable rationale for this arbitrary approach {see infra), Professor Elzinga’s methodology cannot withstand scrutiny.

107. There is no basis for confining the “market” to only U.S.-based companies with “global needs.” Professor Elzinga conceded that the “general decision calculus” for a business considering the purchase of HRM and FMS software “would be similar” for Global 2000 customers located outside the United States. (Tr. 2278:2-13 (K. Elzinga).) Moreover, Professor Elzinga could think of “no systemic difference” between the HRM and FMS buying behavior of members of the Global 2000 that are located in the United States and those that are located in other countries. (Tr. 2279:24-2280:2 (K. Elzinga).) In sum, buyer behavior provides no basis to exclude the majority of the Global 2000 from the market.

108. SAP, Oracle and PeopleSoft design and sell their products into a global market. As Professor Hausman explained, SAP, Oracle and PeopleSoft are designing and selling their products to companies based both inside and outside of the United States. (Tr. 3797:24-3798:13 (J. Hausman).) Oracle and PeopleSoft design software primarily in the United States and ship it throughout the world. SAP primarily designs its software outside of the United States and ships it into the United States. In such circumstances, “there’s no doubt... that the Elzinga-Hogarty test is satisfied for a worldwide market.” (Tr. 3801:3-5 (J. Hausman).) In addition, Professor Elzinga admits that looking at the software code, the Elzinga-Hogarty test would lead to the conclusion that the relevant geographic market is broader than the United States. (Tr. 2280:13-23 (K.Elzinga).)

109. The similarity in pricing inside and outside of the United States confirms that there is a global market. Professor Elzinga admitted that he had not conducted any study of relative prices in the United States versus other regions. (Tr. 2284:12-15 (K. Elzinga).) Professor Elzinga did agree, however, that similar prices in different regions would suggest that there is “one true economic market,” because in a single geographic market “the buying and selling of those products... would collapse those products into one price if they’re homogenous or into an array of prices that just compensate for quality differences or transportation differences between the products.” (Tr. 2283:19-2284:11 (K. Elzinga).) Professor Hausman did conduct a comparison of PeopleSoft’s price discounts in the United States and outside of the United States. That analysis demonstrated that discounts were essentially the same outside of the United States (45.1%) and within the United States (45.2%). (Tr. 3804:25-3806:4

(J. Hausman).) The virtually identical discounts demonstrates that the relevant geographic market is global.

110. Unlike many products, sales of software are not constrained by transportation costs. Professor Elzinga repeatedly conceded that software code is a “textbook example of something that can be shipped around the world... almost instantaneously at almost no cost.” (Tr. 2155:4-8 (K. Elzinga), see also Tr. 2204:18-20, 2155:9-12 (K. Elzinga).) The lack of meaningful transportation costs further confirms that the relevant geographic market is global.

111. Professor Elzinga did not perform any critical loss analysis. In addition to failing to conduct any study of relative prices in different regions of the world (Tr. 2284:12-15 (K. Elzinga)), Professor Elzinga admitted that he did not perform any sort of critical loss analysis. (Tr. 2290:3-2291:16 (K. Elzinga).) In other words, Professor Elzinga performed no analysis of how much U.S. business would have to be lost to overseas business (e.g., additional sales by SAP AG if split into a European company and a US company as the government’s geographic market definition requires) to make a price increase unprofitable.

112. The availability of lower priced programming talent in India and other countries confirms that the market is global. Members of the Global 2000 are increasingly turning to programmers in countries such as India and the Philippines for assistance in supporting EAS. (See, e.g., Tr. 602:19-603:21 (L. Bradley); Tr. 1214:17-24,1217:6-8 (C. Peters).) Those programmers are high-quality and much cheaper than programmers in the United States. (Tr. 603:2-6 (L. Bradley); Tr. 1215:16-22 (C. Peters); see also Tr. 3821:10- 3824:8 (J. Hausman).) That pool of lower-priced, high-quality programmers enables Global 2000 companies to maintain legacy systems. (Tr. 1215:23-1216:10 (C. Peters).) It also provides members of the Global 2000 with options for implementation and maintenance services that are used leverage in negotiations with SAP, Oracle and PeopleSoft. (Tr. 3240:4-22 (K. Harris).)

113. If Professor Elzinga’s “presence” test is applied, then the relevant geographic markets are very local, and not the “U.S.” Professor Elzinga conceded that if the relevant geographic market is where every vendor has a “presence,” then under his methodology, “you could think of looking at this merger on not just a Bay Area versus a Detroit area market,

but on a customer-by-customer basis” in which each customer is its own geographic market. (Tr. 2277:3-6 (K. Elzinga);⁴ see also Tr. 3802:18-3803:5 (J. Hausman).)

114. If Professor Elzinga’s “relationships” test is applied, then each procurement is its own separate market. Professor Elzinga attempted to justify the government’s geographic market definition by asserting that relationships are crucial in the business applications software industry and “you’re buying a relationship, and that relationship is localized” (Tr. 2280:24-2281:10 (K. Elzinga).) Of course, if Professor Elzinga is correct, then each procurement is its own separate market because of the relationship between the local salesperson and the customer. However, the fact that sales people in this industry develop a relationship with their customers does not justify defining a U.S.-only market; in fact, many worldwide markets involve such relationships. (Tr. 3801:12-3802:17, 3803:9-20 (J. Hausman).) Moreover, Professor Elzinga admitted that SAP can and does redirect personnel and other resources around the globe in response to competitive opportunities. (Tr. 2283:2-12 (K. Elzinga).)

115. A reliable unilateral effects analysis requires an accurate assessment of market shares. Professor Elzinga acknowledges that “if you get the shares wrong, you are going to get a misleading inference” that will bias the unilateral effects analysis. (Tr. 2269:2-8 (K. Elzinga), see also Tr. 2270:25-2271:3 (K. Elzinga) (same).)

116. Limiting the geographic market to the U.S. vastly understates SAP’s significance. A U.S.-only geographic market produces incorrect and biased market shares because of SAP’s strength in the worldwide market. (Tr. 3798:14-21, 3792:21-24 (J. Hausman).) Given SAP’s strong market position outside of the United States, Oracle and PeopleSoft’s market shares decrease in a worldwide geographic market. (Tr. 2263:4-25 (K. Elzinga), see also Tr. 2302:20-25 (K. Elzinga) (same); Tr. 3793:3-11 (J. Hausman.)

117. SAP is the leader in the global applications market. SAP’s Richard Knowles testified that SAP has the largest share of the global EAS market with approximately 16%. (Tr. 2834:8-13 (R. Knowles).) He also confirmed that SAP has the largest share of the EAS market in the United States with approximately 7%. (Tr. 2834:17-21 (R. Knowles).)

118. SAP is the leader in the global HRM market. IDC estimates the worldwide HRM market share leaders between 2000 and 2002 as: SAP (11.86%), PeopleSoft (11.29%), ADP (8.56%), followed by Kronos. (Ex. D5815 (July 23, 2003 IDC Report at 6); see also Ex. D5646R (SAP Market Analysis) at 1 (SAP has number one share position in world wide EAS revenue attributable to HR in 2002); Tr. 2867:25-2868:10 (R. Knowles).)

119. SAP is the leader in the global FMS market. SAP’s Final Market Analysis for 2002 presents the relative market shares for financials from 2001 to 2003 based on license revenue for five vendors: SAP, Oracle, PeopleSoft/JD Edwards, Microsoft, and Intuit. (Exhibit D5646R (SAP Final Market Analysis) at 41.) SAP has the largest share of the market. (Id.) Even though Oracle and PeopleSoft/JD Edwards are second and third, their relative market shares combined would not equal SAP’s share. (Id.) PeopleSoft acknowledged that SAP is its number one competitor for HRM. (Exhibit D6011 (PeopleSoft 2003 Campaign Definition: Crush SAP) at 1.)

120. A combined Oracle and PeopleSoft have substantially less than a 35% share in global HRM market. A combined Oracle and PeopleSoft would not have a 35% share of the global HRM market. In 2002, PeopleSoft had a 11.29% market share and Oracle had a 3.04% market share. (Ex. D5815 (IDC Worldwide HRM Applications Forecast Analysis) at 6).

121. A combined Oracle and PeopleSoft have substantially less than a 35% share in global FMS markets. In a 2003 analyst briefing, IBM stated that “Finance Applications service offers tied to the Big 3 [SAP, Oracle, and PeopleSoft] represent about 48% of FMS sales globally.” (Ex. P0825 (IBM Financial Management Solutions) at IBM 500845.) Within the 48% market share belonging to the “Big 3,” IBM represents that “SAP FICO sales represent almost 40% of Finance Apps sales followed by Oracle at about 35% and PeopleSoft

around 25%.” Id. Accordingly, even a combined Oracle and PeopleSoft would not have a 35% share of the global FMS market.

122. Market shares for HRM and FMS are further diluted by the “do nothing” option, incumbent systems, and outsourcing. For example, Christy Bass testified that Accenture clients outsource their entire HR departments to Accenture, and cited DuPont, BellSouth and Best Buy as examples. (Tr. 1646:25-1648:19 (C. Bass); Ex. P3198R (C. Bass 4/21/04 Dep.) at 150:5-15, 151:17-19, 154:1-12.). Ms. Bass also confirmed that the outsourcing market is “rapidly growing.” (Tr. 1655:10-13 (C. Bass); Ex. P3198R (C. Bass 4/21/04 Dep.) at 154:13-22.). In addition, Ms. Bass’s testimony also confirms that about 20% to 30% of all software selection processes end up with the customer deciding to “do nothing.” (Tr. 1673:14-21 (C. Bass); Ex. P3198R (C. Bass 4/21/04 Dep.) at 166:7-9.). Vendor market share calculations are “overstated” where a customer chooses to maintain its current ERP system unless vendor percentage shares are proportioned over the maintenance period. (Tr. 2712:9-2713:5 (T. Cambell); see also, e.g., Tr. 2513:22-2514:9 (P. McAfee).)

123. Professor Elzinga’s market share calculations, which fail to account for “do nothing” and outsourcing options, are not reliable. A customer’s option of relying on its incumbent systems and not purchasing HRM or FMS must be accounted for in a proper market share calculation. As Dean Campbell testified, “It is wrong to assign zero” to the market share calculations for the do nothing option in such an industry (Tr. 2759:11-14.) In an industry where customers consider doing nothing even down into the final stages of an evaluation process, vendors’ market shares are “overstated” unless vendor percentage shares are proportioned over the maintenance period. (Tr. 2712:9-2713:5 (T. Campbell).) In calculating market shares, however, Professor Elzinga did not count outsourcing, best of breed solutions, incumbent systems, or other alternatives for customers of HRM and FMS products. (See, e.g., Tr. 2320:25-2321:8 (K. Elzinga); Tr. 2325:1-13 (K. Elzinga).) It may be impossible to tell what the correct market shares may be in this case, but certainly a substantial amount of commerce goes to these other options. That this overstates the included vendors’ market shares is so plain that even Professor McAfee was compelled to adjust Professor Elzinga’s market share

calculations and make some accounting for incumbent systems in his merger simulation.. (Tr. 2759:19-2760:16 (T. Campbell).)

124. SAP is the largest vendor of EAS in the world, and its U.S. shares are increasing. SAP essentially coined the phrase “ERP software.” (Tr. 2810:1-3 (R. Knowles).) Today, SAP is “the world’s leading provider of business software solutions.” (Tr. 2805:15-19 (R. Knowles) citing SAP’s website.) Its market share has continued to increase even as the EAS market has been declining over recent years. (Tr. 2874:8-11; Ex. D5646R (SAP Market Analysis) at 1.) In particular, SAP’s market share in the United States has continued to increase. (Tr. 2874:21-2875:4 (R. Knowles).) More specifically, of the Global 500 companies, SAP’s customers include 86% of those in Consumer Products, 67% in Retail, 87% in Aerospace & Defense, 87% in Automotive, 69% in Engineering & Construction, 97% in High Technology, 100% in Chemicals, 86% in Milling, 100% in Mining, 93% in Oil & Gas, 100% in Pharmaceuticals, 54% in Banking, 48% of Financial Service Providers, 37% in Insurance, 63% in Media, 76% of Service Providers, 81% in Telecommunications, and 78% in Utilities. (D5000 (SAP Customers Within Global 500); Tr. 2825:1-2828:6 (R. Knowles).)

125. SAP has core HRM and FMS functionality. SAP offers core FMS and HRM functionality. (Ex. P3041 (B. Patel Dep.) at 186:6-14; Ex. P3062 (S. Elliot 05/20/04 Dep.) at 33:9-25.) SAP’s core FMS product is called mySAP Financials. (Tr. 2812:4-13 (R. Knowles).) SAP’s core HRM product is called mySAP Human Capital Management or mySAP HR. (Tr. 2812:14-23 (R. Knowles).) SAP offers a bundled product called SAP ERP, which includes financials and human resources functionality as well other additional functionality. (Tr. 2810:8-17 (R. Knowles).) SAP supports global core HRM and FMS demands, and SAP’s product is highly scalable, as well. Market analysts have stated that “SAP has long had the reputation for being the most scalable ERP system on the market.” (Ex. P3065 (Accenture AOL ERP Evaluation) at ACCEN-00001040; Ex. P3065 (Accenture AOL ERP Evaluation) at ACCEN-00001028, 1039.)

126. SAP is a formidable competitive constraint on Oracle, now and for the future. A formidable market force like SAP can discipline Oracle’s pricing single-handedly. (See Tr. 2852:12-16 (R. Knowles).; Ex. D5649R (SAP Discount Approval for Kellog Brown & Root showing a “horse race” between Oracle and SAP); Tr. 2853:3-2854:22 (R. Knowles).). Should Oracle be permitted to acquire PeopleSoft, some analysts predict a coming “battle royale” between a post-merger Oracle and SAP over accounts where both PeopleSoft and SAP are present today. (Tr. 2860:25-2861:6 (R. Knowles).) Plaintiffs’ expert Kenneth Elzinga conceded that he could not identify any objective criterion to support plaintiffs’ contention that SAP would be less appealing than Oracle and PeopleSoft for a particular customer, and Professor Elzinga could not identify any vertical industries in which SAP is not a good substitute for Oracle and PeopleSoft. (Tr. 2367:9-2368:6 (K. Elzinga).)

127. The evidence shows SAP’s competitiveness in core HRM and FMS. At trial, third parties confirmed SAP’s commercial appeal and resulting power. If Cox Communications had not received the price that it wanted on financial management software from Oracle or PeopleSoft, it could have considered other options, including SAP. (Tr. 186:22-187:24 (S. Hatfield) (5/11/04 Dep. at 183:8-25).) Kerr-McGee’s Scott Elliot views SAP as a competitor to Oracle and PeopleSoft. (Ex. P3062 (S. Elliot 05/20/04 Dep.) at 144:23-25.) PepsiCo., which has been an Oracle client since 1997, recently licensed the mySAP Business Suite. (Tr. 2864:6-21 (R. Knowles).) And thirty-five percent of all IBM ERP implementations since 2001 have been SAP implementations. (Ex. P0900 (IBM Deal Summaries) at 1-4.) Verizon considers SAP HRMS and FMS offerings to be nearly equivalent to its counterpart PeopleSoft products, and was even willing to license SAP products in order to obtain leverage against Oracle “just in case” the merger went through. (Tr. 623:4 – 624:9 (L. Bradley).)

128. PeopleSoft recognizes SAP’s dominance as well. PeopleSoft acknowledges that SAP is its number one competitor, including for HRM software in the United States. (Ex. D6011 (PeopleSoft 2003 Campaign Definition: Crush SAP) at 1).) In 2003, PeopleSoft acknowledged that “SAP is our primary competitor for HCM, FMS and SCM.” (Ex. D6011 (PeopleSoft 2003 “Crush SAP”) at 1.) Chuck Teller of PeopleSoft, in discussing SAP’s composite applications strategy, warned that “SAP is going to eat our lunch. We are late to the game.” (Ex. D7009 (PeopleSoft Teller email).) PeopleSoft 2003 Operations Reviews for HCM, stating on slide 2: “While HR is still a ‘local’ game, the battle for worldwide leadership is now clearly between SAP and PeopleSoft.”

129. SAP is Oracle’s closest competitor and PeopleSoft’s closest competitor. SAP– not PeopleSoft – generally is considered Oracle’s closest substitute. Industry consultants recognize SAP is a clear substitute for either Oracle or PeopleSoft. In the course of evaluating EAS providers for Nextel, for example, five different consultants treated SAP or Oracle as the best vendor for Nextel. (Tr. 1096:6-1097:11 (R. Cichanowicz) (authenticating P 4075 R at 23).) Likewise, PeopleSoft recognizes SAP as its primary competitor. (Ex. D6011 (PeopleSoft 2003 Campaign Definition: Crush SAP) at 1.)

130. Oracle discounts are just as deep when SAP is the competition as when PeopleSoft is the competition. Professor Hausman’s regressions demonstrate that the percentage discount offered by Oracle does not depend upon whether it is competing against PeopleSoft or against SAP. (Tr. 3887:16-25 (J. Hausman) (6-28-04) (“So, what this says is, that PeopleSoft on average is not a closer competitor than SAP, because, in fact, they have exactly the same effect, the same estimated effect on Oracle’s average discount.”) This makes sense, as Oracle’s President Safra Catz testified that she does not consider the identity of a customer in determining the discount. (Tr. 3464:4-14 (S. Catz).) Professor McAfee’s regressions show that the presence of SAP in a deal is correlated with the same or higher discounts from Oracle than the presence of PeopleSoft is. (Ex. P4023A, Tr. 2584:10-15 (P. McAfee) (“This regression does predict that the effect of SAP on the discount is larger than the effect of PeopleSoft.”); Tr. 2576:12-13 (P. McAfee).) In fact, Professor Hausman demonstrated that, after correcting a joint endogeneity problem in one of Professor McAfee’s regressions, the effect of PeopleSoft on

Oracle’s discounts almost disappears but the effect of SAP remains robust. (Tr. 3894:3-3896:13 (J. Hausman).) Professor Hausman explained that competition from SAP produces discounts from Oracle that are just as high as those provoked by PeopleSoft, and the customer “can’t get both discounts.” (Tr. 3887:16-3888:11 (J. Hausman).)

131. Government’s experts do not show localized competition between Oracle and PeopleSoft. Dean Tom Campbell concluded that adverse “unilateral effects” are unlikely from this merger, because this industry is characterized by bilateral negotiations with powerful buyers, not “groups of small consumers surrounding a particular node in product or geographic space for which the unilateral effects doctrine was developed.” (Tr. 2702:20-2703:7, 2721:3-8, 2721:20-2722:3 (T. Campbell).) To the contrary, none of plaintiffs’ economic experts made a contrary showing – i.e., that the products of Oracle and PeopleSoft are particularly close substitutes, or that PeopleSoft constrains Oracle’s pricing discretion more than other competitive alternatives do. Professor Elzinga conceded that he did not conduct any demand or cross-price elasticity studies on the products in this market (Tr. 2291:17-2292:25 (K. Elzinga).), nor performed any statistical regressions. (Tr. 2293:1-6 (K. Elzinga).) And his market share calculations afford no basis for inferring that Oracle and PeopleSoft are uniquely close substitutes, or that PeopleSoft forces Oracle to compete more aggressively than do other alternatives.

132. Professor Eliznga’s analysis of discount approval forms (“DAFs”), which intended to show Oracle and PeopleSoft are best substitutes, is unreliable. Professor Elzinga testified that he reviewed approximately 210 Oracle discount request forms. (Tr. at 2316:18 - 2317:11 (K. Elzinga).) More than 188 boxes of these forms were produced to DOJ. (4/16/04 Hearing Transcript at 21; “[Oracle] rented the entire cargo hold of the last redeye on United Airlines...to get the 188 boxes of documents to [DOJ].”) The selection was too narrow to be representative. Moreover, Professor Elzinga concluded based on tabulations of discount approval forms that “do nothing” was not a sufficiently common occurrence for it to be allocated

market share in an analysis of concentration. (Tr. at 2195:20—2197:11 (K. Elzinga). Professor Elzinga, however, ignored discount request forms where no specific competitor was identified (Tr. at 2318:19—2319:5 (K.Elzinga)), which necessarily understates customer’s use of their “do nothing” options. Moreover, Professor Elzinga testified he did not look at any PeopleSoft or SAP discount approval forms, and as a result, he has no idea whether the occasions of direct competition between Oracle and PeopleSoft that he observed constitute a large or a tiny proportion of any overall “high function” market. (Tr. 2351:6-2352:3 (K. Elzinga).) Moreover, Professor Elzinga’s limited review to transactions of $500,000 or more in net license revenue involving U.S. companies, but made no attempt to employ Dr. Iansiti’s “high-function” definition (Tr. at 2314:16—2315:14 (K. Elzinga), nor any attempt to screen out many customers that obviously do not have “high-function” needs, including the YMCA of Greater Hartford, Connecticut. (Tr. 2323:1-16 (K. Elzinga).)

133. Professor McAfee’s results are flawed and do not show PeopleSoft and Oracle are the closest competitors to one another. There are three fundamental flaws with Professor’s McAfee’s results, (a) He includes in his data only eBusiness Suite transactions. As discussed above, eBusiness Suite deals necessarily involve a bundle of software, for which he made no effort to account for or isolate the HRM or FMS dimension of the transaction or apply a “high function test.” (b) Professor McAfee’s summary statistics demonstrate only that Oracle competes with PeopleSoft in “larger transactions about half the time.” (Tr. 2495:4-13 (P. McAfee).) In the set of deals over $5 million in license revenue, SAP appeared three times and PeopleSoft appeared only once. (Tr. 2568:22-2569:1 (P. McAfee).) In the set of deals between $1 million and $5 million in net license revenue, PeopleSoft appeared 28 times but SAP appeared 20 times. (Tr. 2569:2-8 (P. McAfee).) (c) Professor McAfee conceded that he made a mistake in tabulating how frequently “other” firms appeared in his summary statistics tabulation, because he accidentally counted “others” only when neither PeopleSoft nor SAP were present. (Tr. 2569:13-25 (P. McAfee).) Professor McAfee has not recalculated the summary statistics to correct that error, but testified that he would not be surprised to learn that “other” firms appear in more than three-quarters of the deals. (Tr. 2570:12-20 (P. McAfee).). As noted above,

Professor McAfee’s regressions do not demonstrate that Oracle and PeopleSoft are particularly close competitors; to the contrary, they suggest that Oracle discounts more heavily when it is competing with SAP than when it is competing with PeopleSoft. Professor McAfee discussed several selected “case study” vignettes of competition between Oracle and PeopleSoft, but did not attempt to ensure that they were a significant or representative sample of overall competition in the market. (Tr. 2541:3-9 (P. McAfee).) In fact, he conceded that the discount forms he discussed were not “typical or representative of what [he] found overall,” and that he carefully selected examples that would show clear competition between Oracle and PeopleSoft. (Tr. 2548:11-2549:2 (P. McAfee).) And Professor McAfee’s analysis of Oracle’s discount approval forms does not demonstrate that Oracle and PeopleSoft are particularly close or “localized” competitors. Professor McAfee conceded that Oracle necessarily encounters PeopleSoft more often than other competitors, simply because PeopleSoft is bigger and offers more products in more pillars and verticals. (Tr. 2490:16-2491:14 (P. McAfee).) That does not mean that PeopleSoft’s products are uniquely similar to Oracle’s. Professor McAfee also made no effort to ensure that the discount approval forms he reviewed were exclusively or even primarily focused on the products at issue in this case, and in fact conceded that they generally involved other products. (Tr. 2493:19-23, 2491:15-2493:18, 2550:5-13, 2555:25-2556:3 (P. McAfee).) He also made no effort to ensure that those forms dealt with large, complex, or “high function” customers as opposed to small or “mid-market” customers. (Tr. 2549:12-19 (P. McAfee).)

134. The government’s theory of competitive harm fails if following the merger other vendors can enter or reposition to fill any gaps in competition, which they can. Professor Elzinga acknowledged that unilateral effects are unlikely in differentiated products markets if rival sellers could replace any localized competition lost as a result of the merger by repositioning their product lines. (Tr. 2368:15-21 (K. Elzinga).)

135. Nothing in the record suggests SAP would be unable to reposition with particular functionality if and as desired. SAP’s Richard Knowles testified that there is

no vertical or market segment where SAP would be unable to compete effectively with a post-merger Oracle for core FMS or core HR sales. (Tr. 2964:16-21 (R. Knowles).) Professor Elzinga conceded that he is unaware of any reason, barrier, or ‘resource constraint’ that would prevent SAP from repositioning to replace any localized competition lost as a result of this merger, “[o]ther than the lack of skill and management to do so....” (Tr. 2368:22-2370:8 (K. Elzinga).) In fact, SAP has ample resources, with some 8,000 software developers (Tr. 2871:17-19 (R. Knowles)) and a research and development budget exceeding $1 billion per year (Tr. 2871:13-16 (R. Knowles) (SAP spent $1.3 billion on research and development in 2003 and $900 million in 2002).)

136. Even in verticals where it is alleged to be relatively less able to complete, SAP has significant established customers. For example, Deloitte & Touche, PricewaterhouseCoopers, and Accenture are SAP customers in the professional services sector. (Tr. 2829:2-11 (R. Knowles).) SAP has engineering firm customers including Flour Daniel and Halliburton. (Tr. 2829:12-25 (R. Knowles).) SAP counts 11% of U.S. large enterprise utilities as customers, including PG&E. (Tr. 2830:3-13 (R. Knowles).) SAP’s telecommunications customers in the U.S. include SBC, T-Mobile, and MCI. (Tr. 2830:14-19 (R. Knowles).) AOL is an SAP customer in media. (Tr. 2830:22-25 (R. Knowles).) In retail, The Home Depo, Starbucks, Barnes & Noble, and Nike are all SAP customers. (Tr. 2831:12-19 (R. Knowles).) SAP currently competes for large complex federal procurements and counts among its customers U.S. Customs, Navy, Army Logistics, IRS, Department of Interior, and NASA at the federal level. (Tr. 2881:14-2882:3 (R. Knowles).) At the state level, SAP’s state-wide implementations include Pennsylvania, Florida, Arkansas, and Washington. (Tr. 2883:8-16 (R. Knowles).)

137. SAP can and does target verticals where it has a below-average share of the market. SAP’s efforts to penetrate under-exploited vertical markets does not require improvements to its “core” FMS and HRMS offerings. (Tr. 2963:25-2964:15 (R. Knowles) (discussing focus on financial services vertical).) However, SAP can and does develop industry specific functionality to enter and expand specific verticals. (Tr. 2963:25-2964:15 (R. Knowles).) For example, SAP has also developed a new billing solution specifically targeting

insurance companies. (Tr. 2880:4-20 (R. Knowles); Ex. D6289 (PeopleSoft E-mail re: “SAP – FSI”) at 1.) In the public sector, North Dakota initially viewed SAP as no competitive threat to a post-merger Oracle, but sees SAP as a more credible threat now that SAP has developed higher education functionality. (Tr. 1587:12-18 (C. Wolfe); see also Tr. 2882:18-2883:7 (R. Knowles) (SAP has sold its recently-developed student administration system to Higher Education customers including Duke University, Johns Hopkins, and the University of Kentucky).)

138. SAP specifically has targeted the North American financial services market. SAP has recently targeted the North American financial services sector by (1) entering into an alliance with Accenture to develop additional industry specific solutions, (2) developing a national financial services practice with dedicated sales resources, and (3) recruiting new leadership for the team. (Tr. 2875:5-2876:16 (R. Knowles); Ex. D6289 (PeopleSoft E-mail re: “SAP – FSI”).) As a result of SAP’s efforts, PeopleSoft sees SAP as posing “a real threat” in the financial services space and recognizes SAP as “an incredibly strong new competitor” in that space. (Ex. D6289 (PeopleSoft E-mail re: “SAP – FSI”) at 1; see also Ex. D6014 at 1 (PeopleSoft E-mail stating “SAP is coming into two of our current strongholds – Higher Ed and Banking.”) at 1.) SAP now has customers among North American banks, insurance companies, and capital markets firms, including Allstate Insurance, American Express, Visa, Bank One, Farmer’s Insurance, Employers Reinsurance Company, MassMutual, New York Life, and the New York Stock Exchange. (Tr. 2878:10-2879:6 (R. Knowles).)

139. Acquisitions provide another means for SAP to enter and exploit emerging markets. IBM’s internal assessment of Oracle’s acquisition of PeopleSoft describes the applications market as a “predators’ ball” and identify SAP, along with Microsoft, Accenture, and Hewlett Packard, as potential acquirers. (Ex. D5240R (IBM Analysis: Implications of PeopleSoft/Oracle Merger) at IBM 500367 (potential acquisition targets include Intuit, McKesson, Fiserv, Sungard Data Systems, Dassault Systems, PeopleSoft, Siebel, Synopsys, Amdocs, Cadence Design, Ceridian, Autodesk, Reynolds & Reynolds, Cerner, Hyperion, Ariba, Kronos, Mentor Graphics, Lawson, Peregrine, MSC Software, Freemarkets, and the Manugistics Group); see also (Ex. 3034 (D. Morea 5/17/04 Dep.) at 46:2-19 (SAP & AMS discussed a

possible merger in June, 2000); Tr. 700:1-25 (P. Maxwell) (SAP could reposition through acquisition to improve its offerings to the retail vertical).)

140. Lawson is already in the large enterprise market, and its vertical focus allows it to target any area where it sees significant revenue potential. Lawson’s verticals-focus business strategy is ideally suited to enable Lawson to reposition and constrain a merged Oracle-PeopleSoft’s unilateral conduct. As Professor Hausman testified, Lawson does not plan to fully replace PeopleSoft after the transaction, nor does it need to; what is important is that Lawson can and will bring to bear its strategy of vertical strength to compete for large enterprise customers. (Tr. 3839:1 - 3840:13 (J. Hausman).) (“[Lawson doesn’t] have to be everywhere. They don’t have to replace PeopleSoft. They just have to have the ability if price goes up they can get into that segment and compete and bring the prices back down, and of course, they will have every economic incentive to do so, and they’ve demonstrated that they can do it in the verticals they’ve gone after.”); Tr. 3654:16-3655:1 (J. Coughlan).) Lawson also has significant support from system integrators. BearingPoint already has 35-40 individuals dedicated to Lawson (Tr. 881:13-14 (P. Keating)), Deloitte & Touche has a dedicated Lawson team of about 70 individuals; and IBM is Lawson’s largest implementor. (Tr. 515:10-16 (N. Thomas).) Lawson has a strategic alliance with Deloitte, which has already completed many large, complex Lawson implementations, including implementations for has strong relationships with Cap Gemini.

Lawson has been very successful in the verticals it has focused on, and nothing would prevent it from repositioning if an opportunity arose. (Tr. 3841:3-13 (J. Hausman), see also Tr. 3911:9-15 (J.Hausman) (“If the price goes up in a vertical, Lawson is very likely to move in.”).) Professor Hausman explained that Lawson does not have to be a broad-based competitor like PeopleSoft in order to constrain Oracle’s pricing; they can focus on particular verticals if prices increase there. (Tr. 3839:4-3840:21, 3841:14-3842:2 (J. Hausman).)

141. Microsoft is capable of expanding into any part of the market it chooses. BearingPoint regards Microsoft, IBM, and Siebel as potential entrants to “a specific market for supplying EAS (e.g., HR, FM) to large companies (e.g., companies with more than 10,000 employees or turnover in excess of 1 billion euros)” within the next two years. (Ex. P0203R (BearingPoint E.U. Questionnaire Response) at 14-15.) In BearingPoint’s opinion, the only conditions that prevent market entry by these firms are “self choice.” (Id.) All three are presently active in “all markets” – “in some cases, small but growing market shares.” (Id.) IBM’s internal analysis of the Oracle acquisition makes it clear that IBM sees Microsoft competing head-to-head with Oracle and SAP in the evolving applications market. (Ex. D5242RA (IBM Analysis: Oracle Acquisition of PS) at IBM 500666, 500668.) PeopleSoft’s

Richard Allen testified that it would simply take time and a “boat load of money” for Microsoft to enter the “up-market.” (Tr. 827:24-828:7 (R. Allen).) And he acknowledged that Microsoft has the required “boat load of money.” (Tr. 828:8-10 (R. Allen).) Indeed, the day after Oracle’s tender offer for PeopleSoft was announced, Microsoft began considering a potential acquisition of SAP. . (Exs. D5324R (Gates-Ballmer email entitled “PeopleSoft and SAP and us??”), 5326R (Bates email to Ballmer, Gates and others entitled “Mensa/Sagittarius Combination”); Tr. 3069:4-3072:22 (D.Burgum).) A company that seriously considered spending many tens of billions of dollars to acquire the leading EAS supplier in the world cannot be discounted as a potential entrant.

142. IBM has the resources to compete in the software applications market. IBM executive Steve Mills recently stated that IBM is not presently in the EAS applications business, but he made it clear that IBM’s software business is “one of the major... building blocks of the company going forward.” He added: “The IBM company has been producing software since there was such a thing. Even back when our business was entirely mainframe centric there was... many, many thousands of IBM developers that were... the principal architects and inventors [of] a good deal of base technology that’s used across the industry today. So the skills have always been there.” Mills noted that market analysts and competitors did not initially take IBM’s software business seriously, but “we’ve been relentless at this” and today “more than 45,000 people work in IBM software,... it’s about $15 billion of revenue,... we have more than 10,000 dedicated sales people, [there are nearly 1,000 people in telesales, 1,000 people in marketing, [and] literally tens of thousand of ISV and invaluable partnerships.” (Ex. D5252 (IBM Lehman Brothers Presentation) at 4, 6.) In a prime example of how quickly IBM can enter and dominate a software market, IBM went from a market entrant to a market leader in the Web portal business with its WebSphere product in a mere eighteen months – and, in that instance, IBM wrote the software code from scratch. (Ex. D5252 (IBM Lehman Brothers Presentation) at 23-24.)

143. IBM considered buying PeopleSoft itself. After Oracle announced that it intended to acquire PeopleSoft, IBM’s management considered reentering the applications market. IBM’s strategic internal documents raise the question directly: “do we need to purchase PeopleSoft and reenter the application business to stay in the market or be locked out by SAP, Oracle, and [Microsoft].” (Ex. D5242RA (IBM Analysis: Oracle Acquisition of PS) at IBM 500672.)

144. Though AMS has been most successful in the public sector, it is not foreclosed from commercial space. Indeed PeopleSoft recognizes AMS as a potential threat outside the public sector. In a July, 17, 2003 e-mail attaching a competitive profile of AMS, Scott Rollins of PeopleSoft states that salespeople in PeopleSoft’s E&G group “knew of AMS and competed against them in pretty much every deal.... [I]t is important to note they have gained some traction in the commercial space (especially in financial services and telecom) and should be taken seriously anytime you run into them in the field.” (Ex. D5935 (PeopleSoft e-mail to E&G Management Team).)

145. JD Edwards sold core HRM and FMS to large customers in 2001, and in 2003. As early as May of 2001, JD Edwards had an installed base of at least 260 “large enterprise” customers (D6162 (Developing a Product Market Strategy for JD Edwards, dated May 3, 2001, at 18).), including Dole Foods Company with $5 billion in revenue, LVMH with $12 billion in revenue, and Shell International Petroleum Company with over $75 billion in revenue. (P4875 (JD Edwards Product Marketing Strategy: Making it Actionable, dated May 31, 2001, at 10).) Prior to its merger with PeopleSoft, JD Edwards sold core HRM and FMS software to numerous large and complex enterprises. In 2003, in response to CID No. 22504 in connection with the proposed Oracle/PeopleSoft merger, JD Edwards identified 504 customers meeting the government’s definition of “large and/or complex” organizations. (D7163 (Fed. R. Evid. 1006 Summary of Exhibit D6158/JD Edwards Response to CID 22504, Answer 4(a)(2)/Exhibit 4a2: Products Sold to Large and/or Complex Organizations).)

146. External factors and management issues, not technology limitations limited JD Edwards’ ability to execute on its strategy to enter the up-market. In 2001, JD Edwards decided to focus on certain “core” verticals in the “mid-market,” which JD Edwards defined to include companies with between $500 million and $3 billion in revenue. (P4875 (JD Edwards Product Market Strategy: Making it Actionable) at JDEC-01-030441.) That decision, however, was not driven by any inability of JD Edwards to meet the needs of large enterprise customers. Instead, the decision was driven by the fact that the large enterprise market was “saturated,” and therefore the “mid-market” presented significantly more “new potential ERP customers” and greater revenue opportunities for JD Edwards than the up-market. (D6163 (Developing a Product Market Strategy for JD Edwards/May 7, 2001) at JDEC-04-002691, JDEC-04-002695, JDEC-04-002697, JDEC-04-002699, JDEC-04-002710).) Indeed, the McKinsey and Company analysis used by JD Edwards in formulating its mid-market strategy specifically noted that enterprise customers within JD Edwards’ “core” verticals needed “similar functionality” as mid-market customers, and therefore JD Edwards’ products would have to address the same feature/functionality “gaps” for either market. (D6163 (Developing a Product Market Strategy for JD Edwards/May 7, 2001) at JDEC-04-002711).) McKinsey also concluded that JD Edwards’ technology “appears to be scalable,” although JD Edwards had not yet convinced large enterprise customers. (D6163 (Developing a Product Market Strategy for JD Edwards/May 7, 2001) at JDEC-04-002711).)

147. Negotiating with two vendors is optimal for buyers. If it were consistently preferable to have three vendors rather than two, then buyers would never “voluntarily restrict the number of potential vendors” by down-selecting to two or even one — as customers regularly do in this industry. (Tr. 2723:11-2724:3 (T. Campbell); see also Ex. 3255R (O. Ayala 5/18/04 Dep.) at 136:13-19 (testifying that one cannot generalize that a customer will pay more money by down-selecting to two than if they down selected to three.) In this case, the witnesses demonstrate the adequacy of two vendors competing: core HRM and FMS buyers

usually “whittle” down the competitors to two (sometimes three) finalists and then use them as leverage against each other in negotiations. (Tr. 2848:21-2849:12 (R. Knowles).) “Generally what happens is that a prospect says to two vendors ‘We think either of you can do it.’” Then they may say ‘“Give me a price proposal and a contract and let’s dialogue about legal terms and conditions.”’ (Ex. P3179 (Ciandrini 1/16/04 Dep.) at 147:16-22.) Evaluations are costly and two finalists yields a sufficiently competitive outcome; “I think it is fair to say that we didn’t anticipate getting a sufficiently better outcome with SAP [as a third vendor] to overcome the cost of adding them to the evaluation.” (Tr. 184:5-185:7 (S. Hatfield)) IBM’s Nancy Thomas testified that software purchases “very quickly” go down to one or two vendors for selection of software packages. (Ex. D5238 (IBM Package Selection Services) at 4; Tr. 556:5-558:2 (N.Thomas).) Experienced EAS buyer Ken Harris testified that he prefers to narrow down to two solutions to examine in further detail. “To get to an optimal solution, in my opinion, as long as you have at least two viable competing options, you can get there, and so it’s really a matter of, by and large, a matter of efficiency why you would narrow it down.” (Tr. 3215:12-15 (K. Harris).) Pitting two vendors against each other allows for competitive pricing while maximizing efficiency. (Tr. 3254:5-10, 3252:21-24 (K. Harris).) Professor Campbell explains why: “So the answer is, there are many characteristics in bilateral negotiation where the product is different, and that’s the key here. It’s not that less is better than more. It is that you can get a better combination of functionalities where you have a negotiation carried out between two parties that you couldn’t if it was between one party and a plethora or substantially larger on the other side.” (Tr. 2730:6-12 (T. Campbell).) Professor Hausman also explained that, if more competitors produced higher discounts, we would expect to see customers inviting as many vendors to compete for their business as possible. (Tr. 3897:21-3898:10 (J. Hausman).) But in approximately 75-80% of the buying situations Professor Hausman observed, customers voluntarily down-selected to 2 or fewer vendors. (Tr. 3791:8-21, 3878:12-3880:24 (J. Hausman).) In the Oracle discount approval forms selected by Professor McAfee, Oracle, SAP and PeopleSoft were all competing for the customer’s business only 1.8% of the time. (Tr. 3883:5-16 (J. Hausman).) Professor Hausman’s statistical regressions demonstrated that having

three vendors rather than two competing for the customer’s business does not produce higher discounts. (Tr. 3791:14-21, 3853:6-16, 3888:9-3889:7 (J. Hausman).)

148. The evidence shows large and complex customers down-select to two finalists for negotiations (and sometimes just one). Third-party customers’ experiences (and trial testimony) confirm the norm of selecting down to two vendors:

149. The two finalists are not always software vendors. The two finalists can include a credible “do nothing” option, outsourcing, and best of breed alternatives. DaimlerChrysler narrowed its choices for its United States payroll system down to two choices: its legacy payroll system and PeopleSoft. (Tr. 1415:9-18 (M. Gorriz)) DaimlerChrysler narrowed its choices for its United States HRM for blue collar workers down to two choices: its homegrown legacy system and PeopleSoft. (Tr. 1417:17-1418:3(M. Gorriz)) Had WBP not been able to obtain a satisfactory price on core HRM and FMS software, it could have delayed purchasing any new system, and considered upgrading its existing system. (Ex. D8105 (R. Cushman 5/18/04 Dep.) at 73:18-25.) Emerson Electric’s Charles Peters explained that “[n]o deal” with Oracle would have meant “in-house integration of best-of breed products and... continual maintenance of legacy systems.” (Tr. 1245:11-19 (C. Peters).) Similarly, CH2M Hill’s negotiating strategy with Oracle was to say that if they couldn’t strike a reasonable deal, CH2M Hill was prepared to stay with what it had in place. (Tr. 233:24-234:2 (R. Bullock); Tr. 1467:16-18,1490:7-15, 1490:23 – 1491:6 (M. Glover) (Greyhound on two occasions decided not to purchase new core HRM software because they could not obtain a satisfactory price; on both occasions they decided to upgrade some systems and otherwise use their legacy systems.)

150. These are negotiations, not “bids.” Procurements in this industry are not conducted as bidding situations or auctions, but as negotiations in which “the very good in question, or service in question, changes during the course of deciding how much will be paid for it, and what components will be offered into it.” (Tr. 2695:8-12 (T. Campbell), see also Tr. 2699:2-5 (T. Campbell) (“you’re actually defining the product and service and the price at the same time.”). The customers’ testimony confirmed this. (See Tr. 629: 19-20 (Bradley)(“We engage in very, very complex, tough negotiations with every one of our major vendors.”); Tr. 705:19-706:5 (Maxwell) (Neiman Marcus used professional negotiators for price discussions with the vendors); (Tr. 1408:7-9 (Gorriz) (DaimlerChrysler also relied on “specialized people for negotiating”).) Professor Hausman explained that the procurements in this industry are “structured as negotiations, not auctions.” (Tr. 3900:25-3901:6 (J. Hausman).) Professor Hausman explained that if these procurements were in fact bidding situations rather than

negotiations, customers would never voluntarily down-select to a limited number of competitors. (Tr. 3901:15-3902:7 (J. Hausman).)

151. Negotiations with two alternatives includes a variety of combinations. Outsourcers such as ADP have appeared in “bake-offs” with software vendors such as PeopleSoft, SAP and Oracle. (Tr. 4095:10-19 (J. Rising), 1415:9-18 (M. Gorriz).) Alternatively, some head-to-head competitions feature a software vendor competing against a company’s current legacy systems. (Tr. 1245:11-19 (C. Peters), 233:24 – 234:2 (R. Bullock).) And some companies down-select to two finalists, one being a Tier 1 vendor competing against a Tier 2 vendor. For instance, ExpressJet down-selected to a bake-off between Lawson and Oracle, and Esselte down-selected to a competition between Microsoft and SAP. (Tr. 2841:23 – 2842:6 (R. Knowles), Ex. D7159 (L. Spund 05/25/04 Dep.) at 50:22 – 51:10).)

152. Economics cannot predict results of bilateral negotiations. The economic theories relied upon by the government to predict that prices will increase if the number of vendors declines from three to two are not applicable in situations like this, involving bilateral negotiations and substantial buyer power. (Tr. 2722:4-2723:7, 2737:8-20 (T. Campbell).) Economic theory makes no determinate predictions about the results of such negotiations. (Tr. 2703:8-22 (T. Campbell).) In fact, fewer vendors can be preferable in bilateral negotiations, because it may lead to a more flexible and complete discussion of potential options. (Tr. 2730:16-2731:20 (T. Campbell).)

153. Negotiations are sometimes just bilateral, with no other “second” alternative. Given the leverage and tactics employed by buyers, the presence of competing vendors is not necessarily needed to obtain a favorable price. Esselte, for example, only evaluated and negotiated with one vendor, Microsoft, in its ERP procurement process, and nevertheless received favorable terms. (Ex. D7159 (L. Spund 05/25/04 Dep.) at 227:17-228:4; 230:22-231:3; 231:8-12.)

154. Customers have the negotiating power. SAP’s Richard Knowles observed that customers “control the selection process, and I would characterize it that it’s the

customers, along with maybe their consulting arm or maybe one of the analyst information services that they’ve subscribed to that’s coaching them, they have the power. They control the information flow, they control the negotiations. So we [at SAP] don’t characterize it that we’re in a position to take advantage of anybody.” (Tr. 2850:21-2851:7 (R. Knowles).) They are “sophisticated negotiators.” (Tr. 2850:13-16 (R. Knowles).) Similarly, Professor Elzinga conceded that customers and their chosen integrators – often the world’s largest consulting firms – structure the procurement process for the customer’s advantage. (Tr. 2345:13-17 (K. Elzinga).) As Laurette Bradley of Verizon made clear, customers are not afraid to use threats as a negotiating tactic with software vendors. (Tr. 644:17 – 645:15 (L. Bradley).) Customers have significant negotiation leverage and “the negotiation process... if done right, is skewed dramatically... in the buyer’s favor.” (Tr. 3225:14-15 (K. Harris).)

155. Customers control the information flow. Dean Campbell observed that, in bilateral bargaining, the purchaser controls what information is disclosed to vendors. (Tr. 2700:12-21, 2714:16-24, 2766:24-2767:4 (T. Campbell).) Dean Campbell testified that the documents he reviewed convinced him that buyers “control how much information and when it was released, sometimes not informing the vendor... of a competing bid even, or maintaining the presence of a possible competing bid from [another] vendor, even though it had long since disappeared.” (Tr. 2700:12-2701:4 (T. Campbell).) The witnesses confirmed this. Except in the rare cases where bidding is all public, customers have much more information than do vendors. (Ex. 3034 (D. Morea 5/7/04 Dep.) at 53:1 – 55:1) For example, North Dakota controlled the information flow and monitored the amount of information obtained by the vendors. (Tr. 1586:15-16 (C. Wolfe).) In its ERP procurement, MetroNorth “didn’t tell any other vendors anything about another vendor.” (Ex. P3061 (S. DeSimone 5/19/04 Dep.) at 167:13-14) Thus, “Oracle would never have the information that the customer would have that would be – that would be required to develop a valid ROI.” (P3179 (Ciandrini 1/16/04 Dep.) at 90:18-91:19; see also Ex. P3179 (Ciandrini 1/16/04 Dep.) at 102:1-10 (“a buyer is not going to tell me exactly how much they can spend.”)) SAP’s Richard Knowles testified that vendors never have perfect information about their competition and that “the information is cloudy oftentimes.” (Tr.

2844:3-7 (R. Knowles).) Customers have significant negotiation leverage and “the negotiation process . . . if done right, is skewed dramatically... in the buyer’s favor.” (Tr. 3225:14-15 (K. Harris).)

156. Customers control the product evaluation process. At the RFI stage of the procurement cycle, there are typically still eight to 15 vendors involved; the customer is still trying to understand its alternatives (including outsourcing) and which ones are viable. This will be determined by a number of factors including budget, functionality, costs, desired “productivity gains,” objectives, etc. (Ex. P3179 (Ciandrini 1/16/04 Dep.) at 96:3-101:22.) Several months into the procurement cycle, however, sales people are “probably talking to the prospect about what they think the functional requirements are to address the recognized need.” (Ex. P3179 (Ciandrini 1/16/04 Dep.) at 95:4-15.) Greyhound, for example, scored its potential vendors based on how well they met Greyhound’s functional requirements and on whether each requirement was a must-have, something that would save Greyhound “some headache,” or something that really did not matter. (Tr. 1476:9-16 (M. Glover).)

157. Customers may reveal identity of finalists as a negotiating tactic. SAP Richard Knowles lamented that, after “whittling” the number of competitors down to two (or sometimes three) finalists, customers may disclose the finalists’ identities, but they do so to use as leverage in negotiations. (Tr. 2848:21-2849:12 (R. Knowles).) The State of Arizona’s Timothy Boncoskey confirmed that customers have control over the bidding process and can release the names of competitors at will. (D8101 (T. Boncoskey 5/21/04 Dep.) at 62:17-22.) WBP disclosed to PeopleSoft and Lawson the fact that they were competing against each other as finalists in WBP’s ERP selection process, viewing this as advantageous because it allowed each to highlight relative strengths and weaknesses as compared to the other. (Ex. D8105 (R. Cushman 05/18/04 Dep.) at 66:24-67:12.) Nextel made sure PeopleSoft and Oracle knew they

were bidding against each other using “the logistics of the process” ensuring they would pass each other on the way to demos, etc. (Tr. 1074:12-18 (R. Cichanowicz)) In any event, the competitors’ identities are unhelpful to vendors without some idea of the customer’s relative views of each – which rarely is revealed. (Tr. 3464:4-14 (S. Catz).)

158. Customers do not reveal their preferences or finalists’ “performance scores” to finalists. To help them make a final selection, customers often subject finalists’ demonstrations to complex, multi-criteria scoring. Neither these scores nor any other indication of which finalist is in the lead is likely to be shared, however. For example, Metro North did not disclose its perception of its chosen finalists’ functional differences to the vendors. (Ex. P3061 (S. DeSimone 5/19/04 Dep.) at 171:25-172:4). North Dakota did not tell Oracle how it was faring “from a competitive perspective.” (Tr. 1586:2-4 (C. Wolfe).) Cox’s Scott Hatfield admitted that Cox withheld its “bake off scoring results and explained that these are the sort of data Cox would keep from any vendor. (Tr. 176:15-177-20 (S. Hatfield).) And Nextel continued negotiating with both Oracle and PeopleSoft, even though Nextel had decided to select PeopleSoft. Nextel did not tell PeopleSoft it preferred their product, and Nextel “[was] in control of the process.” (Tr. 1106:5-1107:15 (R. Cichanowicz).)

159. Customers tell finalists about the weaknesses of their proposals to create greater leverage. Customers are much more likely to be forthcoming when it comes to telling finalists why they may lose, to encourage further vendor concessions. “I would say, we, forever, if I’m in a negotiating process, I am forever going to tell the vendor where their product is weak, where I don’t need their product, where I’d prefer not to have their product, where I prefer not to have their product because I don’t want to have to put in a new release that includes a whole bunch of tests I don’t want to have to test for, et cetera. So I am forever going to be telling the vendor about the flaws, and I don’t necessarily always tell the vendor how important the flaws are or are not. I make it sound all very important.” (Tr. 3239:12-22 (K. Harris).)

160. Customers do not reveal their internal valuations of HRM and FMS product features to finalists. Customers do not tell finalists what value they place on the products that are subject to negotiation. (Tr. 3517:25-3518:9 (S. Catz.).) Dean Campbell

testified that in bilateral bargaining, the customer has a strong economic interest in controlling dissemination of information to the vendors – “particularly the value that the software constitutes to the purchaser,” which determines the customer’s reserve price. (Tr. 2700:22-2701:4, 2714:16- 24 (T. Campbell).) Professor Hausman similarly explained that the customers in this industry are “very sophisticated bargainers,” and would never “tell you what their value is, because they know you’ll use that information against them.” (Tr. 3873:9-12, 3877:6-19 (J. Hausman).) Thus, when CH2M Hill was negotiating with Oracle, Oracle was not told how it had scored relative to PeopleSoft. (Tr. 239:11-14 (R. Bullock).) Customers will tell vendors through the RFP and product demonstration process what business processes they would like to consider automating, what features they would like their software solution to have, and whether their current incumbent systems do or do not address those business needs, (Tr. 3518:19-3520:18 (S. Catz).) Customers do not share with vendors their internal valuation of satisfying those needs or alternatives, however, or even what particular needs or features are the drivers of the customer decision. (Tr. 3518:19-3520:18 (S. Catz.); Ex. P3061 (S. Desimone 5/19/04 Dep.) at 171:14- 172-4.) Customers recognize how they internally value products is extremely valuable information to vendors, and protect the information accordingly. (Tr. 177:6-20 (S. Hatfield); see also                                                                   Vendors learn as much as they can about a customer’s “needs” in the evaluation process, but vendors cannot know how much the customer actually values the various aspects of a vendor’s software.

161. Customers often will time negotiations for end of a vendor’s financial quarter. Customers know that if they wait until the end of the fourth quarter, discounts will go up. (Ex. P3050 (T. Kender 5/11/04 Dep.) at 244:17-245:11.) One PeopleSoft executive lamented: “We have created a monster in the software industry – the quarter end deal. . . . Many customers are now being coached or have experience with the quarter/year end software offers.” (Ex. D5956 (PeopleSoft Field Observations email).) Customers Aimco, Cox and CH2M Hill all acknowledged the effectiveness of this tactic. (Tr. 1144:1-7 (S.Wesson)); 179:11-24 (Hatfield); 243:4-7 (Bullock).) Charles Peters testified that Emerson “often negotiate[s] around end of quarters” and that when Emerson signed its deal for Oracle’s broad EAS suite near the end of a

quarter or fiscal year, there was “a great deal of negotiation about quarters and whose revenue would count in what quarter, and the like.” (Tr. 1204:5-10 (C. Peters).) Ken Harris recalled that, when he was negotiating with finalists, “I kept a calendar on my desk of the year-end for every vendor... whether it should be this way or not, the marketplace is such that software vendors try to get sales in the current quarter, so they can report good results to the marketplace.” (Tr. 3235:2-8 (K. Harris).) In the Oracle deal that Harris negotiated for the Gap, after letting two quarters more pass, in the third quarter, he waited and finally the Oracle salesperson “got in a cab and raced over to their office with the signed document, and I got my last concession from them at 11:15... you get an awful lot of power from timing.” (Tr. 3235:11-17 (K. Harris).) Systems integrators, like Deloitte, actually advise their clients on timing purchases to a software vendor’s year and quarter end as a negotiating strategy. (Ex. P3186 (D. Dortenzo 5/18/04 Dep.) at 320:23-327:1.)

162. Customers employ various other negotiating tactics to secure a good prices. A common customer negotiation tactic is to tell the vendor that the customer needs more products and/or services to be provided, given the total amount of money willing to be spent. Customers also use reference value as a bargaining chip. As Verizon’s Laurette Bradley observed, “[i]t is the fact that PeopleSoft is in a market where it cannot dare to put Verizon at a significant disadvantage with poor maintenance or exorbitantly high costs because it would risk losing us as a reference, formal or informal, for a tremendous number of other customers that it is seeking for its software.” (Tr. 596:2-7 (L. Bradley).) Similarly, DaimlerChrysler’s Michael Gorriz acknowledged that Daimler uses its value as a references to bargain for lower prices. (Tr. 1410:2-9 (M. Gorriz).) Esselte CIO Lani Spund has a “portfolio of negotiation tactics” that he uses, including the “good guy/bad guy approach,” and utilizing surrogate potential third parties that he does not plan to buy from to put pressure on other vendors. (Ex. D7159 (L. Spund 05/25/04 Dep.) at 235:16-236:12.) Other customers demand multiple “final” offers (Tr. 1560:22-1561:3 (C. Wolfe)); use their large size as a negotiating advantage (Tr. 1407:15-1408:1 (M. Gorriz)); and use idle threats of doing nothing to extract vendor concessions in the course of negotiation. (Tr. 1434:2-6 (M. Gorriz). Ken Harris confirmed that customers let the vendor

know how they might solve the problem otherwise, the other things they can do — from doing nothing for the time being, choosing point solutions, choosing other vendors, limiting the scope of the package to be acquired from the vendor at hand (Tr. 3225:14-23 (K. Harris).) “The objective that you are always trying to reach and you keep ringing the bell about is that there are always other things that you can do, and you make that vendor know that.” (Tr. 3234:13-16 (K. Harris).) Thus, Harris has had two negotiations going on in proximity, in two separate conference rooms with glass walls between them, so that each vendor can see the negotiations occurring with their competitors. (Tr. 3235:18-24 (K. Harris).) He has had one competitor’s negotiation convene early so that the other would see its competitor leave early and think that the deal had been wrapped up and that they were in danger of losing the opportunity (Tr. 3236:1-13 (K. Harris).) and he has left marketing literature out from other companies. (Tr. 3236:14-15 (K. Harris); Professor Elzinga explained that purchasers “mak[e] threats... all the time when they’re negotiating to buy a capital good.” (Tr. 2355:15-24 (K. Elzinga).) He also acknowledged that “as a matter of bargaining strategy,... knowing what I know about this market [were I] a purchasing agent looking at high function enterprise software, I probably would, to some extent, give strong hints that, hey, I may just stay with my incumbent system.” (Tr. 2356:1-17 (K. Elzinga).)

163. Customers can and do hire negotiations specialists and other industry experts to assist them. Neiman Marcus used the services of a “negotiation support company,” which they believe saved their corporation $500,000 in negotiating price with Oracle. (Tr. 705:19 – 706:23 (P. Maxwell); see also Ex. P3186 (D. Dortenzo 5/18/04 Dep.) at 320:23-327:1 (stating Deloitte advises its clients on various strategies for obtaining the best possible prices from software vendors, including timing purchases to a software vendors year and quarter end); Ex. P3198R (C. Bass 4/21/04 Dep.) at 133:24-135:2 (stating Accenture does not engage in price negotiations; however, about 30% of the time Accenture is asked by clients to advise on strategies for obtaining the lowest price or to put pressure the vendor to reduce its price).)

164. Customers can “do nothing” or credibly threaten to do nothing. During negotiations, customers can, and often do, threaten to not procure an HRM or FMS

solutions, and instead continue to use the customer’s existing systems. (See, e.g., Tr. at 1245:14-19 (C. Peters). This threat is legitimate, as a significant percentage of customers decide to “do nothing,” or not to purchase a new core HRM or FMS product, even after going through a vendor selection process. (Tr. 2710:13-23 (T. Campbell); Tr. at 1490:7-22 (M. Glover) (Greyhound evaluated an HR system from 2001 through 2003 and again elected to do nothing for reasons unrelated to the tender offer).) Accenture estimates that as many as 20-30 percent of its clients evaluate software but elect to do nothing. (Tr. 1673:14-21 (C. Bass).) Vendors recognize that doing nothing is an option for customers, and will discount accordingly. (Tr. at 1792:17-24 (P. Wilmington).)

165. The initial sale is just the starting point. Negotiations with vendors continue long after the initial purchase is made. (Tr. 3241:25-3242:19 (K. Harris).) As Ken Harris explained, “[T]he process of getting value out of software – getting solutions for business – is not equivalent to buying apples. When you actually finalize a negotiation to buy some software you are actually at the very beginning of a pretty painful and extensive process called ‘implementing’ software, changing the business using the software. And so, the fact that you have now negotiated to acquire some technology is really only the starting.” (Tr. 3242:7-15 (K. Harris).) The incumbent vendor enjoys advantages, such as making follow-on sales more readily (Tr. 3245:7-8 (K. Harris)), but incumbency can be a double-edged sword because if the incumbent’s technology is problematic, or the customer is getting bad service it operates as a real negative, and the vendors will not only lose any other business from the customer, the word will get out to other CIOs and people in the industry about the negatives. (Tr. 3218:19-3219:1; 3242:25-3243:10 (K. Harris).) Thus, customers retain leverage over incumbent vendors even after the initial purchase.

166. The loss of a customer is very costly to a vendor in the present and in the future. The loss of a customer or potential customer results in significant lost revenues and revenue opportunities. For instance, if Oracle lost a sales opportunity, it would lose revenue from not only the initial software license, but also lose revenue that could be derived from the

sale of additional modules, annual maintenance contracts, and consulting services. (Tr. at 3458:1-22 (S. Catz).) Jay Coughlan of Lawson characterized it as a “zero sum game,” stating “you either win and get it all or you get zero.” (Tr. at 3774:24-3774:3 (J. Coughlan).) Indeed, Oracle committed to support PeopleSoft products in an effort to maintain the PeopleSoft customer base post-acquisition. Maintaining these customers will enable Oracle to sell additional application software, database, application servers, and outsourcing services. (Tr. at 3448:23-3450:2 (S.Catz); Tr. at 4263:20-4264:7 (L. Ellison).)

167. Vendors value customers on a total value basis, with little regard for the actual modules and functionality purchased and more with an eye to the long term prospective opportunities. The factors that Oracle includes in approving a discount include: whether the customer is in a vertical or industry in which Oracle is interested in establishing a presence, whether the customer would be a good reference, whether this is a new Oracle customer, and whether there is potential to sell additional products. (Tr. 3460:11-20; 3462:16- 3463:21 (S. Catz); Tr. 2845:12-25 (R. Knowles).) Discounting is common in this industry- in fact, Knowles from SAP testified that during his tenure, he never denied a discount. Mr. Knowles’ behavior would be irrational if the only thing at stake were the initial license fee; however, what is at stake is the (much larger) total value of the customer, including revenue from maintenance fees, and implementation services, and indirect benefits such as references.

168. Alternatives to PeopleSoft exist now and will remain post-merger. For example, Emerson Electric Company signed its deal for Oracle’s broad ERP suite, including FMS and HRM and SCM functionality, after considering the options of best of breed, continuing with its legacy system, SAP, JD Edwards, and Oracle. (Tr. 1200:11-1201:1; 1264:19-21 (C. Peters).) Emerson’s Charles Peters does not believe that the acquisition of PeopleSoft would result in higher prices for software because Emerson had been “very, very successful” in its

negotiations for Oracle’s ERP suite and, as he did not use PeopleSoft for leverage in that negotiation, PeopleSoft’s acquisition “provides no negatives for me.” (Tr. 1235:11-24 (C. Peters).)

169. SAP is and will continue to be an alternative. SAP will continue to be an alternative. SAP is the acknowledged leader in the EAS industry, and is strong in every vertical. (Tr. at 2834:8-21, 2825:13-15 (R. Knowles); D5000 (SAP Customers Within Global 500).) SAP will continue to be the market leader after this transaction. (Ex. D5633 (SAP Acquisition Scenarios Analysis) at SAP29027.) The government has suggested that SAP is less strong in certain verticals such as financial services; however, there is no evidence that this is due to a lack of core HRM or EMS functionality, and in fact, SAP does compete heavily for deals in the financial services space. (Tr. at 2875:5-2876:16 (R. Knowles); Tr. at 950:10-951:12 (P. Keating).) Moreover, SAP is investing in plans to move into this segment with its Accenture partnership. (Tr. at 1632:25-1636:12 (C. Bass); Tr. at 2875:16-2876:4 (R. Knowles).)

170. Lawson is and will continue to be an alternative – Law son has the demonstrated ability to serve large enterprise customers in its verticals. (Tr. 3599:18-3623:7 (J. Coughlan).); Tr. 2096:8-10 (M. Iansiti).) Lawson’s business strategy is to develop deep expertise in certain industry verticals. (Tr. 3592:20-3595:7 (J. Coughlan).) This strategy is well-suited to serve as a constraint on a merged Oracle-PeopleSoft post-merger, in part because the verticals in which Lawson is strongest (healthcare, retail, and increasingly financial services) are areas where SAP is not already dominant. (Tr. 3597:19-25, 3602:22-3603:7, 3608:11-20, 3617:4-3618:2 (J. Coughlan); Tr. 2942:16-25 (R. Knowles).); Tr. 3839:4-3840:21, 3841:14-3842:2 (J. Hausman).) Lawson has experienced recent wins with this strategy, including its selection by the State of Arizona over PeopleSoft and SAP. (Ex. D8101 (T. Boncoskey 5/21/04 Dep.) at 29:12-17.) Lawson also has the incentive and the opportunity to reposition itself into other verticals, post- transaction.

171. AMS is and will continue to be a viable alternative in the public sector. AMS has been particularly effective in the public sector with more successful implementations in public sector customers over the past five years than all of its competitors

combined (Ex. 3034 (D. Morea 5/7/04 Dep.) at 101:7 – 18). Indeed, the Department of Justice recently chose AMS’s FMS over Oracle, PeopleSoft, and SAP (Ex 3034 (D. Morea 5/7/04 Dep.) at 21:22-23:21).

172. None of the other alternatives will disappear if PeopleSoft goes away. Point source or “best-of-breed” will continue to be an alternative; outsourcing will continue to be a viable alternative; Doing nothing will continue to be an alternative.

173. Oracle’s acquisition of PeopleSoft will not prevent CIOs from negotiating effectively in their EAS purchases. Debra Anderson of Novell testified that even if Oracle and PeopleSoft were to merge she would continue using the numerous negotiation techniques at her disposal, including using other vendors such as SAP as leverage in the negotiation process. (Ex. D8107 (D. Anderson 06/03/04 Dep.) at 181:1-14.) Anderson does not believe that a merger between Oracle and PeopleSoft would prevent Novell from obtaining the ERP products it needs at a competitive price. (Ex. D8107 (D. Anderson 06/03/04 Dep.) at 180:18-22.) Verizon also can obtain EAS products at a competitive price after the merger, since they licensed SAP products “just in case” the merger went through, thus locking in a low price for redundant software. (Tr. 623:4 – 624:9 (L. Bradley).) Nextel’s Richard Cichanowicz testified that he had been told by the “Big Five” consulting firms that SAP could meet Nextel’s functional needs. (Tr. 1103:18-21 (R. Cichanowicz).) Harris also testified that he did not believe the merger would ruin his ability to negotiate effectively because he “find[s] the solutions I need. It would simply be a different fact pattern, and the fact pattern is changing all the time. It is always changing. So, I don’t think would have any negative effect whatsoever.” (Tr. 3248:1-6 (K. Harris).)

174. Vendors still desire reference customers. Remaining vendors will still desire, and make price and other concessions to get good references from large customers.

Existing customers may provide references for a vendor. (Ex. P3184 (D. Dortenzo 5/5/04 Dep.) at 194:9-195:20 (it is “critical” for a vendor to have client references in order to be considered a viable alternative); Tr. at 2845:12-2846:22 (R. Knowles); Tr. at 3463:1-6 (S. Catz).) Customers use their value as a reference to get concessions out of vendors. (D8101 (T. Boncoskey Dep.) at 51:18-25; (Ex. D7159 (L. Spund 5/25/04 Dep.) at 56:19-58:17; 75:8-21.).)

175. Vendors still need to tend reference customers. It remains in vendors’ interests to develop and maintain good reference accounts. (See, e.g., Ex. P3041 (B. Patel 6/3/04 Dep.) at 107:2-14 (vendors have to value a company like Ford for its reference value as well as the purchases it makes”).) As Ken Harris explained, “When you’re doing good by me, I have a vested interest in being a reference account for you because I want you to succeed, and you have a very vested interest in having me be a positive reference.” (Tr. 3219-12 (K. Harris).) Indeed, poor service and support from an incumbent vendor can lead customers to switch. Harris testified that while he was at the Gap, and using some portions of PeopleSoft FMS, Gap had a “very difficult time getting support from PeopleSoft” and that was a factor in Gap ultimately choosing Oracle for FMS. (Tr. 3205:2-15 (K. Harris); see also Tr. 3243:4-6 (K. Harris) (CIO’s communicate frequently, whether at industry conferences or directly by telephone); Ex. P3041 (B. Patel 6/3/04 Dep.) at 112:23-113:9.) (loss of a customer like Ford would result in the loss of much more than just immediate license fees”).)

176. Customers still use the references. Customers use the references in their procurements, and the vendors know it. (Ex. P3184 (D. Dortenzo 5/5/04 Dep.) at 194:9-195:5 (“Maybe 65, 70 percent of the time clients ask for references” and in perhaps “60 percent of those occasions they follow through in calling those references.”).) Lawson’s reference customers were important to the State of Arizona in its ultimate selection of Lawson. (Ex. D8101 (T. Boncoskey 5/21/04 Dep.) at 27:3-9; see also 3230:25-3231:5 (K. Harris) (“For vendors of technology, the acquisition of a big customer, the power of incumbency is so great that as a CIO of a large company, I get constantly — I get all kinds of companies coming at me offering to make all kinds of deals if I will, you know, use their technology, and be a good reference for them.”);

177. Pricing and discounting is done on a bundle or “blended” basis. EAS is most often sold and priced in a bundle of modules from one or more pillars along with associated services. (Tr. 3453:16-3454:2; 3455:14-3456:10 (S. Catz).) The discount negotiated between supplier and customer can be a blended discount across the various parts of the bundled software. (Tr. 3813:12-3814:22 (J. Hausman).) What really matters to customers is the total amount of the deal, including services, not the price of any individual item. (Tr. 3774:19-23 (J. Coughlan).) License fees and discounts are negotiated for the cluster of modules a customer is considering, not on a module-by-module basis. (Tr. 3772:10-3774:23 (J. Coughlan); Tr. 3453:16-3454:2 (S. Catz); (Tr. 3814:8-3815:12 (J. Hausman).)

178. Vendors still want to make add-on sales. Remaining vendors will still desire, and make price and other concessions to protect the prospect of additional future sales to a customer. (See Tr. at 2846:24-2847:25 (R. Knowles) (SAP considers the possibility of follow- on consulting sales and the potential for future product sales in making price discount decisions); Tr. at 2557:24-2558:10 (P. McAfee) (McAfee conceded that the possibility of selling Oracle consulting services to a customer “would make [Oracle] more aggressive on the price generally”); see also Tr. 3463:7-12 (S. Catz) (prospect for future sales influences discounting decisions).) Control of follow-on business is one of the reasons that buyers have greater bargaining power than vendors. Tr. 2359:19-24 (Elzinga).); Tr. 361:1-362:25 (R. Bergquist); (Ex. 3034 (D. Morea 5/7/04 Dep.) at 56:20-24).)

179. Government expert agrees future sales are very important to vendors. The government’s expert, Professor Elzinga, acknowledged vendors have the “prospect or the hope . . . that once they get their foot in the door” they may be able to sell additional modules, including modules outside the FMS and HRM pillars—and that that prospect will constrain vendors’ pricing strategies. (Tr. 2356:25-2358:8 (K. Elzinga).) Professor Elzinga further explained that it would be costly to a vendor if it attempted to raise prices for FMS and HRM because it could “tick[] off” customers and “lose sales” that might have been made to that customer in other pillars. (Tr. 2361:4-2362:1 (K. Elzinga).) Professor Elzinga has studied the

importance of reputation as a constraint on opportunistic pricing in connection with his expert work for Kodak, and he acknowledged the powerful effect of a vendor’s reputation for fair dealing in securing future sales in this industry. (Tr. 2363:11-2366:1 (K. Elzinga).) Professor McAfee also acknowledged that the possibility of future follow-on sales “could be relevant to the discount that’s struck” and that “from an economic perspective . . . you would want to account for that” though he had been “unable to do that.” (Tr. at 2551:6-17, 2556:13-2557:3 (P. McAfee).)

180. The desire for future sales of products and services beyond the FMS and FRM pillars provide customers with significant post-license leverage. As Professor Elzinga confirmed, the desire for future sales of products “outside the FMS and HRM pillars” (i.e., CRM, SCM, etc.) impacts vendor pricing strategies. Professor Teece agreed and testified that Oracle’s behavior with respect to the applications market and pricing in that market is going to be conditioned on the fact that Oracle wants to sell the customers other things in additiona to applications. (Tr. 4394:7-4395:6 (D. Teece).) Professor Teece explained that under the basic economics of the this industry “when Oracle’s trying to sell [customers] other things from the stack . . . [i]t would be the most foolish thing in the world, they’d be shooting themselves in the foot if they then disaffected those customers through over-pricing or whatever strategy the government contemplates, because the idea is to sell them Oracle application service, Oracle databases, and to generate the revenues to fuel innovation.” (Tr. at 4394:18-4395:1 (D. Teece); Tr. 3463:13-21, 3523:5-24 (S. Catz) (Oracle’s discounting decisions influenced by prospect for future database sales and opportunity to expand Oracle infrastructure “through that enterprise more and more, so it will be an Oracle platform company”).)

181. The identity of competitors may be relevant, but it is not a significant factor in a vendor’s ultimate pricing decisions. SAP is “competitor agnostic” in making

pricing decisions – the identity of competitors is “insignificant” because “it is not about the competitor, it’s about winning.” (Tr. 2848:1-10 (R. Knowles).) In the ExpressJet competition, Knowles testified that SAP’s discounting would not have been different if the short-list had included Microsoft and Lawson rather than Oracle and Lawson. (Tr. 2848:11-20 (R. Knowles).) Similarly, Oracle’s discounting decisions are based on winning the deal, not on who the competitor is. (Tr. 3436:4-14, 3523:25-3524:17 (S. Catz).)

182. Vendors cannot identify vulnerable, inelastic customers. As, Professor Elzinga testified, while “price discrimination” might occur in a literal sense whenever a seller sells a product at two different prices where the difference is not accounted for by differences in cost (Tr. 2221:23-2222:2 (K. Elzinga)), he was clear to explain that a vendor cannot engage in a price discrimination strategy to prevent marginal customers from disciplining the prices charged to infra-marginal customers, unless it can reliably distinguish between those customers with elastic demand for the product and those with inelastic demand: (Tr. 2222:24 - 2224:11 (K. Elzinga); see also (Tr. 2715:11-13 (T. Campbell) (Dean Campbell and Professor Hausman explained that price differences—and even price differences unrelated to differences in costs— are not the same thing as a systematic price discrimination strategy).) Prices and discount vary in this market for a number of reasons unrelated to any systematic price discrimination strategy. (Tr. 2717:10-24 (T. Campbell) (because the components being purchased are different in every transaction; Tr. 2718:4-9 (T. Campbell) (because some customers have greater bargaining power and can use it to secure a better price); Tr. 2718:10-16 (T. Campbell) (because some customers are particularly valuable to vendors as a potential reference); Tr. 2729:4-10 (T. Campbell) (because vendors are more eager to sell at the end of their fiscal year or quarter); Tr. 2718:17-21 (T. Campbell) (because some customers are just better negotiators than others, or hire sophisticated consultants to help them negotiate the best possible price).)

183. There are no criteria by which vendors can identify inelastic customers. Professor Elzinga explained that customers are “inelastic” if they have fewer realistic options (Tr. 2338: 23-25 (Elzinga), but when asked what the three best markers to distinguish an inelastic customer would be, Professor Elzinga conceded that he does not know.

(Tr. 2341:23-2342:2 (K. Elzinga).) Ms. Catz confirmed that Oracle does not have full information and cannot identify which customers might be vulnerable such that Oracle could take advantage and price discriminate. (Ex. P3031 (S. Catz 5/20/04 Dep.) at 98:1-99:3).)

184. Large customers tend to get better prices. Professor Elzinga admitted that he had not studied whether a customer’s size and complexity indicate its inelasticity. To the contrary, he suspects that, generally speaking, larger enterprises get higher discounts. (Tr. 2336:18-2337:15 (K. Elzinga).) Professor McAfee also acknowledged that customers “who have big deals get a bigger discount just for that reason.” (Tr. 2500:11-2501:5 (P. McAfee).) Professor Elzinga also acknowledged that he has not attempted to study whether there is any correlation between how many countries, languages, or currencies a customer operated in and the discounts they received. (Tr. 2343:7-2344:5 (K. Elzinga).) This is borne out in the vendors’ behavior. SAP, for example, provides automatic standard discounts off of the list price based on the amount the customer will spend. (Tr. 2836:25-2837:7 (R. Knowles).) While the government suggests that large companies pay a higher total amount for software than do smaller purchasers, there is no proof that these differences are the result of anything other than the fact that larger purchasers are buying more “units” - that is, the right for more of its employees to use the software. The government has not shown, for example, that large companies who buy on a centralized basis get larger or smaller discounts than large companies who buy on a decentralized basis. (Tr. 3816:15-21; 3869:24 - 3870:8 (J. Hausman); Tr. 3465:16 - 3466:3 (S. Catz).)

185. Vendors make no effort to identify inelastic customers. The remaining vendors recognize that large complex enterprises are not necessarily “inelastic” customers. SAP makes no effort to analyze whether a customer is “vulnerable” when it makes discounting decisions and is not in a position to “take advantage of anybody.” (Tr. 2850:17-2851:7 (R. Knowles).) Oracle cannot price discriminate because it cannot determine which customers are so vulnerable that would allow Oracle to increase its prices. (Tr. 3458:1-22 (S. Catz).)

186. Vendors can not dependably isolate the “prices” for FMS or HRM because it’s sold in a bundle that is of a total negotiated value. Dean Campbell explained that Oracle’s discount approval forms sometimes request a 100% discount on certain products, “but

obviously the Oracle company is... making up a zero price for [that] software with a non-zero price in servicing, or in a non-zero price for another product.” (Tr. 2717:19-24 (T. Campbell).) Professor McAfee agreed that the discounts offered in these transactions are often “blended” across all of the products being sold in the bundle, including products that are not in the relevant markets for this case. (Tr. 2493:4-18 (P. McAfee).) When asked by the Court whether the discount request forms he reviewed in fact combine “apples, oranges, pears and kumquats,” Professor McAfee conceded that “[t]hey do include some things that, in a perfect world, would not be present.” (Tr. 2493:19-23 (P. McAfee).)

187. Vendors risk the total value of the transaction (beyond the HRM and FMS fees) if they “guess wrong” about the customer’s elasticity. Professor Elzinga acknowledged that in calculating the “critical loss” of business that would render a hypothetical price increase unprofitable in this industry, it would be appropriate to count profits lost on the entire bundle of products that would have been sold to that customer. He has not, however, attempted to calculate how much business Oracle would need to lose on other modules or pillars in order to render an attempted price increase on core HRMS or FMS unprofitable. (Tr. 2361:4- 2362:1 (K. Elzinga).) Ms. Catz confirmed that Oracle would not be able to charge some customers higher pries post merger because “the customers have many alternatives, because we don’t have full information of who even these quote, vulnerable customers would be so that in the event that we guess wrong, we would lose much more than we could possibly gain by raising their price say ten percent or whatever it is, and the risk of being wrong so outweighs any potential value of increasing prices on these customers.” (See (Tr. 3466:8-3467:18 (S. Catz); Ex. P3031 (S. Catz 5/20/04 Dep.) at 98:1-99:3).)

188. Core HRM and FMS are not sold in an “oral ascending auction.” Though his model assumes that competition is characterized as an “English auction or oral ascending auction” that assumes that there are “multiple rounds of bidding” (Tr. 2511:12-22,

2586:9-20, 2587:15-19 (P. McAfee)) Professor McAfee testified that he had not in fact reached a professional conclusion based on his own economic analysis that that was true in the sale of HRM and FMS software. (Tr. 2540:3-20 (P. McAfee); Professor Hausman explained that if procurements in this industry genuinely resembled auctions or “bidding,” then customers would never limit the number of vendors that they invited to compete. (Tr. 3901:3-3902:7, 4070:18-23 (J. Hausman).)

189. McAfee model does not apply to negotiations. Professor McAfee acknowledged that auctions are an appropriate economic model for bidding situations, and that bidding is distinct from negotiation or “bargaining.” (Tr. 2509:9-2510:22, 2589:16-2590:21 (P. McAfee).) Professor McAfee conceded that he was not familiar with any economic literature discussing the economic effects of mergers in markets characterized by simultaneous bilateral negotiations. (Tr. 2591:5-11 (P. McAfee).)

190. HRM and FMS software is sold through negotiations, not bidding. Professor McAfee testified that “many, not all, but many [discount approval forms] show... multiple rounds of bidding. That is, it’s not a one-time best and final offer, but instead an iterative process in which the pieces are adjusted and readjusted.” (Tr. 2459:14-18 (P. McAfee).)

191. The McAfee model incorrectly assumes the vendors know the prices other vendors are offering. Professor McAfee’s model assumes the vendors know what prices the other vendors are offering to the customer throughout the “bidding.” (Tr. at 2588:3-15 (P. McAfee).) However, this information generally is not provided to the vendor. (Ex. D7166 (D. Morea 5/17/04 Dep.) at 51:24-52:4) (uncommon for customers to share with AMS the competing vendor’s prices or details of their offerings); Ex. P3198R (C.Bass 4/21/04 Dep.) at 133:8-135:2 (price negotiations are typically extremely confidential between the software vendor and the customer buying the software); Tr. at 1586:11-16 (C. Wolfe) (North Dakota did not reveal final prices to the finalists, resulting in Oracle bidding $3 million less than PeopleSoft); Ex. D8109 (S. Holland 5/24/04 Dep.) at 77:21-78:1 (CDI did not disclose to either competing vendor – Microsoft and Deltek –the price being offered by the other vendor.); Tr. at 1205:3-5 (C. Peters) (Emerson Electric did not reveal to Oracle that JD Edwards had increased its price during early

negotiations); Ex. P3179 (Ciandrini 1/16/04 Dep.) at 152:11-153:9 (Oracle has “very incomplete potentially inaccurate information” when negotiating and “very rarely” knows “whatever [the] competitor is offering”).)

192. The McAfee model incorrectly assumes the vendors have “full information” about how the buyer values the software. Professor McAfee’s model assumes that all participants “know the values that the buyer places on the software” of each vendor. (Tr. at 2518:21-2519:11, 2561:9-16, 2588:20-2589:2 (P. McAfee).) However, even Professor McAfee recognized that this assumption was not realistic. (Tr. at 2561:1-4 (P. McAfee); see also Tr. at 2700:22-2701:4, 2714:16-24 (T. Campbell) (testifying that in bilateral bargaining, the customer has a strong economic interest in controlling dissemination of information to the vendors – “particularly the value that the software constitutes to the purchaser,” which determines the customer’s reserve price).) And in fact it is not true – vendors do not know what value customers place on the software. (Tr. at 2849:16-2850:3 (R. Knowles) (customers do not reveal vendors the “reserve value” or what the software is really worth to the customer); Tr. at 3873:9-12, 3877:6-19 (J. Hausman) (customers in this industry are “very sophisticated bargainers” and would never “tell you what their value is, because they know you’ll use that information against them”).)

193. The McAfee model has only two “real world” inputs, neither of which is factually established. Professor McAfee’s auction model depends on only two “real world” inputs: market shares and a parameter estimating the proportion of the value of these products that is captured by customers (i.e., consumer surplus). (Tr. 2516:17-2517:3, 2591:12-24 (P. McAfee).) Professor McAfee did not use any actual prices or transaction-level data from this market in his model. (Tr. 3899:3-20, 3905:24-3906:7 (J. Hausman).) Neither of the inputs that Professor McAfee used to calibrate his model are adequately supported by the evidence.

194. McAfee assumptions on price/value relationship are unsupported. To estimate the consumer surplus parameter, Professor McAfee simply assumed that Oracle’s list prices are equal to the total value of these products, at least to smaller buyers. (Tr. 2517:5-2518:12 (P. McAfee); Ex. P4024A (McAfee Demonstrative).) Professor McAfee provided no

evidence to justify that assumption, and he conceded that Oracle’s eBusiness discounts off of list price are essentially automatic for deals over $250,000, and that if he had used the discounted list price to estimate the consumer surplus parameter then his model would have predicted smaller price effects. (Tr. 2593:9-2594:4 (P. McAfee).)

195. The McAfee model dependence on Elzinga shares is unsupported. Professor McAfee also conceded that he used market shares to proxy consumers’ preferences for different products in his simulation, and that he relied entirely on Professor Elzinga’s market share calculations. (Tr. 2563:6-9, 2513:11-2514:9, 2522:21-2523:5, 2599:11-2600:3 (P. McAfee), Ex. P4024A (McAfee Demonstrative).) Professor McAfee assumed that a vendor is the second or third choice of customers in the same proportion that it is their first choice—in other words, in proportion to that vendor’s market shares. (Tr. 2616:14-17 (P. McAfee).)

196. The McAfee model is not accurate if market shares are not accurate. Professor McAfee agrees that the predictions of his model are only as good as the market share numbers. (Tr. 2600:13-15 (P. McAfee).) For example, Professor McAfee’s model predicts much smaller price effects in the FMS market solely because Professor Elzinga’s market shares for SAP are higher in FMS than in HRM. (Tr. 2531:6-19 (P. McAfee).)

197. Professor McAfee concedes that his results are unrealistic. The price increase predictions of Professor McAfee’s model vary wildly with changes in these mathematical assumptions. (Ex. P4024A (McAfee Demonstrative).) Professor McAfee’s auction simulation projects price increases that are not credible since no Oracle customer will pay double what it pays now before switching to SAP or some other alternative. (Tr. 2596:23- 2597:11, 2597:20-23 (P. McAfee); (Tr. 4147:20-21 (D. Kutnick); (Tr. 4254:6-4258:4 (L. Ellison); D5613R, page 31 (Financial Model) (showing decrease in average sales price over time).) Professor McAfee admitted that the 30% price increase predicted by his model – across all customers – under certain assumptions is “jarring” and unrealistic. (Tr. 2598:5-11 (P. McAfee).); (Ex. P4024A (McAfee Demonstrative).) Ultimately the only conclusion Professor McAfee was willing to draw from his simulation was that “a significant economic effect is consistent with basic economic analysis,” and that it is not possible to say “from a theoretical

perspective” that the numbers produced by his regressions are “too large” to be credible. (Tr. 2531:24-2532:14 (P. McAfee).)

198. McAfee model always predicts a price increase. Professor McAfee conceded that his auction model will predict a price increase from any merger. (Tr. 2600:16-20 (P. McAfee).)

199. McAfee model does not predict any post-merger harm for customers who view SAP as a close alternative. Professor McAfee acknowledged the economic theory of his simulation is that the merger would affect prices only for those customers who currently rank Oracle and PeopleSoft as their first and second choices. (Tr. 2595:6-21 (P. McAfee).) Professor McAfee explained that the price in an auction is ultimately set by “the value offered by... the best losing bidder.” (Tr. 2512:8-11 (P. McAfee).) Customers for whom SAP (or any other alternative) is their first or second choice will not, under Professor McAfee’s economic theory, be affected at all. (Tr. 2595:11-2596:3 (P. McAfee).)

200. Even if SAP is a close third, Professor McAfee predicts no conomic harm from the merger. Even a customer who currently ranks Oracle and PeopleSoft #1 and #2 will not be harmed if the gap between the #2 and #3 options for that customer is not substantial. (Tr. 2596:4-12 (P. McAfee).) Further, he says, “[i]f these products are all very similar, then the market is going to be pretty competitive, and you don’t have to go very far down the list in order to reach the third” bidder. (Tr. 2529:11-15 (P. McAfee).) In order to produce average price effects as large as 15-30%, therefore, Professor McAfee’s model is in fact projecting that the individual customers who currently rank Oracle and PeopleSoft #1 and #2 “would pay double what they are paying now before they would turn to SAP.” (Tr. 2596:23-2597:11, 2597:20-23 (P. McAfee).)

201. Oracle’s new market position would spark enhanced competition between Oracle and SAP. In a post-merger world, SAP anticipates “a much more highly competitive environment to try to win business.” (Tr. 2860:6-7 (R. Knowles); see also 2858:9- 21 (R. Knowles).) The merger would allow Oracle to leap past SAP as the number one EAS provider in the United States. (Tr. 2859:1-2860:7 (R. Knowles).) Richard Knowles testified that

SAP’s “analysis and our intelligence tells you that that’s going to create a hugely more competitive environment, because Oracle would gain No. 1, and they’ll want to maintain that No. 1 position. So, they’ll do anything that they are going to do to keep that position.” (Tr. 2859:1-2860:7 (R. Knowles).) Indeed, some analysts are predicting a coming “battle royale” between a post-merger Oracle and SAP over accounts where both PeopleSoft and SAP are present today. (Tr. 2860:25-2861:6 (R. Knowles).)

202. Professor McAfee relied on Professor Elzinga’s flawed market shares and some profound assumptions about causation. Professor McAfee’s statistical regressions afford no basis for inferring that Oracle and PeopleSoft are uniquely close substitutes, or that PeopleSoft forces Oracle to compete more aggressively than other alternatives. At most, Professor McAfee’s regressions demonstrate only that the presence of PeopleSoft in a deal is modestly correlated with higher discounts from Oracle. (Tr. 2497:7-9 (P. McAfee).) Professor McAfee assumes that correlation is synonymous with causation, so that if PeopleSoft were not available as a competitor, Oracle’s prices would therefore go up. (Tr. 2450:8-11, 2497:7-9 (P. McAfee).)

203. Professor McAfee acknowledges that customer reference value and the prospect for future sales – not just PeopleSoft’s presence — are likely cause for Oracle to discount. But Professor McAfee’s regressions did not account for many other factors that can – in some instances by his own admission – affect discounts, such as whether the customer had value as a reference customer or a buyer of follow-on products or services, whether the sale took place late in the quarter, whether the customer had had a good or bad previous experience with Oracle. (Tr. 2551:6-17, 2556:13-2557:3 (P. McAfee) (acknowledging that “from an economic perspective. . . you would want to account for” whether a customer was a good prospect for further sales, but that he was “unable to do that”); 2564:18-21; 2557:24-2558:17; 2562:6-22; 2564:22-2565:1; 2578:14-19 (P. McAfee).) So it is no surprise that his regressions explain no more than a third of the overall variation in Oracle’s discounts. (Ex. P4023A, Tr. 2576:14- 2577:5, 2585:6-10 (P. McAfee), Tr. 2504:3-20 (P. McAfee) (acknowledging that PeopleSoft could be “proxying for” some other unidentified factor in the data).)

204. Professor McAfee did not isolate sales of core HRM and FMS in his regressions. Professor McAfee’s regressions were based on “bundle” deals that included lots of products outside of the market. (Tr. 2499:6-9 (P. McAfee) (“All of the data concerns the E-Business Suite product.”), Tr. 2573:16-2574:25 (P. McAfee) (conceding that it is “possible” that the customer acquired E-Business Suite only for SCM or CRM functions); see also id. at 2582:22-2583:2 (P. McAfee) (conceding that “I didn’t have a systematic way of identifying the extent in which deals were of – of the relevant products for this case.”), see also Tr. 2550:5-13, 2575:1-7, 2578:20-2579:16 (P. McAfee).) Professor McAfee conceded that the discounts offered in these forms are often “blended” across all of the products being sold in the bundle, including products that are not in the relevant markets for this case. (Tr. 2493:4-18 (P. McAfee).) When asked by the Court whether the discount request forms he reviewed in essence combine “apples, oranges, pears and kumquats,” Professor McAfee conceded that “[t]hey do include some things that, in a perfect world, would not be present.” (Tr. 2493:19-23 (P. McAfee).) In fact, Professor McAfee made no effort to exclude from the sample he regressed deals that (like the Hanover Compression transaction discussed in his testimony) appeared to involve no HRM or FMS software at all. (Tr. 2580:9-2581:23 (P. McAfee).)

205. Professor McAfee did not apply Professor Iansiti’s definition or any other screen for “high-function” software. Like Professor Elzinga, Professor McAfee did not attempt to apply Dr. Iansiti’s definition of “high-function” or otherwise try to ensure that the discount forms he reviewed pertained to “high function” as opposed to “mid-market” customers. (Tr. 2549:12-19 (P. McAfee).)

206. Professor McAfee did not use “typical” or representative case studies from Oracle, nor did he review discount forms from PeopleSoft or SAP. Professor McAfee discussed several selected “case study” vignettes of competition between Oracle and PeopleSoft. He did not attempt to ensure that these vignettes were a significant or even a representative sample of overall competition in the market. (Tr. 2541:3-9 (P. McAfee).) In fact, he conceded that the discount forms he discussed were not “typical or representative of what [he] found overall,” and that he carefully selected examples that would show clear competition between