MEMORANDUM OPINION RE GOOGLE'S MIL NO. 6 REGARDING ORACLE'S DAMAGES EXPERT JAMES MALACKOWSKI

WILLIAM ALSUP, District Judge.

INTRODUCTION

In this copyright infringement action, the accused infringer moves to exclude portions of the testimony of the copyright owner's damages/disgorgement expert, James Malackowski. The final pretrial order GRANTED IN PART AND DENIED IN PART defendant Google Inc.'s motion to exclude portions of Malackowski's testimony. Specifically, it excluded Malackowski's conclusions that Oracle was entitled to recover the entire contribution of the Android platform to the revenue streams he considered, and that no further apportionment was possible. It also excluded Malackowski's lost profits analysis to the extent it relied on a revenue forecast beyond 2011. Finally, it excluded certain testimony regarding Project Acadia.

This memorandum opinion explains the reasons for that ruling. It also imposes requirements on Malackowski's testimony regarding the causal nexus between the alleged infringement and his gross revenue estimate.

STATEMENT

Plaintiff Oracle America, Inc., offers James Malackowski as an expert to testify regarding its claim for damages due to defendant Google Inc.'s use of the declaring code and structure, sequence, and organization of 37 API packages from Oracle's copyrighted works, namely Java 2 Standard Edition Version 1.4 and Version 5.0. Malackowski, who specializes in the valuation of intangible assets including intellectual property, conducted two separate analyses in his opening report, first addressing Oracle's actual damages (lost profits issue) and then addressing Google's profits attributable to the infringement (disgorgement issue).

With regard to actual damages, Malackowski relied on a strategic forecast of growth in licensing revenue from Java Micro Edition, an alleged derivative work of the asserted copyrighted works, which Sun Microsystems, Inc. (Oracle's predecessor), and Oracle licensed for use in numerous devices, including phones. Sun created the forecast in 2007 or 2008, and it included "high," "mid," "low," and "strategic" forecasts of revenue growth from Java ME licensing in 2009 and 2010. Malackowski adopted the "strategic" growth projection, which predicted 8.3% growth over those two years. He then assumed that licensing revenue would have continued to grow at that rate through 2015 but for Google's alleged infringement and compared that projection to actual revenues. He concluded, after deducting expenses, that Oracle suffered $475 million in lost profits due to Google's alleged infringement.

Malackowski also discussed Project Acadia, Sun's (and later Oracle's) unsuccessful attempt to launch a full-stack Java-based mobile platform. He concluded that "Sun and later Oracle's actual losses attributable to the lost Acadia opportunity could be quite significant, and, potentially best measured by the apportioned Android profits attributable to the Infringed Java Copyrights" (Malackowski Rpt. ¶ 217).

In his disgorgement analysis, Malackowski examined four categories of Google's Android-related revenue: (i) advertising and search on Android devices, (ii) app sales, (iii) digital content sales, and (iv) hardware sales. He concluded that from 2008 through the end of 2015, Google earned approximately $29 billion in revenue from advertising and search on Android devices, and $11.6 billion in revenue from the remaining three sources.

In his rebuttal to Google's damages expert, Malackowski undertook a calculation of Google's costs and expenses relating to the $40.6 billion in gross revenues he addressed in his opening report and opined as to the portion of the resulting profits figures attributable to the accused infringement.

As to the gross revenue of $29 billion from advertising and search on Android devices, he subtracted $6.3 billion in "traffic acquisition costs," resulting in a gross profits calculation of $22.6 billion. He then applied a "platform contribution factor" of 35.6%, which he derived as a weighted average of the portion of profits that Google shared with competing mobile operating systems in exchange for incorporating Google's search and advertising services in those platforms. Thus, he concluded $8.0 billion in advertising and search profits from Android devices was properly attributable to the contribution of the overall Android platform, with the balance of the gross profits attributable to other elements of the business, including Google's search engine and its advertising servers (neither of which used the accused code) (Malackowski Opp. Rpt. ¶¶ 279, 280-85, Fig. 12, Exh. 7).

Malackowski noted that Google had "commingled" the infringing and non-infringing elements of Android "in a way that makes it difficult or impossible to separate out the respective contributions of each to overall profits attributable to the accused works." Thus, he stated that under the "legal theory of commingling," it was appropriate to award Oracle the entire contribution of the Android platform to the gross profits. He further opined, based on his expertise in valuing intellectual property and based on his understanding of the testimony of Oracle's technical experts, that the allegedly infringing code should be viewed as a "gating item," without which neither Android nor its contribution to the gross profits could have existed. Accordingly, Malackowski concluded that Oracle should reap the full $8.0 billion of the platform contribution to search and advertising revenue on Android devices (less deductible expenses) (id. ¶¶ 18, 272, 285).

With regard to Google's other streams of revenue from Android (namely, sales of apps, digital media, and hardware), Malackowski subtracted the costs of sales totaling $8.2 billion in generating the revenue of $11.6 billion, resulting in a calculation of $3.4 billion in gross profits. Google shared revenue with its partners (such as hardware manufacturers and app developers), which Malackowski accounted for in his calculation of the costs of sales. Accordingly, Malackowski did not apply a separate "platform contribution factor" to his calculation of profits from these sources, because his subtraction of the costs of sales already isolated the contribution of the platform to those revenue streams (id. ¶¶ 276-78, 286, Exh. 7.7).

Malackowski found, as above, that these additional sources of revenue all critically depended on the allegedly infringing code, so he concluded that Oracle was entitled to disgorge the full $3.4 billion of the platform contribution to profits from sales of apps, digital media, and hardware (less deductible expenses) (id. ¶¶ 18, 288).

Finally, after summing the $8.0 billion in gross profits from advertising and search and $3.4 billion from other revenue streams (for a total of $11.4 billion), Malackowski subtracted $2.7 billion in expenses to arrive at a final calculation of $8.8 billion in profits, which he found entirely attributable to Google's alleged infringement and thus subject to disgorgement (id. ¶ 280, Fig. 12).

Below is a table summarizing Malackowski's disgorgement calculations:

Although Oracle proposes to offer Malackowski's actual damages analysis as well as his disgorgement analysis, at oral argument on this motion, counsel for Oracle acknowledged that Malackowski's analyses overlapped. Counsel stated that to avoid double recovery, Oracle would only seek to recover the larger figure (after verdict).

Google moved to exclude portions of Malackowski's testimony. The parties fully briefed this motion, and the Court heard oral arguments on all motions in limine at five hearings each (for a total exceeding twenty hours), with two hearings dedicated exclusively to damages/disgorgement experts.

ANALYSIS

District courts are charged as the gatekeepers who evaluate the admissibility of expert opinion testimony. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141 (1999). Rule 702 of the Federal Rules of Evidence permits a court to admit expert testimony that is (1) based upon sufficient facts or data, (2) the product of reliable principles and methods, and (3) delivered by a witness who has applied the principles and methods reliably to the facts of the case. The admissibility of expert testimony turns on "whether expert testimony proffered in the case is sufficiently tied to the facts of the case that it will aid the jury in resolving a factual dispute." Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 591 (1993) (citation omitted). The proponent of expert testimony has the burden of demonstrating its admissibility. Lust v. Merrell Dow Pharms., Inc., 89 F.3d 594, 598 (9th Cir. 1996).

Google contends that Oracle failed to proffer sufficient evidence of a causal connection between Google's infringement and Malackowski's $29 billion gross revenue figure for disgorgement of advertising and search profits. Google further contends that Malackowski offered impermissible legal opinions regarding apportionment of Google's profits and further offered an improper apportionment analysis. Finally, Google argues that Malackowski used an unreliable methodology in calculating Oracle's lost profits. Google does not dispute Malackowski's qualifications to offer any of his opinions. It only disputes the legal and factual sufficiency of the bases for his opinions.

1. CAUSAL NEXUS.

Section 504(b) of Title 17 of the United States Code provides that a copyright owner is entitled to recover "any profits of the infringer that are attributable to the infringement and not taken into account in computing the actual damages." This is referred to as disgorgement. Section 504(b) contemplates a two-step burden-shifting analysis for determining the infringer's profits:

In cases involving indirect profits (that is, profits with an attenuated nexus to the infringement, rather than the direct sale of an infringing product), our court of appeals has interpreted the two-step framework of Section 504(b) to require the following analysis:

Polar Bear Productions, Inc. v. Timex Corp., 384 F.3d 700, 711 (9th Cir. 2004) (as amended). Thus, a copyright owner may not satisfy its burden under Section 504(b) by "toss[ing] up an undifferentiated gross revenue number." Ibid. Rather, it is "the duty of the copyright plaintiff to establish a causal connection between the infringement and the gross revenue reasonably associated with the infringement." Id. at 715. This may be shown by "proffer[ing] some evidence . . . [that] the infringement at least partially caused the profits that the infringer generated as a result of the infringement." Ibid. (quoting Mackie v. Reiser, 296 F.3d 909, 911 (9th Cir. 2002)).

In Polar Bear, the owner of the copyright in a video sought disgorgement of the infringer's profits from its use of that video in its marketing materials without the owner's permission. Specifically, the copyright owner sought to recover three categories of profits: (i) sales at trade shows where the video was displayed, (ii) sales due to a promotion that included a still image from the copyrighted video, and (iii) the overall enhancement of the brand prestige. Our court of appeals held that the copyright owner had satisfied its burden to show a causal nexus as to the first two categories by proffering evidence that the copyrighted work was displayed prominently in trade booths and during the promotion, that purchasers would have seen the copyrighted work, and that the accused infringer profited from the resulting sales. Polar Bear, 384 F.3d at 712-13.

On the other hand, the copyright owner failed to offer evidence to support its theory the infringement drove consumer enthusiasm and thereby permitted the accused infringer to increase its prices and generate more revenue. The "unqualified success" of the aforementioned promotions were not enough to supply the missing link. Thus, our court of appeals held the copyright owner could not recover any infringer's profits based on a theory of increased brand prestige. Id. at 713-14.

In Mackie, the owner of the copyright in a sculpture sought disgorgement of the profits from subscriptions to a symphony series after the symphony used an image of the sculpture in a montage in a multi-page promotional brochure. The copyright owner offered no concrete evidence that the accused infringement caused individuals to subscribe to the series. Absent such evidence, our court of appeals noted "we can surmise virtually endless permutations to account for an individual's decision to subscribe to the Pops series, reasons that have nothing to do with the artwork in question." Indeed, even assuming some causal link to the promotional brochure, there remained only "[r]ank speculation" of a link to the copyright owner's work. Mackie, 296 F.3d at 916.

Google argues that Oracle failed to proffer sufficient evidence of a causal nexus between Google's use of the declaring code and SSO of 37 API packages in Android and the $29 billion in gross advertising and search revenues from Android devices that Malackowski offers in his opening report. Thus, Google argues, Malackowski's testimony regarding the gross revenues will not aid the trier of fact in resolving any dispute of fact in the case and should therefore be excluded. (Google does not challenge the admissibility of Malackowski's testimony regarding the causal nexus between the accused infringement and the $11.4 in revenue from sales of Android apps, digital media, and hardware.)

In his rebuttal report, Malackowski assumed that Oracle will be able to establish that the alleged infringement served four critical purposes in the success of Android, and he found support for that assumption through his review of the reports from Oracle's technical experts and Google's internal documents. Specifically, Malackowski assumed that Google's use of the declaring code and SSO of 37 Java API packages enabled faster development of applications for the Android platform, gave Google access to the community of Java developers, increased Google's speed to market, and the lack of available alternatives. Oracle contends and Malackowski assumed that these factors facilitated Android's success and facilitated Google's control over a platform with a significant share of mobile Internet users. With that advantage, Google could avoid sharing its advertising and search revenue with the operators of other mobile platforms, and it could direct mobile Internet users to conduct searches using Google's search engine and to view advertisements delivered by Google's services, rather than lose those search and advertising opportunities to competitors like Microsoft and Apple.

Although the final pretrial order ruled that Malackowski could proffer the full $29 billion of Google's revenue from search and advertisements on Android devices in support of Oracle's disgorgement case, this order now requires Malackowski to testify as well to the "platform contribution factor" in his rebuttal report as a further condition. The $29 billion in search and advertising revenue was generated in large part by technology that operates entirely independent from the Android platform, namely, its search engine and its advertising servers. Indeed, Google generated significant revenue from those technologies on other platforms such as desktop web browsing and Apple Inc.'s iOS mobile operating system. Thus, there were "virtually endless permutations to account for an individual's decision" to conduct a Google search or to visit a site that uses one of Google's advertising services. See Mackie, 296 F.3d at 916. The relationship between that revenue and the availability of a rich app ecosystem on Android is simply too attenuated to satisfy Oracle's burden under Polar Bear.

In his rebuttal report, Malackowski offered a further analysis that, at least at this stage, may meet Oracle's burden. Specifically, Malackowski calculated a "platform contribution factor," a weighted average of the revenue that Google shared with partners in order to promote Google's services on other platforms. Malackowski's theory is that Google need not share revenue from advertising and search on Android devices, so increased demand for Android allegedly driven by Google's use of the declaring code and SSO from 37 Java API packages enabled Google to retain the entire sum of revenue without needing to share with any partners. The proffered causal connection to this reduced sum of $8.0 billion is less speculative, although even if Oracle's gross revenue proffer is limited to that figure, it remains a close call, and the trier of fact will ultimately decide if the causal nexus required by Polar Bear has been met. This may be revisited at the Rule 50 stage. Thus, this order holds that, in order to meet its duty "to establish a causal connection between the infringement and the gross revenue reasonably associated with the infringement" with regard to advertising and search revenue, Oracle must affirmatively elicit the entirety of Malackowski's calculation that arrives at the platform contribution factor of $8.0 billion during direct examination, though it remains Google's burden to prove its deductible expenses and to prove that the $8.0 billion is attributable to non-infringing factors.

It is worth noting here that Malackowski's deposition occurred after Oracle served his rebuttal report (in which his platform contribution analysis appeared), so Google will suffer little prejudice if Malackowski is required to offer opinions from his rebuttal report during direct examination. Additionally, it is true that Malackowski's platform contribution figure accounts for Google's traffic acquisition costs, although it is Google's burden of proof to offer evidence of its deductible expenses. This oddity is the result of Oracle's own overreaching in its proffer of gross revenues, and thus, it is appropriately borne by Oracle. (This is the only aspect of Malackowski's rebuttal report to be advanced to his initial testimony.)

To be clear, this order holds only that Oracle has proffered sufficient evidence of a causal nexus to present its case (subject to the above limitations). Oracle must still establish a causal connection as a threshold matter. To repeat, this ruling is without prejudice to a Rule 50 motion directed at this issue after the presentation of evidence.

2. APPORTIONMENT.

If Oracle carries its burden of demonstrating "a causal connection between the infringement and the gross revenue reasonably associated with the infringement," Section 504(b) shifts the burden to Google to prove its deductible expenses and to show that some of the profits at issue were attributable to non-infringing factors.

As described above, in his rebuttal report, Malackowski undertook his own calculation of Google's deductible expenses and isolated the share of the remaining profits associated with the Android platform. As to advertising and search profits, he concluded that Android accounted for 35.6% of the revenue for a total of $8.0 billion. (This order requires Malackowski to advance this calculation to his initial testimony.) As to sales of apps, digital media, and hardware, he concluded that the deduction of the cost of sales accounted for the contribution of any non-Android factors, and thus concludes Android accounted for $3.6 billion in profits from those revenue streams.

Malackowski then undertook an "apportionment" analysis in which he concluded that zero dollars of the respective contribution of the Android platform could be attributed to any factor other than the declaring code and SSO of 37 API packages from Java and thus opined that Oracle should recover the full $8.8 billion in profits. In reaching this conclusion, Malackowski referred to the "legal theory of commingling," which he understood to mean that "when the infringer has mixed the infringing and non-infringing attributes in a way that makes it difficult or impossible to separate out the respective contributions of each to overall profits attributable to the accused work" the copyright owner it entitled to recover the entire amount of profits attributable to the overall work (Malackowski Opp. Rpt. ¶ 18). Malackowski further concluded that the collection of Java API packages was "properly viewed as a `gating item' to the Android Platform" (id. Rpt. ¶ 272). That is, Malackowski concluded that the allegedly infringing code was the sine qua non of Android. Thus, he apportioned 100% of the profits attributable to the platform to the declaring code and SSO of 37 APIs, and opined that Oracle could disgorge the full $8.8 billion.

Where, as here, "an infringer's profits are not entirely due to the infringement, and the evidence suggests some division which may rationally be used as a springboard it is the duty of the court to make some apportionment." Cream Records, Inc. v. Jos. Schlitz Brewing Co., 754 F.2d 826, 828-29 (9th Cir. 1985) (citation omitted). This amount need not be calculated to mathematic exactness, "only a reasonable approximation" is necessary. Id. at 829 (quoting Sheldon v. Metro-Goldwyn Pictures Corp., 309 U.S. 390, 408 (1940)).

It is Google's burden to apportion the disgorgement figure advanced by Oracle. Nevertheless, to the extent Malackowski engages in the exercise of apportionment, he must offer a methodology tied to the purpose of making a reasonable approximation of the profits attributable to the allegedly infringing elements of Android. By nullifying the contribution of critical components such as the Linux kernel, the Dalvik virtual machine, the other core libraries, and the implementing code for the 37 API packages, Malackowski will simply invite the trier of fact to adopt an apportionment methodology that is contrary to the facts, and as such, his apportionment analysis is inadmissible.

Again, it is Google's burden to provide a method of apportionment, so Oracle need not offer Malackowski's apportionment opinion at all (except, as discussed above, that it must advance his full calculation of platform contribution to advertising and search profits as his initial proffer). Moreover, it may rest on Malackowski's $8.8 billion (or even the $11.6 billion before deducting general expenses) in hopes that Google fails to carry its burden of demonstrating that further apportionment or expense deduction is appropriate. But Oracle may not offer Malackowski's opinion that simply ignores the numerous critical non-infringing elements of Android to which some of the profits should be apportioned and thus eviscerates the duty to apportion where the evidence plainly shows that the profits are partially attributable to non-infringing elements. That would risk the "certainly unjust course of giving the plaintiffs everything. . . ." Cream Records, 754 F.2d at 829 (citation omitted).

To be clear, Malackowski's calculation of the "platform contribution factor" from advertising and sales must be advanced to his direct examination; however, as a procedural matter, this order holds that Oracle must present Malackowski's specific critiques of Google's own apportionment methodology and expense deductions (as well as his own further expense deductions) after Google has offered its apportionment analysis. (Yes, this means Malackowski will have to appear twice.) This arrangement will assist in comprehending the evidence within the framework of the burden-shifting analysis.

3. LOST PROFITS.

Google also contends that Malackowski's calculations of Oracle's lost licensing revenue from Java ME, an alleged derivative work of the copyrighted works, is too speculative. Malackowski opined that Oracle lost profits totaling $475 million due to Google's use of the declaring code and SSO of 37 API packages in Android. Google contends that Oracle has failed to establish the "necessary, immediate and direct causal connection" between Google's alleged infringement and Google's diminished revenue. Google further contends that even if the requisite causal connection can be established, Malackowski's calculation of the lost profits is based on an unreliable forecast made by Sun.

As to the causal connection, a reasonable jury could find a direct causal connection between Google's alleged infringement and Oracle's loss of licensing revenues from Java ME. Malackowski details several licensing deals that fell through or were renewed at a substantially lower value, and he notes that many of Oracle's former, current, and prospective partners subsequently released products that implemented Android. Malackowski's calculation of the lost profits suffers from several serious flaws and must therefore be limited. As above, this holding is without prejudice to a Rule 50 motion after the evidence has been presented.

Malackowski did not calculate lost profits based on the value of the actual deals that fell through or that suffered from significant pricing pressure. Rather, Malackowski calculated lost profits based on a forecast of licensing revenue for Java ME that Sun produced in 2007-08. The "strategic" component of the forecast projected licensing revenue growth of 8.3% from 2009 through 2010, and Malackowski assumed that absent the infringement Oracle would have realized the same rate of growth in its licensing revenues through 2015 — five years beyond the scope of the projection. He calculated the difference between the projected growth and actual results (after the deduction of expenses) as $475 million.

Malackowski supposedly vetted the projection by discussing it with Oracle personnel knowledgeable about Sun's business strategy at the time of the forecast, by comparing it to actual licensing revenues from previous years, and by considering another (more optimistic) forecast made later in the year (Malackowski Rpt. ¶¶ 189, 196-97). Malackowski did not, however, perform any analysis regarding whether Oracle underperformed relative to the forecast due to market shifts or events unrelated to Google's alleged infringement. Nor did he offer any analysis to support extending the growth projections five years beyond the end of the forecast.

A prior order held that Oracle could offer expert testimony that extended the projected growth through 2011 and that Google could test the propriety of conclusions based on the forecast through cross-examination (Dkt. No. 685 at 5-7). This order holds that the forecast may be used to that extent, but it cannot be used to project revenue growth any further out than 2011 because Malackowski offered no analysis regarding whether the forecast (which did not itself purport to project so far into the future) would be a reliable basis for making such projections despite changes in market conditions, inter alia. Accordingly, Malackowski may offer conclusions about the amount of Oracle's alleged lost profits only through 2011. He may note that his calculation omits four years of the damages period, but he may not opine that it would be proper for the jury to assume that same rate of growth through 2015.

Google also moved to exclude Malackowski's conclusion that the actual damages attributable to the failure of Project Acadia is "potentially best measured by the apportioned Android profits attributable to" the alleged infringement, because Malackowski failed to evaluate whether Acadia could have been as successful and profitable as Android but for the alleged infringement. Oracle responds that Malackowski did not an attempt to quantify the lost Acadia opportunity, he merely sought to explain why his lost profits calculation was conservative.

Malackowski may offer testimony that his lost profits calculation did not account for the profits Oracle could have realized from Project Acadia, but he may not in any way suggest that the proper measurement of those lost profits should be based on Google's apportioned profits. Such testimony would invite the jury to expand the scope of its lost-profits finding beyond that which can be supported by the evidence.

CONCLUSION

For the reasons stated above, Google's motion to exclude Malackowski's testimony is GRANTED IN PART AND DENIED IN PART. Oracle may present its causal nexus case, however, it must also affirmatively elicit Malackowski's calculation of the platform contribution to advertising and search profits (resulting in an initial proffer of $8.0 billion). The viability of Oracle's causal nexus case also remains subject to a Rule 50 motion after the evidence. Malackowski may not offer his opinion that no further apportionment of the platform contribution in any category of revenues is possible or appropriate. Malackowski may not rely on the Sun 2007-08 forecast to calculate Oracle's lost profits beyond 2011, and he must adjust his totals to reflect that limitation. Finally, Malackowski may not offer testimony that Oracle's lost profits from the failure of Project Acadia can be measured in any way based on Google's profits attributable to the alleged infringement.

IT IS SO ORDERED.