Oracle America, Inc. v. Google Inc.

Filing 227

Letter from Steven C. Holtzman and Scott T. Weingaertner re Non-Mobile Data. (Holtzman, Steven) (Filed on 7/21/2011)

B O I E S ,

S C H I L L E R

&

F L E X N E R

L L P

1999 HARRISON STREET* SUITE 900* OAKLAND, CA 94612* 510-874-1000 FAX [phone]

July 21, 2011
VIA ECF
The Honorable Donna M. Ryu
United States Magistrate Judge
Northern District of California
1301 Clay Street, Courtroom 4, 3rd floor
Oakland, California 94612
Re:

Oracle America, Inc. v. Google Inc., No. 3:10-CV-03561-WHA (N.D. Cal.)
Joint Letter Regarding Oracle’s Request for Google’s Production of Non-Mobile Data
and Projections

Dear Judge Ryu:
Oracle has requested Google’s production of non-mobile data and projections responsive
to Request for Production (“RFP”) Nos. 168-174. The parties met-and-conferred telephonically
on July 19, 2011, but were unable to resolve the issue. Accordingly, Oracle respectfully requests
the Court’s assistance in compelling Google’s response to the aforementioned document
requests.
Oracle’s Statement
As we have repeatedly discussed with counsel for Google, and as described in the expert
report of Oracle’s damages expert, certain non-mobile data – including specific categories of
data identified by Oracle in the course of meeting and conferring – are highly relevant to the
assessment of damages in this case. Despite our explanations, and without any substantive, nonconclusory rationale, Google has refused to produce any non-mobile data.
Over the course of the last five months, Oracle has asked for documents concerning nonmobile data and projections through a series of requests, including RFP No. 128 (seeking
“Google’s website generated revenues . . . that relate to mobile revenue or Android revenue,
profitability or costs”) and RFP No. 160 (“documents reflecting or relating to any impact of
Android on Google’s non-mobile business, sales, revenue, profitability, market share,
competition, or competitive advantage” (emphasis added)). During and after numerous meetand-confers that began on March 7, 2011,1 Google insisted that non-mobile data and projections
were not responsive to Oracle’s original document requests. Accordingly and to leave no doubt
1

The extensive meet-and-confer process was detailed in Oracle’s May 2, 2011 letter to Google.
Oracle further corresponded on May 12, 16, and 31, 2011 and the parties engaged further on
these issues in senior counsel meet-and-confers on or about May 11, May 23, June 1, June 10,
June 27, and July 1, 2011.
www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 2 of 8
as to responsiveness, Oracle served RFP Nos. 168-174 on May 2, 2011 (nearly three months
before the discovery cutoff) specifically seeking non-mobile data and projections. Google knew
throughout the process that Oracle sought such data and projections in order to complete its
damages expert report, which Judge Alsup required be submitted on May 20, 2011, seventy days
before the end of fact discovery.
In a response dated June 6, 2011, Google referred Oracle exclusively to mobile data and
SEC filings in response to our requests for non-mobile data related to ad revenue (RFP No. 168),
traffic acquisition costs (RFP No. 169), ad revenue projections (RFP No. 170), click-through
rates and similar search related metrics (RFP No. 171), and search volume (RFP No. 172). In a
letter dated June 8, 2011, Oracle explained that these documents do not address Oracle’s requests
for non-mobile data in RFP Nos. 170-172 and fail to provide non-mobile specific data on the
regular intervals sought by RFP Nos. 168-169. On June 21, 2011, Google responded that the
requests for non-mobile data were “late,” “overly broad,” and “irrelevant.”
The relevance of the limited non-mobile data sought by Oracle is made clear in the expert
report of Dr. Iain Cockburn dated May 20, 2011 (“Cockburn Report”). In the report, Dr.
Cockburn explains that Android is part of a network market, in which adoption of Android feeds
the growth of Google’s brand generally and non-mobile business specifically, particularly as
mobile (and Android) usage becomes increasingly significant. Cockburn Report ¶¶ 92-104.
Google’s refusal to produce non-mobile data has precluded Dr. Cockburn from quantifying these
effects, which likely represent very significant benefits to Google of the patent and copyright
infringement at issue in this case. Id. at ¶¶ 227-29. Moreover, Google documents from before
the infringement began make clear that Google specifically intended Android – for which it
relied on Oracle’s intellectual property – to provide precisely this strategic benefit to Google’s
core non-mobile business. For example, Google’s 2007 “Mobile” presentation, explains that
mobile “increases ROI [Return on Investment] . . . [a]cross entire universe: Web, PC, phone”.
GOOGLE-22-00226487 at 519.
The data responsive to RFP Nos. 168-69 and 171-72 (revenue, costs, ad metrics, and
search volume) would enable Dr. Cockburn to quantify Android’s impact on Google’s overall
growth. To be economically reliable, that time series requires data on a no greater than monthly
basis – data that Google possesses but refuses to produce. The non-mobile revenue projections
responsive to RFP No. 170 inform the magnitude of these effects that Google anticipated at the
time of the hypothetical negotiation that provides the basis for patent and copyright infringement
damages in this case. See Cockburn Report ¶¶ 121.
Google has had since January 25, 2011 to produce this important data, but has done
nothing and then declared Oracle’s requests untimely five months after they were first served.
To justify its delay, Google has asserted that non-mobile data is unrelated to Android and
therefore non-responsive to Oracle’s requests because (1) Professor Cockburn purportedly failed
to apportion damages and (2) network effects are unrelated to the hypothetical negotiation to set
a reasonable royalty for Google’s infringement. As to apportionment, Oracle explained in
Opposition to Google’s Daubert Motion that Google improperly assumes that a damages expert
must weigh in with technical expertise before the technical expert has delivered his report. See

www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 3 of 8
Pl. Opp. Br. at 13-16. That is inconsistent with Judge Alsup’s case and expert discovery
schedule. Further, as explained more fully in Oracle’s Opposition, Professor Cockburn did, in
fact, carefully consider the value of the technology at issue. Id. Google’s disagreement that the
patents and copyrights in suit drive demand for Android is simply another factual dispute that
must be resolved by the jury.
As to network effects, Google’s contention of irrelevance is belied not only by Dr.
Cockburn’s report regarding network effects, but also Google’s own documents, as noted above.
Google’s unjustified and absolute refusal to provide any non-mobile data is particularly
prejudicial to Oracle because Professor Cockburn’s efforts to date to obtain third-party analyst
data that might be able to replace Google’s own data have been stymied by the third party’s
refusal to even sell that data for use in this case because of concerns about the analyst’s
dependence on and relationship with Google.
Google admits that it possesses responsive data, but objects that it “maintains thousands
of internal dashboards and to manually export data from each one would certainly impose
significant burdens on Google.” However, Oracle seeks not all of the dashboards, but only those
responsive to the limited categories listed in its request for productions. Google has refused
Oracle’s repeated requests for a list of dashboards that would permit Oracle to further narrow the
request for data.
Oracle has narrowed its request to non-mobile projections produced from Q1 2008
through Q1 2009. Though it lacks a list of dashboards, Oracle proffered the following list of
non-mobile datasets it believes to exist after examination of comparable mobile dashboards:












Total search volume, by operating system (and by keyword if available)
Number of banner ads served by DoubleClick
Click-through-rates for display ads (by keyword)
Click-through-rates for search ads (by keyword)
Market share of search volume, by search provider
Average cost-per-click paid by advertisers for search ads
Average cost-per-impression paid by advertisers for display ads
Revenue run rates for search ads (by keyword)
Revenue run rates for display ads (by keyword)
Web content indexed by Google
Number of apps for the desktop (e.g., gmail)

For each metric, we would prefer, if possible, daily frequency (or finest frequency
available), from 2004 to the present, by country.
Google maintains its refusal to produce and responsive data.
For these reasons, Oracle respectfully requests that the Court compel production of the
data responsive to RFP Nos. 168-174.

www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 4 of 8

Google’s Statement
Oracle’s demand for seven years worth of Google’s non-mobile data is an improper
attempt to saddle Google with a burden of substantial irrelevant discovery designed to prop up a
fatally-flawed damages report. This case concerns claims of infringement based on limited
portions of a software platform used exclusively by mobile products. Specifically, Oracle’s
infringement allegations are limited to claims of patent and copyright infringement relating to
certain features of the Android software platform. See, e.g., Amended Complaint, ¶ 13. The
accused Android platform is used solely for mobile products such as smartphones and tablets.
Id. at ¶ 12 (describing Android as “an operating system and software development platform for
cellular telephones and other mobile devices”). Yet, after having demanded and received
extensive discovery relating to Android and related mobile data, Oracle now demands much
more by seeking discovery into aspects of Google’s business that have nothing to do with mobile
devices.
Oracle supports this demand only by pointing to the theories espoused by its damages
expert in his recent report (the “Cockburn Report”). Yet, as described below and in Google’s
Brief in Support of Daubert Motion (Docket No. 171 at 9-10), the theories in the Cockburn
Report are entirely untethered to law or fact and seek to artificially inflate his damages numbers
with advertising dollars earned outside the context of mobile devices. Google respectfully asks
the Court to reject Oracle’s request to compel the production of highly sensitive non-mobile data
in response to Oracle’s overly broad, unduly burdensome and untimely RFP Nos. 168-174.
1. Non-mobile data is irrelevant to this case.
The requested non-mobile data is irrelevant to this case, including any legally cognizable
damages theory.
Oracle’s infringement claims are based on specific, identified aspects of the Android
mobile device platform. Amended Complaint, ¶¶ 12-13. Oracle bases its claims on seven
patents and two copyrights. Id. at ¶¶ 16-47. The patents relate to specific, discernable features
of Android, including security and optimizations for a virtual machine. See, e.g., U.S. Patent No.
6,192,476 at 21-25; U.S. Patent No. 6,125,447 at 23-25; U.S. Patent No. 6,910,205 at 1:10-13;
U.S. Patent No. 7,426,720 at 1:8-11; U.S. Patent No. 6,061,520 at 4-6; U.S. Patent No. RE38,104
at 15-19. Oracle’s patent infringement contentions allege that only portions of Android code
referring to security and portions of Android code relating to the Dalvik virtual machine infringe
the asserted patents. Oracle’s copyright infringement claims are similarly based on relatively
small portions of the Android source code. Amended Complaint, Exhibits I & J.
Oracle nonetheless seeks to expand the scope of discovery far beyond issues relating to
the intellectual property or technology at issue. Oracle’s RFP Nos. 168-174 seek data relating to
Google’s non-mobile advertising revenues and revenue projections (RFP Nos. 168, 170), nonmobile traffic acquisition costs (RFP No. 169), non-mobile costs per click and costs per milli
(RFP No. 171), total search volume (RFP No. 172), total market share (RFP No. 173), and total
number of advertisers and ads served (RFP No. 174). Neither Google’s search technology nor its
advertising business is an accused product, and the value of the accused Android platform
www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 5 of 8
software and of Google’s search technology and advertising business are entirely separate. The
Android software provides for certain phone functionality, whether or not the user is viewing ads
or running searches, and Google’s ads and search results are viewable on any software and are
not uniquely enabled by Android.
Oracle has articulated no reasonable justification for its far-fetched and over-reaching
demands. Oracle explains that its damages expert, Dr. Iain Cockburn, requires Google’s nonmobile data to quantify Android’s impact on Google’s overall growth, but such an analysis
would only further compound the serious flaws underlying the Cockburn Report. As just one
example of its unreliability, described in further detail in Google Inc.’s Brief in Support of
Daubert Motion (Dkt. No. 171), the Cockburn Report does not tie Cockburn’s royalty
calculation to the Android platform software. It does not value the individual patents or
copyrights, the technology embodied in the asserted copyrights or claims of the patents-in-suit,
or the relative contribution of that technology to the overall Android platform, and it does not
apportion the value of patented and non-patented features. Instead, Cockburn improperly
includes in his royalty calculation all of Google’s worldwide advertising revenue from Android
phones. There is simply no credible and sufficient economic proof that the patented invention or
copyrighted works drove demand for Android, and Cockburn offers none. See Cornell Univ. v.
Hewlett-Packard Co., 609 F.Supp. 2d 279 (N.D.N.Y. 2009) (Rader, J.).
Yet, Cockburn’s already hyper-inflated multi-billion dollar damages calculation is not
enough to satisfy Oracle, which seeks to increase the damages calculation by laying out, at
considerable effort, speculative theories as to direct and indirect revenue that Google purportedly
attained or could have anticipated with respect to Android. Having improperly assumed without
any basis that the asserted patents and copyrights account for the entire value of Android, Oracle
should not be permitted to exacerbate the problems with the Cockburn Report by incorporating
into its damages analysis data that bears no reasonable relation to the issues in this case.
2.

Google has produced substantial information responsive to Oracle’s
demands.

Even though the information is not relevant, Google has already produced, in response to
these and other RFPs, the most relevant information in its possession, namely mobile and
Android P&L statements, monthly mobile revenue figures, and dozens of Android- and mobilerelated datasets. Indeed, Google has already exceeded its obligations in this regard. Although
Oracle has only accused a portion of the Android platform of infringing its intellectual property,
Google has produced financial information and data related to not just Android devices, but also
mobile devices that do not even run Android. Moreover, Oracle’s assertion that Google has
refused to produce any non-mobile data is incorrect. In response to RFP Nos. 168-169, Google
produced high-level financial data in the form of Form 10-K and Form 10-Q SEC filings. To the
extent Oracle believes it is necessary to evaluate Google’s non-mobile revenues and costs, it can
derive such information simply by subtracting the mobile data that Google has already produced
from the financial data for the entire company as disclosed in Google’s public SEC filings.
Oracle’s vague reference to documents that purportedly show Google’s hopes and
aspirations for Android and to Cockburn’s speculative theories about so-called network effects
do not satisfy any legal standard for expanding its damages to non-mobile revenues or warrant
the production of highly sensitive non-mobile data, particularly when Oracle already has access

www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 6 of 8
to far more relevant Android and mobile data. See, e.g., ResQNet.com, Inc. v. Lansa, Inc., 594
F.3d 860, 869 (Fed. Cir. 2010) (“[T]he trial court must carefully tie proof of damages to the
claimed invention’s footprint in the market place.” (emphasis added)); DSU Medical Corp. v.
JMS Co., Ltd., 471 F.3d 1293, 1309 (Fed. Cir. 2006) (“To prevent the hypothetical from lapsing
into pure speculation, this court requires sound economic proof of the nature of the market.”
(citation omitted)). Perhaps realizing that the information already produced by Google cannot
possibly support its excessive damages calculations, Oracle has resorted to fishing for a wide
range of data far removed from the intellectual property or accused technology at issue in this
case. It should not be permitted to do so.
3.

Oracle’s demands for non-mobile data are not only irrelevant but unduly
burdensome.

RFP Nos. 168-174 are not only overbroad and irrelevant. They would also impose
significant burdens on Google for the production of extremely sensitive, highly confidential
information. As explained to Oracle repeatedly throughout the meet-and-confer process, Google
maintains thousands of dashboards from which the requested data, to the extent it exists, can be
exported. Searching through these dashboards to find the non-mobile data that Oracle might be
seeking would consume enormous amounts of time and would subject Google to a tremendous
drain on its internal resources. These burdens are especially pronounced in view of Oracle’s
failure to offer a reasonable or compelling need for irrelevant non-mobile data and in view of the
specificity of the non-mobile data that Oracle seeks (e.g., “by keyword”).
Oracle’s supposedly “narrowed” requests offer little assistance. The list of non-mobile
datasets it seeks remains just as irrelevant and burdensome as the non-mobile data sought by the
original RFPs. It remains unclear how requesting worldwide non-mobile data broken out “by
country” and “by keyword” narrows the issues or reduces the burdens at all – or how this level of
specificity is even relevant. Nor does Google understand what Oracle means by “Web content
indexed by Google” and “Number of apps for the desktop (e.g., gmail)” or how they relate to any
of the RFPs. According to Oracle, these requests fall within the scope of RFP Nos. 160 and 173
because they “indicate user activity” and those RFPs “seek documents showing Google’s market
share.” Google disagrees. As noted below, RFP No. 160 is tied to Android. And while RFP No.
173 may relate to market share, it is not clear how these requests relate to market share.
Furthermore, although Oracle provided a chart with examples of specific mobile
dashboards it believes are “comparable” to the requested non-mobile datasets, the chart provides
little clarity or guidance. Many of the identified mobile datasets do not break out data “by
country.” For several categories (e.g., “Web content indexed by Google,” “Average cost-perimpression paid by advertisers for display ads,” “Market share of search volume, by search
provider”), Oracle provided no examples of “comparable” mobile datasets at all. For other
categories (e.g., “Click-through-rates for display ads,” “Click-through-rates for search ads”), it is
not clear how the identified mobile datasets include data broken out “by keyword.” For the
“Total search volume, by operating system” category, Oracle acknowledges that it has already
received non-mobile data.

www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 7 of 8
4.

Oracle’s demands for extensive non-mobile data are untimely in any event.

Oracle should not be rewarded for its untimely service of RFP Nos. 168-174. The Court,
in paragraph nine of its November 18, 2010 Case Management Order, required that “[a]ny party
seeking monetary damages must serve all of its damages report SEVENTY CALENDAR DAYS
before other non-damages opening reports are due,” i.e., May 20, 2011. (Dkt. No. 56.) Oracle
states that the requested non-mobile data are highly relevant to its assessment of damages, yet
Oracle waited until May 2, 2011, less than one month before the deadline for serving its
damages report, to serve RFP Nos. 168-174. Realizing it had waited too long to serve these
requests, Oracle attempted to unilaterally impose its own set of discovery rules on Google,
demanding that Google respond to the requests within two days and produce the requested data
within four days.
Oracle claims that the requested non-mobile data falls within the scope of other RFPs
served earlier in discovery (e.g., RFP Nos. 112, 120, 128, 129, 130 and 160), and that it served
RFP Nos. 168-174 only in an abundance of caution “to leave no doubt” that it was seeking nonmobile data. A fair reading of the earlier RFPs shows that they are all clearly limited to Android
and/or mobile information. RFP No. 112 requests “revenue derived from the sale, use, licensing
or other activities relating to Android by any third party.” RFP No. 120 requests documents
relating to “any link, connection or relationship between Android and any actual and projected
Google or third party revenues or profits.” RFP Nos. 128-130 request “Google’s websites
generated revenues,” “Google’s partner sites generated revenues,” and “Google’s other
revenues,” respectively, “that relate to mobile revenue or Android revenue, profitability or
costs.” Each of these requests is limited to mobile- and/or Android-related material. And
although RFP No. 160 refers to non-mobile information, the request is limited to “any impact of
Android on Google’s non-mobile business . . . .” Oracle’s late, overly broad demand for
Google’s raw non-mobile financial data completely outside the context of any issues in this case
falls well beyond the scope of these requests.
Oracle did not express an interest in non-mobile data until April 21, 2011, when it sent
Google a letter demanding the production of at least 50 separate sets of data, many of which
addressed information unrelated to Android or mobile revenue and none of which Oracle
correlated to any particular discovery request. Oracle’s decisions to wait until April 21, 2011 to
first raise the issue of non-mobile data; to wait until May 2, 2011 to serve RFPs explicitly
directed at non-mobile data; and to wait until the end of June to raise this issue with the Court
undermine its claim that such data is highly relevant to its damages analysis.
Faced with a Daubert motion that exposes the severely flawed methodology of its
damages expert and its baseless claim for billions of dollars in damages, Oracle only now – over
one month after it served its damages report – seeks the Court’s assistance in obtaining
additional, irrelevant non-mobile data from Google. The Court should reject Oracle’s improper
attempt to use untimely RFPs as an excuse to supplement a defective damages report and
circumvent the Court’s clear schedule regarding damages reports.
***

www.bsfllp.com

Honorable Donna M. Ryu
July 21, 2011
Page 8 of 8
Google has already far exceeded its discovery obligations by producing mobile data and
high-level financial information, in addition to Android financial data. Oracle has provided no
reasonable justification for its expansive and overreaching demands for highly confidential,
commercially sensitive non-mobile data. Accordingly, Google respectfully requests that the
Court summarily reject Oracle’s unreasonable demand that Google produce additional irrelevant
non-mobile data in response to RFP Nos. 168-174.
Respectfully submitted,
BOIES, SCHILLER AND FLEXNER LLP

KING & SPALDING LLP

By: /s/ Steven C. Holtzman
Steven C. Holtzman

By: /s/ Scott T. Weingaertner
Scott T. Weingaertner

Attorneys for Plaintiff
ORACLE AMERICA, INC

Attorneys for Defendant
GOOGLE INC.
ATTESTATION OF FILER

I, Steven C. Holtzman, have obtained Mr. Scott T. Weingaertner s concurrence to file
this document on his behalf.
Dated: July 21, 2011

BOIES, SCHILLER & FLEXNER LLP
By: /s/ Steven C. Holtzman
Steven C. Holtzman
Attorneys for Plaintiff
ORACLE AMERICA, INC.

www.bsfllp.com