Oracle America, Inc. et al v. Procore Technologies, Inc. et al

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
UNITED STATES DISTRICT COURT
NORTHERN DISTRICT OF CALIFORNIA
ORACLE AMERICA, INC., et al.,
Plaintiffs,
v.
PROCORE TECHNOLOGIES, INC., et al.,
Defendants.
Case No. 24-cv-07457-JST
ORDER DENYING MOTIONS TO
STRIKE OR EXCLUDE
Re: ECF Nos. 289, 340, 344
Before the Court are three motions by Defendants Procore Technologies, Inc. and Procore
Payment Services, Inc. to strike or exclude portions of the expert reports of Melissa Bennis and
Andrew Crain. ECF Nos. 289, 340, 344. The Court will deny all three motions.
I. BACKGROUND
Plaintiffs Oracle America, Inc., Oracle International Corporation, and Textura Corporation
(“Oracle”) filed the complaint on October 25, 2024. ECF No. 1. Oracle alleges a single count of
trade secret misappropriation under the Defend Trade Secrets Act (“DTSA”), 18 U.S.C.
§§ 1836(b), 1839 et seq.1 Id. ¶¶ 73–92. In short, Oracle alleges that Procore misappropriated its
trade secrets related to Textura Payment Management (“TPM”), construction project payment
software, to develop a competitor product called Procore Pay. Id. ¶¶ 47, 66, 74. The alleged
misappropriation largely stems from an Oracle employee named Mark Mariano’s move to Procore
in October and November of 2021. ECF No. 349-34 ¶¶ 51, 639.
Oracle’s Fifth Amended Trade Secret Disclosures, served on March 16, 2026, disclosed
more than 450 alleged trade secrets. See generally ECF No. 345-3. Two weeks later, Oracle
1 A second count for breach of contract was originally alleged against an individual defendant
named Mark Mariano, but he has since been dismissed. ECF No. 1 ¶¶ 93–105; ECF No. 165.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 1 of 20

2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
II. JURISDICTION
The Court has jurisdiction under 28 U.S.C. § 1331.
III. LEGAL STANDARD
A. Daubert
Under Federal Rule of Evidence 702, the District Court acts as a “gatekeep[er],”
preventing expert testimony from reaching the jury that is irrelevant or does not result from the
application of reliable methodologies or theories to the facts of the case. Kumho Tire Co. v.
Carmichael, 526 U.S. 137, 147–48 (1999) (citing Daubert v. Merrell Dow Pharms., Inc., 509 U.S.
579, 589 (1993)). To survive a Daubert challenge, “[t]he expert’s testimony must be ‘based upon
sufficient facts or data,’ ‘the product of reliable principles and methods,’ and the expert must
‘appl[y] the principles and methods reliably to the facts of the case.’” Stilwell v. Smith & Nephew,
Inc., 482 F.3d 1187, 1191–92 (9th Cir. 2007) (quoting Fed. R. Evid. 702). “Expert opinion
testimony is . . . reliable if the knowledge underlying it has a reliable basis in the knowledge and
experience of the relevant discipline.” City of Pomona v. SQM N. Am. Corp., 750 F.3d 1036, 1044
(9th Cir. 2014) (citation omitted). The reliability of non-scientific testimony may “depend[]
heavily on the knowledge and experience of the expert, rather than the methodology or theory
behind it.” Hangarter v. Provident Life & Acc. Ins. Co., 373 F.3d 998, 1017 (9th Cir. 2004)
(emphasis removed).
“Shaky but admissible evidence is to be attacked by cross examination, contrary evidence,
and attention to the burden of proof, not exclusion.” Primiano v. Cook, 598 F.3d 558, 564 (9th
served a pared down list of 110 alleged trade secrets. See ECF No. 345-4. Oracle then further
narrowed to 70 and then 34 trade secrets on June 1 and June 5, respectively. See ECF Nos. 331,
345-5, 345-6.
Oracle’s alleged trade secrets are organized into six categories: (A) TPM Solution (51, now
10 trade secrets); (B) Oracle Enterprise Resource Planning (“ERP”) Integrations (5, now 3 trade
secrets); (C) Third-Party ERP Integrations (39, now 10 trade secrets); (D) Cross Platform
Integrations (9, now 6 trade secrets); (E) Customer Information (3 trade secrets); and (F) Internal
Business Information (3, now 2 trade secrets).
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 2 of 20

3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
B. Rules 26 and 37(c)
Under Federal Rule of Civil Procedure 26(a), “a party must, without awaiting a discovery
request, provide to the other parties . . . a copy—or a description by category and location—of all
documents, electronically stored information, and tangible things that the disclosing party has in
its possession, custody, or control and may use to support its claims or defenses.” Fed. R. Civ. P.
26(a)(1)(A)(ii). Expert reports must disclose, among other things, “a complete statement of all
opinions . . . and the basis and reasons for them” and “the facts or data considered by the witness
in forming them.” Fed. R. Civ. P. 26(a)(2)(B)(i)–(ii). Rule 26(e) imposes a continuing duty to
supplement those disclosures “in a timely manner.” Fed. R. Civ. P. 26(e)(1)–(2).
Federal Rule of Civil Procedure 37(c) provides that “[i]f a party fails to provide
information or identify a witness as required by Rule 26(a) or (e), the party is not allowed to use
that information or witness to supply evidence on a motion, at a hearing, or at a trial, unless the
failure was substantially justified or harmless.” Fed. R. Civ. P. 37(c)(1). “Rule 37 ‘gives teeth’ to
Rule 26’s disclosure requirements by forbidding the use at trial of any information that is not
properly disclosed.” Goodman v. Staples The Off. Superstore, LLC, 644 F.3d 817, 827 (9th Cir.
2011) (quoting Yeti by Molly, Ltd. v. Deckers Outdoor Corp., 259 F.3d 1101, 1106 (9th Cir.
2001)). It is “a self-executing, automatic sanction designed to provide a strong inducement for
disclosure.” Id. (internal quotation marks and citations omitted).
“[T]he burden is on the party facing sanctions to prove harmlessness.” Yeti by Molly, 259
F.3d at 1107. In assessing harmlessness, courts consider: “(1) prejudice or surprise to the party
against whom the evidence is offered; (2) the ability of that party to cure the prejudice; (3) the
Cir. 2010), as amended (Apr. 27, 2010). “Simply put, ‘[t]he district court is not tasked with
deciding whether the expert is right or wrong, just whether [their] testimony has substance such
that it would be helpful to a jury.’” Canty v. DePuy Orthopaedics Inc., No. 14-CV-05407-JSW,
2024 WL 2852213, at *2 (N.D. Cal. June 5, 2024) (quoting Alaska Rent-A-Car, Inc. v. Avis
Budget Grp., Inc., 738 F.3d 960, 969–70 (9th Cir. 2013)).
It is Oracle’s burden to establish that its expert testimony is admissible under FRE 702.
See Lust v. Merrell Dow Pharms., Inc., 89 F.3d 594, 598 (9th Cir. 1996).
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 3 of 20

4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
IV. DISCUSSION
A. Motion to Exclude the Expert Opinion of Melissa A. Bennis
Melissa Bennis is Oracle’s damages expert. The DTSA “authorizes three separate
measures of damages: (1) ‘damages for actual loss caused by the misappropriation of the trade
secret’; (2) ‘damages for any unjust enrichment caused by the misappropriation of the trade secret
that is not addressed in computing damages for actual loss;’ or (3) ‘in lieu of damages measured
by any other methods, the damages caused by the misappropriation measured by imposition of
liability for a reasonable royalty for the misappropriator’s unauthorized disclosure or use of the
trade secret.’” Sonoma Pharms., Inc. v. Collidion Inc., No. 17-CV-01459-EDL, 2018 WL
3398940, at *6 (N.D. Cal. June 1, 2018) (citing 18 U.S.C. § 1836(b)(3)(B)).
Procore seeks to exclude Bennis’s report under Daubert on four grounds: (1) that it fails to
apportion damages by individual asserted trade secret, rather than by trade secret category; (2) that
its price erosion opinion lacks any independent basis; (3) that its reasonable royalty opinion is
unreliable; and (4) that the asserted damages period lacks any reliable basis. ECF No. 344. None
of these objections justifies exclusion.
1. Failure to Apportion Value or Damages Among the Asserted Trade
Secrets
A jury verdict requires a reasonable basis in the evidence. See Flores v. City of
Westminster, 873 F.3d 739, 757 (9th Cir. 2017). As such, should the jury conclude that fewer than
all asserted trade secrets have been misappropriated, it must be able to calculate the damages
attributable only to the misappropriated ATSs. See Trinseo Eur. GmbH v. Kellogg Brown & Root,
L.L.C., 165 F.4th 399, 410 (5th Cir. 2026) (“Trade secret damages—whether measured by a
reasonable royalty or lost profits—must, like patent damages, ‘reflect the value attributable to the
infringing features of the product, and no more.’”). 02 Micro Int’l Ltd. v. Monolithic Power Sys.,
likelihood of disruption of the trial; and (4) bad faith or willfulness involved in not timely
disclosing the evidence.” Gryglak v. HSBC Bank USA, N.A., 2023 WL 3243998, at *2 (9th Cir.
May 4, 2023) (internal quotation marks omitted).
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 4 of 20

5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
2 This case has the following subsequent history: amended sub nom. O2 Micro Int'l Ltd. v.
Monolithic Power Sys., Inc., 420 F. Supp. 2d 1070 (N.D. Cal. 2006), aff’d, 221 F. App’x 996 (Fed.
Cir. 2007), and aff’d, 221 F. App’x 996 (Fed. Cir. 2007).
3 Oracle states that “Bennis structured her opinions and calculations so that if particular trade
secrets within a category or entire categories are dropped, any adjustment can be transparently
made.” ECF No. 385-3 at 8 (citing ECF No. 345-7 ¶¶ 280, 297–298, 306, 313, 316, 332–334,
336; ECF No. 345-8 at 248). This does not accurately characterize the report, which does not
enable apportionment among individual trade secrets within a category. Paragraph 280 is an
introductory summary paragraph asserting Bennis’s intent to “consider[] what should occur in the
event that fewer than all the Oracle trade secrets are found to be misappropriated and [to]
provide[] calculations for those scenarios.” ECF No. 345-7 ¶ 280. Paragraphs 297 and 298
discuss options for apportioning Procore’s profits among three trade secret categories (TPM
Solution, Third-Party ERP Integrations, and Customer Information) based on which Procore Pay
functionalities they relate to, but these paragraphs do not discuss methods for apportioning profit-
based damages among specific trade secret files should fewer than all in a category be found to
have been misappropriated. Id. ¶¶ 297–98. Paragraph 306 calculates ill-gotten profits for each of
Procore Pay’s Accounting Integrations, but does not apportion Procore’s profits among Oracle’s
allegedly misappropriated trade secrets. Id. ¶ 306. Paragraph 313 likewise apportions profits
across Procore’s tools, but not across Oracle trade secrets. Id. ¶ 313 (“To the extent that only a
subset of the tools are found to be misappropriated, my calculation allows for a tool-specific ill-
gotten profit margin.”). Paragraph 316 is a table showing ill-gotten profits by Procore product and
Oracle Trade Secret category, but not by particular trade secret. Id. ¶ 316. Paragraph 332
discusses R&D total expenditures for the TPM Solution trade secret and the integrations trade
secrets collectively. Id. ¶ 332. Paragraphs 333 and 334 discuss apportioning the R&D for
integrations trade secrets equally across third party ERP platforms, but not how to apportion them
among individual trade secret files. Id. ¶ 333. This is especially problematic given that different
Inc. illustrates this principle. 399 F. Supp. 2d 1064, 1076 (N.D. Cal. 2005).2 There, the damages
expert “provided the jury with a damages calculation based on an assumption that all of the trade
secrets were misappropriated,” but the jury “found that only five of the eleven trade secrets were
misappropriated, and that only the misappropriation of one of them resulted in [the defendant]
being unjustly enriched.” Id. Although there was some evidence in the record that the one trade
secret resulting in unjust enrichment was of particular importance to the plaintiff, that evidence did
not interact with the expert’s damages analysis and as such did not provide a basis for
apportioning the total amount. Id. at 1076–77. The court granted the defendant’s motion for
judgment as a matter of law that the plaintiff failed to prove unjust enrichment damages. Id. at
1077.
Bennis’s report apportions damages among the six trade secret categories disclosed by
Procore, but not by individual trade secret file within those categories.3 This is particularly
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 5 of 20

6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
numbers of asserted trade secret files relate to each third-party ERP: at least four CMIC trade
secret files are asserted at this stage, but only 2 Sage files, one Yardi file, etc. ECF No. 349-15 at
2–3. Paragraph 336 states that “[a]s illustrated below, I have considered what should occur in the
event that fewer than all the Oracle trade secrets are found to be misappropriated and have
provided calculations for those scenarios.” Id. ¶ 336. The accompanying chart, however,
apportions avoided R&D only by trade secret category, not individual trade secret. Id. Finally,
the table at page 348 of Exhibit 6 (ECF No. 345-8) likewise apportions ill-gotten profits by trade
secret category but not trade secret.
problematic, in Procore’s view, because Bennis’s damages report analyzed the 110 trade secrets at
issue at that stage, rather than the maximum of 24 trade secrets that will remain when the case
goes to trial. ECF No. 345-2 at 9. The Court will nonetheless not strike Bennis’s report on this
basis.
First, in the ordinary case, insufficient damages apportionment among trade secrets does
not necessarily mean that a damages report is unhelpful, unreliable, or based on insufficient facts
or data. See Fed. R. Evid. 702. The court in 02 Micro International acknowledged as much:
rather than excluding the damages report pretrial, it cautioned the plaintiffs that “you can have that
kind of damage evidence if you want to. I’m just warning you that if you do and you don’t win on
all 12 of your trade secrets, you’ll have no damages.” 399 F. Supp. 2d at 1076. Adequate
evidence from outside the damages report may also provide a means for apportioning damages.
For instance, the Ninth Circuit recently upheld a jury award of half of the total unjust enrichment
amount where only one of nine asserted trade secrets was misappropriated. See EchoSpan, Inc. v.
Medallia, Inc., No. 24-4751, 2025 WL 3046753, at *2 (9th Cir. Oct. 31, 2025). Because the jury
had before it evidence “including an expert opinion regarding [the defendant’s] total unjust
enrichment from creating a 360-degree review product that it could sell, expert guidance that the
jury should consider ‘value drivers,’ and testimony about the main drivers of value in [the
plaintiff’s] 360-degree review product,” its award was upheld. Id.
Courts exclude damages opinions for failure to apportion where they lack any utility to the
jury. LivePerson, Inc. v. [24]7.AI, Inc. is one example. No. 17-CV-01268-JST, 2018 WL
6257460, at *2 (N.D. Cal. Nov. 30, 2018). In that case, the damages expert’s opinion was
“excluded because he [did] not apportion trade secret misappropriation damages among particular
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 6 of 20

7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
alleged trade secrets, and offer[ed] no methodology for the jury to calculate trade secret
misappropriation damages on fewer than all of the 28 alleged trade secrets in the case.” Id. This
was especially damning because “the initial trial is a bellwether trial in which only 15 of the 28
alleged trade secrets will be at issue,” so “the jury’s verdict will necessarily encompass fewer than
all of the alleged trade secrets.” Id.
Procore contends that this case is exactly like LivePerson because Oracle will be allowed
to assert only 24 of the 110 trade secrets discussed in Bennis’s report. ECF No. 345-2 at 9. But
unlike in LivePerson, Oracle does offer some “methodology for the jury to calculate trade secret
misappropriation damages on fewer than all of the . . . alleged trade secrets in the case.” 2018 WL
6257460, at *2. The report provides both (1) justifications for awarding the total damages for the
full category even if only some trade secrets contained therein are misappropriated and (2) in some
cases, means for apportioning damages further within a category.
Oracle’s primary argument is that further apportionment may not be warranted if Procore
“still has no alternatives to the remaining trade secrets (lost profits, royalty).” See ECF No. 385-3
at 9; see ECF No. 345-7 ¶¶ 204, 368. Where a given trade secret is demonstrably necessary to
Procore’s ability to develop and sell the competing product, awarding the full lost profits or ill-
gotten sales may be justified if those profits or sales would not have occurred but for the
appropriation of that trade secret. ECF No. 385-3 at 15 (“Because Procore could not offer Procore
Pay without any one of the Procore Pay-related trade secrets, misappropriation of any one of those
trade secrets justifies the full amount of lost profits; further apportioning is not “economically
logical” because Oracle “cannot lose a portion of a sale.” (quoting Sprint Commc’ns v. Charter
Commc’ns, No. 17-1734-RGA, 2021 WL 982732, at *13 (D. Del. Mar. 16, 2021)). Indeed, as
courts have held in the patent context, “[l]ost profit damages do not depend on the number of
[trade secrets misappropriated] by one single product.” See Crystal Semiconductor Corp. v.
TriTech Microelecs. Int’l, 246 F.3d 1336, 1357 (Fed. Cir. 2001). As to avoided R&D costs,
Oracle’s expert and fact witnesses explain “that development of the TPM technology was an
iterative process, such that the TPM components and integrations build on each other and earlier
R&D,” which could justify awarding aggregate, unapportioned damages figures for
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 7 of 20

8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
2. Reliability of Price Erosion Opinion
Bennis opines that Oracle lost $ million due to price erosion because it would have
raised its subcontractor fee percentage for TPM from % to % of contract value in January
2023 but did not do so because of Procore Pay’s anticipated market entry. ECF No. 345-7 ¶¶ 261–
73. Procore argues that this price erosion opinion ought to be excluded (1) because it is based on a
made-for-litigation Oracle spreadsheet reflecting historical Textura sales and (2) because it does
not account for elasticity of demand. ECF No. 345-2 at 13–17. Neither argument justifies striking
this portion of the report.
4 Procore also suggests that the Bennis report could lead the jury to double-count the same
damages if more than one trade secret is found to be appropriated, but nothing about Bennis’s
analysis requires double counting and juries are capable of understanding the basis for a damages
estimate and avoiding crediting the same basis more than once when inappropriate. ECF No. 345-
2 at 11. This issue can also be addressed through the jury instructions if the Court concludes that
it is necessary.
misappropriation of mere subsets of trade secrets. ECF No. 385-3 at 12–13; ECF No. 345-20
¶¶ 543, 547.
The Bennis report also provides evidence enabling some apportionment. For instance, if
the jury finds that only a certain subset of Oracle trade secrets were misappropriated such that only
one or two of Procore’s workflow, budget, and direct costs Project Financials tools were
developed as the result of misappropriation, then Bennis explains how to divide Procore’s ill-
gotten sales appropriately. ECF No. 345-7 ¶ 313. As another example, Oracle asserts that
Bennis’s R&D estimates can easily be adjusted by time frame, e.g., if “2021 files are dropped . . .
the 2021 R&D spend can be removed from [Bennis’s] calculation.” ECF No. 385-3 at 13–14.
Because it is not possible to peer into alternate universes to compare profits and losses
given each permutation of misappropriated trade secrets, trade secret damages are inherently
imprecise. The Bennis report is precise enough to enable a jury to award damages in a variety of
scenarios. At the same time, the Court does not suggest that the Bennis report will necessarily
provide sufficient evidence. Indeed, like the plaintiffs in 02 Micro International, Oracle may
come to the end of trial and face reversal of a jury award that the evidence did not justify. But
Procore has not justified excluding the report now based on apportionment issues.4
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 8 of 20

9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
a. Reliance on Oracle’s Sales Spreadsheet
In calculating price erosion for TPM, Bennis relied on testimony from Mike Antis, Global
VP of Textura, that Oracle considered including the subcontractor percent-of-contract value fee
basis points, from 0.22% to or , but declined to do because of Procore Pay. ECF
No. 345-7 ¶¶ 262; ECF No. 385-8 at 8–9. She considered data from 2011, 2014, and 2019
showing that before Procore Pay entered but while other competitors like GCPay and Flashtract
were on the market, Oracle increased its percent-of-contract-value pricing basis points.
ECF No. 345-7 ¶¶ 260–62. Antis explained that little to no attrition or customer pushback resulted
from those price increases, reflecting an inelastic market. Id. ¶¶ 262, 270. Bennis also considered
the general inflation rate across the relevant period, concluding that Oracle’s proposed fee increase
would have been in line with inflation. Id. ¶ 269. Bennis then analyzed data provided by Oracle
applying the higher fee to its actual sales. Id. ¶¶ 264–66. She also interviewed and reviewed the
deposition of Michael Russo, the Oracle employee who prepared the data. Id. ¶¶ 261–70; ECF
No. 385-6 at 10. Procore also deposed Russo. See ECF No. 385-10.
Procore raises several objections. First, Procore repeatedly implies that the sales data upon
which Bennis bases her price erosion opinion is unreliable because it was compiled for the
purpose of litigation. ECF No. 345-2 at 13–14. That data was compiled for litigation does not
alone make it unreliable and Procore does not identify any authority to the contrary. Procore cites
In re Imperial Credit Industries, Inc. Securities Litigation, for instance, but that case concluded
that an expert may not “rely upon excerpts from opinions developed by another expert for the
purposes of litigation” because such opinions lack “independent indicia of reliability.” 252 F.
Supp. 2d 1005, 1012 (C.D. Cal. 2003), aff’d sub nom. Mortensen v. Snavely, 145 F. App’x 218
(9th Cir. 2005). Here, Oracle does not rely on “excerpts from opinions developed by another
expert for the purposes of litigation,” but merely “facts and data” which happen to have been
compiled for the purpose of litigation. Id. Notably, had Bennis somehow gathered the sales data
herself, it still would have been compiled for the purpose of litigation. And although Procore
deposed Russo, it does not raise any specific concern about the veracity of the data he put
together. See ECF No. 385-10.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 9 of 20

10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
3. Reliability of Reasonable Royalty Opinion
Bennis’s royalty damages opinion of $ million is the sum of two numbers: $
million of allegedly avoided R&D for the TPM Solution, Oracle ERP Integrations, Third Party
ERP Integrations, Cross-Platform Integrations, and Customer Information categories, and $
Second, Procore asserts that “Bennis did not scrutinize Oracle’s spreadsheet, and instead
adopted it in full and projected further damages using the same sales figures from 2025.” ECF
No. 345-2 at 13. “She did not isolate the impact (if any) of the alleged misappropriation and did
not account for customer attrition, reduced usage, or lawful competition.” Id. at 13–14. But as
outlined above, Bennis did consider multiple factors, including evidence of the historic behavior
of Oracle’s customers, in determining whether it would be appropriate to estimate price erosion by
simply multiplying actual sales by the would-be price increase. See also ECF No. 385-6 at 9. The
fact that Oracle’s employee applied that multiplication to a spreadsheet before she did is
immaterial.
Third, Procore argues that reliance on the spreadsheet may have resulted in an inflated
price erosion estimate, because Bennis did not take into account evidence of the pricing pressure
exerted by competitors like GCPay and the discounts Oracle was offering to customers. ECF No.
345-2 at 14. But these are mere fact disputes, not reasons to exclude Bennis’s report as unreliable
under Daubert. See Alaska Rent-A-Car, 738 F.3d at 969 (“[T]he judge is supposed to screen the
jury from unreliable nonsense opinions, but not exclude opinions merely because they are
impeachable.”).
b. Elasticity of Demand
In assuming inelastic demand, Bennis relied on testimony from Oracle fact witnesses
concerning Oracle customers’ historic behavior given similar price increases. ECF No. 345-7
¶¶ 262, 270; ECF No. 385-6 at 10. The extent to which Bennis’s price erosion calculation should
have nonetheless accounted for demand elasticity is a quintessential fact question to be raised in
cross-examination and resolved by the jury. See Ericsson. v. Harris Corp., 352 F.3d 1369, 1378–
79 (Fed. Cir. 2003) (upholding damages award based on evidence of inelastic market and noting
that the jury presumably resolved any elasticity issues).
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 10 of 20

11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
4. Reasonable Basis for Damages Period
Procore finally argues that Bennis’s asserted damages period, which runs through trial, is
unsupported by any evidence that Procore would not have been able to develop and launch
Procore Pay independently before that cut-off. ECF No. 345-2 at 20. But Bennis’s damages
5 See Georgia–Pacific Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116, 1120 (S.D.N.Y. 1970).
million for the Internal Business Information category, which is the estimated amount of price
erosion. ECF No. 345-7 ¶¶ 378–379; ECF No. 346-7 at 40, 43–44.
Procore first argues that this figure is unreliable because it propagates the deficiencies
already discussed in the component parts. ECF No. 345-2 at 17–18. The Court has already
concluded, however, that neither the data source and methodology of the price erosion calculation
nor the aggregate, unapportioned nature of the R&D calculation provides a basis for striking those
sections of the report. Procore also argues more broadly that “Bennis’s apportionment failure is
on starkest display in her reasonable royalty analysis,” but the Court has likewise already
explained why the failure to apportion does not here present a basis for striking the report. ECF
No. 345-2 at 18–19.
Procore next suggests that Bennis’s analysis of the Georgia-Pacific5 factors, which
consider what reasonable royalty would result from a hypothetical negotiation between parties, are
conclusory because she uses them to rationalize the figure pre-derived by summing price erosion
and R&D costs. ECF No. 345-2 at 18. To the contrary, Bennis explains why application of the
Georgia-Pacific factors would arrive at the same result. ECF No. 345-7 at 354–374. Courts have
routinely rejected arguments like Procore’s. See, e.g., Willis Elec. Co. v. Polygroup Ltd., 166
F.4th 1363, 1385–86 (Fed. Cir. 2026) (“It is sufficient for an economic expert to explain how
qualitative considerations [under Georgia-Pacific] influence where a royalty should fall within an
already-apportioned numerical range.”); Atl. Inertial Sys. v. Condor Pac. Indus. of Cal., No. 2:08–
cv–02947–CAS (PJWx), 2015 WL 3825318, at *15 (C.D. Cal. June 18, 2015) (applying the
Georgia-Pacific factors to validate a quantitatively-derived royalty figure). Bennis’s royalty
estimate is not unreliable under Daubert.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 11 of 20

12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
The motion to exclude the expert opinion of Melissa Bennis is denied in full. ECF No.
344.
B. Motion to Exclude Opinions of Andrew Crain
Oracle retained Andrew Crain as its forensics expert. ECF No. 286-2 ¶ 1. He has “25
years of experience performing forensic imaging and analysis of computer hard drives and other
electronic storage media, advising clients regarding e-discovery and forensics issues, and
providing expert testimony.” Id. ¶ 2. He has “led and conducted hundreds of computer forensics
investigations (often involving numerous computers, devices, and massive amounts of data in a
single case) . . . in the context of a wide range of legal matters.” Id. ¶ 5. Procore seeks to strike
three aspects of his report under Daubert: (1) his inclusion of “stub” placeholder files in his
description of the files “existing on” Mariano’s devices, (2) his attribution to Mariano of intent to
destroy evidence and a pattern of transferring and using trade secret files, and (3) his imputation of
certain conduct to Procore. ECF No. 340. Each objection is largely semantic and none justifies
excluding portions of the report.
1. Stub Files
Procore seeks to “narrow[ly] exclude Crain’s “repeated use of stub files to inflate the count
of Asserted Trade Secrets that ‘existed on’ a device.” ECF No. 396-2 at 6. A basic review of the
relevant passages, however, shows that the Crain report is fastidious in distinguishing stub files
from full files. Stub files “are not actual files, but rather electronic placeholders for files whose
contents never left the Cloud.” ECF No. 341-2 at 6. The Crain report acknowledges this
distinction, explaining that:
The majority of the iCloud related files on the Procore MacBook
A2251 are placeholder files, or ‘stubs’; these are small files
automatically created by iCloud during synchronization and stored on
period was supported by the Almeroth report, which opined that Procore to this day has not
independently developed the trade secrets, has no alternatives to them, and would not have been
able to achieve Procore Pay in a “commercially reasonable time” without the trade secrets. See
ECF No. 346-9 ¶¶ 582–604. That evidence is sufficient to support her opinion that damages
should be calculated through the date of trial.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 12 of 20

13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
the local computer, indicating that Mr. Mariano indeed retained
access to these Asserted Trade Secret files through the Procore
MacBook A2251 computer. These stubs provide a visual
representation of the files that exists in the cloud, such that the user
can “browse” their iCloud files and folders just as if they were stored
locally. By default, stubs mirror several metadata values from iCloud,
but do not contain the file’s content until the file is opened, previewed,
copied, expressly chosen for download, etc. In such an event, the
computer will retrieve the file’s contents from iCloud and the local
‘stub’ is replaced by the substantive file (lacking the prepended “.”
and the “.icloud” extension).
ECF No. 341-3 ¶ 62. In his deposition, Crain testified that “if the synchronized file exists in stub
form only, we can infer or determine that the user has not done anything with the file.” ECF No.
341-10 at 41.
Procore moves to strike Paragraphs 87, 94, 113, and Exhibit C. Paragraph 87 explains that
“many of the Asserted Trade Secret files exist on the Mariano iPhone in the form of iCloud stubs
(indicating that Mr. Mariano indeed retained access to these Asserted Trade Secret files through
the iPhone), while the substantive versions of two Asserted Trade Secret files are locally stored on
the device, showing they were downloaded from iCloud in order [to] complete an action such as
opening, copying, printing, emailing, etc.” Id. ¶ 87. Paragraph 113 claims that “62 of Plaintiff’s
Asserted Trade secrets existed on the Mariano Mac Mini. . . . [A] significant number of the
Asserted Trade Secret files resided on the Mariano Mac Mini as iCloud stubs, while the
substantive copies (i.e., not stubs) of 4 Asserted Trade Secret files were locally stored on the
device.” Id. ¶ 113. And Paragraph 94 states “that 70 of Plaintiff’s Asserted Trade secrets existed
on the Procore MacBook Pro A2251 . . . via several means, including at least iCloud
synchronization, downloads from Google Drive, Gmail, and LinkedIn,” summarizing Exhibit C.
Id. ¶ 94. Exhibit C, finally, lists storage locations and access history for the asserted trade secret
files, differentiating between full files and stub filed with an “X” or “XX” notation. ECF No. 286-
2 at 76–83; ECF No. 341-2 at 9 n.6.
Procore’s quarrel is with Crain’s description of stub files as “existing on” a certain device.
Based on the definition of stub files provided above, the use of the verb phrase “exists on” is
plausibly accurate. In addition, the identified passages provide the ample additional context
reproduced above, which cures any confusion engendered by the use of the term “exists on.”
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 13 of 20

14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Procore has identified nothing inaccurate or even misleading about these passages. The request to
strike them is denied.
2. Mariano’s Intent
Procore next seeks to exclude two passages on the grounds that they impermissibly opine
on Mariano’s intent. ECF No. 341-2 at 12. In Paragraph 33, Crain explains that a factory reset is
“not a routine maintenance task, nor is it a step that a user would perform in the daily course of
operating a MacBook Pro,” but “a measure that completely and irreversibly erases all data stored
on the machine” and that “cannot be explained by any standard software update, system
optimization, or disk management routine.” ECF No. 286-2 ¶ 33. It “requires a user to take
multiple concrete steps as described above, and this process does not occur accidentally” but
“reflects a decision made to destroy the data and activity history residing on that machine, and it is
consistent with an effort to eliminate evidence of what was stored on or interacted with using the
computer rather than any routine housekeeping or troubleshooting.” Id. Procore takes issue with
the suggestion that Mariano made a “decision . . . to destroy” data. ECF No. 341-2 at 12. But the
passage only explains that a factory reset, by its nature, results from a deliberate decision to
destroy data. It is a technical opinion based on the technical features of a factory reset, drawing
inferences only from the details of its implementation and implications. It does not bear on
Mariano’s reasons for resetting his computer. Procore’s cited orders prohibiting certain experts
from opining on “intent, motive, or state of mind” therefore do not apply. See, e.g., Oracle Am.,
Inc. v. Hewlett Packard Enter. Co., No. 16-cv-01393-JST, 2018 WL 6511146, at *3 (N.D. Cal.
Dec. 11, 2018).
Procore also objects to Crain’s use of the word “pattern” to describe certain conduct by
Mariano. In Paragraph 16(g), Crain opines that “[a]fter starting at Procore, Mr. Mariano engaged
in a pattern of transferring and using Asserted Trade Secret files on his personal devices and
Procore devices,” including because the “Construction Accounting 101.pptx” trade secret “was
modified in his Google Account, as well as downloaded, opened, and used on a Procore MacBook
Pro . . . through at least March 2023.” ECF No. 286-2 ¶ 16(g). Again, Procore has identified
nothing inaccurate or outside of Crain’s expertise. ECF No. 341-2 at 13. Procore may explore
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 14 of 20

15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
any imprecision in the descriptive term “pattern” during cross-examination. See In re Lidoderm
Antitrust Litig., No. 14-md-02521, 2017 WL 679367, at *28 (N.D. Cal. Feb. 21, 2017).
3. Imputation to Procore
Procore finally objects to the section headings6 stating that “Procore permanently
destroyed forensic evidence” and “Procore obtained and used plaintiff’s asserted trade secret
files.” ECF No. 341-2 at 16 (citing ECF No. 286-2 at 3). As Procore argues, “whether Procore’s
employees’ actions, including those of Mr. Mariano, are attributable to Procore the company is a
legal question, not a forensic one.” Id. However, use of the subject “Procore” is a reasonable and
accurate way to capture the facts “that one of the individuals whose electronic devices and
accounts were forensically analyzed (Mr. Mariano) was a Procore Director [who] transferred
Oracle’s documents to his Procore computer and disseminated Oracle’s documents to his Procore
coworkers.” ECF No. 376-3 at 17. Nothing about that factual statement prevents Procore from
parsing and contesting which actions by Mariano or other Procore employees are or are not legally
attributable to the company as a whole. The entirety of the motion to exclude portions of the
expert report of Andrew Crain, ECF No. 340, is denied.
C. Motion to Strike Portions of the Expert Report of Andrew Crain
Procore also moves to strike portions of the Crain report on the grounds that they rely on
forensic images of two Oracle laptops and that those images should have been, but were not,
timely produced to Procore. ECF No. 289.
1. Supplemental Background
Discovery disputes in this case were referred to Magistrate Judge Laurel Beeler. ECF No.
66. In a March 21, 2025 order adjudicating a discovery letter brief, Judge Beeler rejected
wholesale production of device images—which contain voluminous personal, third-party,
privileged, and unrelated confidential information—and ordered the parties to enter into a Forensic
Inspection Protocol Agreement. ECF No. 68 at 2; ECF No. 74. The Agreement, filed on April 8,
6 In a footnote, Procore also identifies paragraphs 14, 23, 54, and 121 as containing similarly
problematic language. ECF No. 341-2 at 7 n.4. To the extent that Procore also seeks to strike
these passages, the request is denied for the reasons explained above.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 15 of 20

16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
2025, governs “documents, files, accounts, and devices . . . subject to forensic examination.” Id.
at 1. The parties jointly engaged “a neutral and independent forensic examiner . . . to carry out”
the Agreement. Id. ¶ 1. The neutral would image certain enumerated “Forensic Devices,”
including Oracle MacBook and Lenovo laptops used by Mariano, maintain the images, and “not
provide or disclose the contents of the digital forensic images to the Parties or to any third party
except as provided herein.” Id. ¶¶ 3, 4(a); see also id. ¶ 21 (“Except as explicitly set forth in this
Protocol, the Forensic Examiner shall not provide any information regarding the Producing Party’s
Devices or the contents thereof to the Requesting Party.”). The Agreement circumscribed what
information could be exchanged: the neutral’s reports based on agreed search terms, id. ¶¶ 6–10,
and the underlying “files, documents, or emails” identified by those reports, id. ¶ 11.
Paragraph 22 of the Forensic Agreement provides that “nothing in this Protocol prevents
any Producing Party from conducting forensic collections, examinations, or inspections on any
Forensic Device or other item, source, thing, or device in its possession, custody, or control,
including forensic collections, examinations, or inspections that may be protected as attorney work
product or other applicable privilege.” Id. ¶ 22.
The parties conducted discovery based on what the Forensic Agreement allowed and the
neutral produced. ECF No. 289 at 8–9. Issues over production requests perceived to be outside of
the Forensic Agreement were aggressively litigated and submitted to Magistrate Judge Beeler for
resolution. See, e.g., ECF No. 253.
Separately, Procore served Request for Production No. 44 on April 15, 2025, demanding:
“All Documents and Things provided by Mr. Mariano to Plaintiffs after his departure, including,
but not limited to, his personal Google Drive and iCloud accounts and the two computers
referenced in paragraph 37 of the Complaint, whether it was forensically collected by Stroz
Friedberg on Your behalf or otherwise.” ECF No. 289-7 at 15 (referring to the MacBook and
Lenovo Oracle laptops Mariano kept after his termination, see ECF No. 1 ¶ 37, both of which are
Forensic Devices under the Agreement).
Oracle had in its possession forensic images of the Oracle laptops, including evidence not
appearing in the Forensic Reports. It never produced the images in response to RFP No. 44,
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 16 of 20

17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Mariano . . . . Certain artifacts from each of these images, as described in this report, form the
basis of [Mr. Crain’s] opinions with respect to these devices.” ECF No. 341-3 ¶ 19.
On April 16, 2026, Procore’s counsel asked Oracle to withdraw the impacted paragraphs of
the Crain report based on the unproduced laptop images (and the emails, asserted trade secret files,
forensic material, and other information they contain). ECF No. 289-3 at 6. Oracle instead
produced the withheld images to Procore on April 17, 2026. Id. at 3; ECF No. 319-3 at 2. On
May 18, 2026, Oracle served its rebuttal expert report which included a forensic analysis of the
forensic images at issue in Procore’s motion, and which responded to Mr. Crain’s opinions on
those images. Id. at 2. On May 27, 2026, Oracle filed a notice with the Court informing it of this
development and arguing that any prejudice to Procore had been cured. Id. Procore did not move
for leave to respond to this notice.
2. Discussion
Oracle violated Rule 26 in allowing its expert to analyze the forensic images of the laptops
without producing them to Procore. Nonetheless, because the untimely production to Procore
appears to have cured any meaningful prejudice, the motion to strike certain portions of the Crain
report is denied. ECF No. 289.
Oracle provided the forensic images to its expert for use in his expert report—to advance
its core claims—while simultaneously failing to produce those images to Procore under Rule 26 or
in response to Request for Production No. 44. In doing so, it violated Rule 26(a)(1)(A)(ii),
requiring that a party provide to the other parties “all . . . electronically stored information, and
tangible things that the disclosing party has in its possession . . . and may use to support its claims
or defenses,” and Rule 26(e)(1)(A), requiring that a 26(a) disclosure be supplemented “in a timely
manner if the party learns that in some material respect the disclosure or response is incomplete.”
United States v. North East Medical Services is instructive. No. 10-CV-01904-CW (JCS), 2014
apparently because of the Forensic Agreement. ECF No. 289 at 9. It nonetheless gave the images
to Crain to rely on in his opening report. See id. at 10. Procore did not learn of this fact until the
Crain report was served on April 8, 2026, including a paragraph stating that “[Crain’s firm] was
provided access to forensic images of Oracle’s corporate-owned devices issued to Mr.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 17 of 20

18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
7 Oracle asserts that it was authorized by Paragraph 22 of the Forensic Agreement to provide the
forensic images to its forensic expert. ECF No. 294 at 11. Procore responds that Paragraph 22
“was intended for consulting or work-product inspections for internal purposes, not to license
sharing party information that was never produced during fact discovery.” ECF No. 289 at 13.
The critical point is this: the Forensic Agreement was designed to cabin and replace unfettered
exchange of the devices and device images. Because of the Forensic Agreement, Oracle did not
provide the device images to Procore. At the same time, Oracle used the device images—not the
reports produced by the neutral’s examination of the images—as the basis for its expert report.
Even if authorized by Paragraph 22, doing so triggered obligations under Rule 26 and Request for
Production 44. Oracle cannot—as it does in its opposition—maintain that the Forensic Agreement
allowed it to use the raw device images as the basis for its own claims but simultaneously
prohibited it from providing those images to Procore in discovery. Any such interpretation of the
Agreement would defy logic.
WL 7208627, at *7 (N.D. Cal. Dec. 17, 2014). There, the Court struck a supplemental production
and corresponding expert report that defendant had introduced after the close of fact discovery
because “Rule 34 required [the defendant] to produce documents responsive to the [plaintiff’s]
requests for production, Rule 26(a)(1) required [the defendant] to produce or disclose documents
relevant to its defenses, and Rule 26(e)(1) required [the defendant] to supplement its production
and/or disclosure in a timely manner if it later learned that it was incomplete.” Id. at *5–7. Oracle
circumvented the Forensic Agreement7 and, like the defendant in North East Medical Services,
and violated the Federal Rules of Civil Procedure by allowing its expert to rely on the forensic
images without producing them to Procore.
Relying on the terms of the Forensic Agreement, Procore understood that both its
discovery and Oracle’s would be cabined by the agreement and limited to the neutral’s
productions. In Oracle’s defense, it explains that it provided the forensic images to the forensic
neutral. ECF No. 294 at 9. But providing the images to the neutral was no substitute for
providing them to Procore, especially because the neutral was not authorized to pass them on to
Procore. See ECF No. 74 ¶¶ 3, 21. Oracle also argues that Procore could have requested
“whatever additional discovery and forensic analysis of those laptops and their images it deemed
necessary.” ECF No. 294 at 10. Paragraph 12 of the Agreement provides that “[s]hould a party
require additional forensic information regarding a specific time-period or file, it shall provide a
written request, specifying (1) the specific forensic information requested, and (2) the basis for
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 18 of 20

19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
The burden is on Oracle “to show that the failure was substantially justified or harmless.”
M.H. v. Cnty. Of Alameda, No. 11-cv-02868-JST, 2015 WL 14073366, at *2 (N.D. Cal. Feb. 1,
2015). Here, it appears that while withholding the images prejudiced Procore, that prejudice has
been cured by the production on April 17 and the likelihood of disruption of the trial is negligible.
See Gryglak, 2023 WL 3243998, at *2 (listing factors courts consider in assessing harmlessness).
In its motion, Procore argues that it was deprived of the opportunity “to depose Oracle’s
witnesses, prepare its own witnesses, or issue discovery based on the Oracle device images.” ECF
No. 289 at 6. Had it timely received the files, it would have “pursued targeted document requests,
questioned witnesses about the artifacts, emails, and files extracted from the laptops, cross-
examined custodians, and prepared its own witnesses accordingly.” Id. But Procore identifies no
specific learnings from the forensic images that would have affected its approach to fact
discovery. It also does not argue that Oracle’s approach to fact discovery was affected by the
forensic images in any particular way. Rather, the primary utility of the forensic images appears
to have been through the forensic experts and both experts had the opportunity to consider the
forensic images in forming their opinions. Under these circumstances, Oracle has met its burden
to establish harmlessness. See, e.g., Brocade Commc’ns Sys., Inc. v. A10 Networks, Inc., No. 10-
CV-03428 LHK, 2012 WL 13234023, at *37 (N.D. Cal. Sept. 18, 2012) (affirming denial of
motion to strike where prejudice was “uncertain”).
Oracle violated the forensic protocol and circumvented the forensic neutral by providing
the forensic images to Crain. Having done so, it exposed those images to the requirements of Rule
26, with which it failed to comply. Nonetheless, because the failure to produce the images was
relevance of the requested information, including any applicable support.” ECF No. 74 ¶ 12. To
the extent that Procore failed to utilize this provision, that failure is irrelevant. Unaware that
Oracle had circumvented the Agreement and provided the forensic images to its expert, Procore
would have no reason to seek discovery of those images in particular. In any case, Oracle’s
insistence that “Procore chose not to request any additional materials or inspections during the fact
discovery process” is confusing at best, given that Procore did request such materials through RFP
No. 44. ECF No. 294 at 12.
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 19 of 20

20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
rendered harmless by their belated production, the motion to strike certain portions of the Crain
report, ECF No. 289, is denied.
CONCLUSION
The motion to exclude portions of the Bennis report and the two motions to exclude and
strike portions of the Crain report are denied in full. ECF Nos. 289, 340, 344.
IT IS SO ORDERED.
Dated: September 25, 2026
_________ _____________ ______________
JON S. TIGAR
United States District Judge
Case 4:24-cv-07457-JST Document 438 Filed 10/14/26 Page 20 of 20