Unreported / Non-Citable
Background
Oracle America, Inc., Oracle International Corporation, and Textura Corporation sued Procore Technologies, Inc. and Procore Payment Services, Inc. in October 2024 under the Defend Trade Secrets Act. Oracle alleges Procore used trade secrets about its Textura Payment Management construction payment software to develop Procore Pay, largely through former Oracle employee Mark Mariano's move to Procore in late 2021. Oracle narrowed its asserted trade secrets from more than 450 to 34 in six categories.
Procore filed three expert motions. One sought to exclude damages expert Melissa Bennis for lack of apportionment by individual trade secret, an unsupported price erosion opinion, an unreliable royalty figure, and a damages period running through trial. A second challenged forensics expert Andrew Crain's treatment of iCloud stub files, Mariano's device reset and file transfers, and headings attributing conduct to Procore. A third sought to strike Crain opinions based on images of two Oracle laptops that Oracle gave Crain but did not produce to Procore, despite a forensic inspection protocol using a neutral examiner.
The Court’s Holding
The court denied all three motions. Apportioning damages by category rather than by individual trade secret did not by itself make Bennis's report unhelpful or unreliable under Rule 702. Unlike the excluded expert in LivePerson, she offered some methodology, including reasons the full category amount might be recoverable if any trade secret in it was misappropriated and, in places, ways to divide profits or avoided research and development costs further. The court cited an unpublished Ninth Circuit memorandum, EchoSpan, upholding a partial award supported by other evidence. It cautioned that the report may still prove insufficient at trial and that a verdict could be reversed. Reliance on an Oracle-prepared sales spreadsheet and an inelastic-demand assumption went to weight. The royalty opinion was not unreliable merely because it summed those components and used Georgia-Pacific factors to explain the result. The Almeroth report supported a damages period through trial. Double-counting concerns could be handled through jury instructions if needed.
The court found Crain's description of stub files as existing on devices plausibly accurate and adequately qualified, treated his factory reset statement as a technical inference rather than testimony about Mariano's motive, and left his word 'pattern' to cross-examination. The Procore headings did not foreclose Procore from contesting legal attribution. On the motion to strike, the court held that Oracle violated Rules 26(a) and 26(e) by giving the images to its expert without producing them. But Oracle produced them on April 17, 2026, both sides' experts addressed them, Procore identified no specific lost discovery, and trial disruption was negligible, so Oracle showed harmlessness under Rule 37(c)(1).
Key Takeaways
- Category-level damages apportionment does not automatically require excluding a trade secret damages report when it offers some methodology for fewer-than-all scenarios and other trial evidence may help the jury apportion.
- Data compiled for litigation is not unreliable for that reason alone, and disputes over elasticity or competitor pricing generally go to weight rather than admissibility.
- A forensic expert may use descriptive terms such as 'existing on' or 'pattern' where the report explains the technical meaning, with any imprecision left for cross-examination.
- Giving raw device images to a testifying expert can trigger Rule 26 and document-request obligations even when a forensic protocol limits what a neutral examiner discloses.
- Belated production cured the prejudice here, but the denial does not decide whether Bennis's damages evidence will ultimately support a verdict.
Why It Matters
The order treats apportionment weaknesses as questions of proof at trial rather than grounds for pretrial exclusion, while distinguishing cases where an expert gives jurors no way to calculate damages for fewer than all trade secrets. It is a district court order, and the Ninth Circuit EchoSpan decision it cites is an unpublished memorandum, so neither is binding precedent.
The ruling also warns litigants that using forensic images from a neutral-examiner process as expert foundation triggers ordinary disclosure duties. Open issues include whether the damages evidence will suffice once the trial set of trade secrets is fixed, any double-counting instructions, and the merits of misappropriation and attribution to Procore, none of which was decided.