Impinj, Inc. v. NXP USA, Inc.

Opinion

1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 IMPINJ, INC., 7 Case No. 19-cv-3161-YGR Plaintiff, 8 v. PRETRIAL ORDER NO. 4, INCLUDING 9 MOTIONS TO EXCLUDE NXP USA, INC., 10 Dkt. Nos. 238, 240, 263, 269-8, and 297 Defendant. 11

12 13 I. Pretrial Issues 14 On Sunday, July 2, the parties sent the Court an email stipulating to excuse Juror Nos. 1, 9, 15 22, 26, 27, and 35. Said jurors were excused. 16 The Court has provided the parties with a draft of jury instructions for purposes of 17 facilitating further instructions. 18 The Court clarified that the parties should be prepared to proceed with opening statements 19 if a jury is chosen quickly. 20 II. Outstanding Motions to Exclude 21 A. Motion to Exclude Kindler (Dkt. No. 240) 22 The legal framework is not in dispute. Federal Rule of Evidence 702 permits opinion 23 testimony by an expert as long as the witness is qualified and based upon that qualification, the 24 witness’s opinion is relevant and reliable. An expert witness may be qualified by “knowledge, 25 skill, experience, training, or education” as to the subject matter of the opinion. Fed. R. Evid. 702. 26 The proponent of expert testimony has the burden of proving admissibility in accordance with the 27 rule. Id., Advisory Committee Notes (2000 amendments). Scientific opinions must be based on 1 Experts assist the factfinder in their own evaluation of the evidence by providing the factfinder 2 with opinions based upon verifiable, scientific, or other objective analysis. Id. at 589–90. 3 1. Overview 4 The two remaining patents at issue in this suit are directed to the shape of the channel 5 between large pads that minimize turbulence when customers attach the ICs of the products to 6 their antennas (the ’302) and improved rectifier design for enhancing read/write performance (the 7 ’597). See Dkt. No. 279-2 (“Oppo.”) at 3. NXP moves to exclude paragraphs 114-183 of 8 Kindler’s report on the grounds that (1) Kindler cannot have provided a reliable Georgia-Pacific 9 analysis because her starting point is arbitrary; (2) Kindler fails to properly apportion what value 10 derives from patented versus unpatented elements of the accused devices; (3) Kindler’s reliance on 11 lay witness and Impinj employee Ron Oliver is not proper. 12 Lauren Kindler is a managing principal at Analysis Group, Inc., which “provides 13 economic, financial, and business strategy consulting to its clients and specializes in the 14 interpretation of economic and financial data and the development of economic and financial 15 models.” Kindler Rpt. ¶ 4. Kindler has provided financial and economic consulting services for 16 over 18 years. See id. ¶ 5. Kindler received her B.A. in Economics from Tulane and her M.A. in 17 Economics from Southern Methodist University. See id. ¶ 6. In forming her opinions, she 18 reviewed legal documents, the patents themselves, deposition testimony, and other documents. 19 See id. ¶ 8. In addition, she held discussions with several Impinj officers and employees, 20 including Ron Oliver, a technical fellow. See id. In summary, Kindler opines that, due to NXP’s 21 sales of its UCODE 8 and UCODE 9 products, Impinj suffered lost profits due to patent 22 infringement, and she also calculated a reasonable royalty rate for sales NXP made for which 23 Impinj is not seeking lost profits. See id. ¶ 10. 24 Ms. Kindler’s report is founded on the premise that Impinj has lost sales of its Monza R6 25 product (RAIN RFID tag chips with a variety of applications). Ms. Kindler’s key method 26 underlying her reasonable royalty analysis is calculating the incremental losses to Impinj’s profits 27 during the time period from October 6, 2017 through the second quarter of 2022, during which 1 placed into products and distributed in the United States, and Impinj also seeks a reasonable 2 royalty for other unit sales. 3 Kindler has calculated a reasonable royalty rate for all NXP sales of the Accused Products, 4 in the event that the jury finds that lost profits are not an acceptable remedy. Kindler has also 5 produced a reasonable royalty rate for those sales on which Impinj fails to recover if the jury 6 awards compensation for some lost sales. 7 2. Criticism of Kindler’s Method 8 According to NXP, Kindler identifies four primary features: (1) “sensitivity 9 improvements,” (2) “big pads,” (3) “auto tune”1 and (4) “memory safeguard.”2 In light of these, 10 NXP argues, Kindler assigned no value to other features. For each given feature, Kindler assigns 11 a percentage value attributable to the teachings of the patent. For example, for sensitivity 12 improvements, that figure is 50% to the ’597, and for big pads, that amount is 75% attributable to 13 the ’631 and 25% attributable to the ’302. NXP’s profit margin is 41.8%, and Kindler applies 14 these percentages to the 57% of allegedly “at risk” sales at that profit margin. So, for example, 15 Kindler arrives at a 3% reasonable royalty rate for the ’597 by multiplying at risk sales (57%) by 16 profit margin (41.8%) by whole divided by one fourth because of the four features (25%) x 50% 17 attributable to the ’597. 18 Ms. Kindler’s reasonable royalty analysis is based on the assumption that the parties would 19 be negotiating in view of potential lost profits. Kindler’s reasonable royalty calculations for the 20 ’302 and ‘597 are based on a hypothetical negotiation concerning a license taking place in or 21 around May 2017. Kindler Rpt. ¶ 18. Kindler sets forth her formula for the royalty rates near the 22 end of her report. Id., n.444 (percentages referenced above)). 23 NXP argues that Kindler lacks a starting point, while Impinj labels this as “semantics.” 24 The Court agrees. Given the ending point, a starting point exists. Kindler arrives at her rates with 25

26 1 “Auto tune” is relevant only to the ’266 patent, and the Court dismissed claims of infringement of the ‘266 in its Summary Judgment Order. See Dkt. No. 339, MSJ Order at 4-6. 27 1 the following equation: (percentage of sales of accused products that would have been at risk 2 without a license, i.e. adjusted market share (Exhibit 7.3 to Kindler Rpt.)) x (profit margin) x (1/4 3 for the four primary, patented features of the accused products) x (the value of the patented feature 4 attributable to the particular patent).3 5 The Court finds Kindler’s formula provides a “classic way to determine the reasonable 6 royalty amount,” as was used in Open Text S.A. v. Box, Inc.: “multiply[ing] the royalty base, 7 which represents the revenue generated by the infringement, by the royalty rate, which represents 8 the percentage of revenue owed to the patentee[.]” No. 13-CV-04910-JD, 2015 WL 349197, at *1 9 (N.D. Cal. Jan. 23, 2015). Here, the market share and the profit margin represent the royalty base, 10 and the fractions Kindler identifies as representing the value of the patented features and the 11 degree to which that value is assignable to a given patent represent the royalty rate. That is the 12 starting point. Kindler’s choice to reveal the starting point towards the end of her report is of no 13 material consequence. 14 3. Criticism of Kindler’s Calculations and Inputs 15 NXP argues that Kindler does not account for the value of unpatented features, and, if she 16 does, that those features may provide value even if they do not drive demand. Impinj contests 17 this, arguing that the 43% of the market that would not be at risk represents demand that would not 18 have deviated from the status quo, i.e. products that did not have the patented features. Kindler 19 argues that the sharp decrease in UCODE 7 products from 2017 to 2021 at the same time as the 20 steep rise in demand for Impinj’s products shows strong demand for products with the patented 21 features. Kindler Rpt. ¶ 13, n.7. Kindler acknowledges that NXP would have continued making 22 sales of non-infringing products, and she assumes that about 10% of the market for NXP’s 23 3 Kindler derives the 57% figure from adjusted market share. NXP’s profit margin on the 24 accused products (specifically the UCODE 8), was 41.8% in 2017. The 25% figure is derived from the fact that, at the outset of the hypothetical negotiation, there would be no need for such a 25 reduction because the parties would be negotiating over patents that cover only one feature. By the end of the negotiations, however, the parties would have come to terms on patents for four 26 features. Notably, however, one of those features is no longer in play in this suit and another is being litigated in the Western District of Texas. For the ‘302, Kindler opines, based on 27 conversations with Oliver, that 25% of the improved sensitivity is due to its teachings. For the 1 products would have come from other sources. 2 Impinj argues, moreover, that Kindler is not ignoring the value of unpatented features that 3 NXP added to the accused products. Rather, she is merely assigning them a value of zero. There 4 is no requirement that an expert find that the non-patented technologies of the accused products 5 have a nonzero value. The expert is required only to provide an apportionment. See Salazar v. 6 HTC Corp., No. 2:16-CV-01096-JRG-RSP, 2018 WL 1783157, at *1 (E.D. Tex. Apr. 13, 2018) 7 (allowing apportionment of zero for nonpatented features where expert did consider them). NXP’s 8 marketing efforts to promote those other features as “key” contrasts with other evidence of their 9 relative worthlessness, such as customers requesting that they be removed from the products. The 10 Court agrees with Impinj that these arguments go to weight, not admissibility. 11 Further, Kindler opines that, absent any one of the patented features, NXP would have 12 failed to make 57% of its infringing sales, and consequently would have been willing to forego its 13 incremental profit margin on 57% of its sales in exchange for the opportunity to sell products with 14 those patented features. Ms. Kindler appropriately hypothesizes about each patent independently, 15 but she reasonably constrains her analysis by assuming that NXP would not have been willing to 16 pay more than its total incremental profits on all allegedly infringing sales. As a result, inputs to 17 Ms. Kindler’s analysis shift accordingly. At each juncture, the market loss risk remains the same, 18 because it is Impinj’s and Kindler’s opinion that each patented feature is sufficiently in demand to 19 warrant buying a product with that feature. Thus, the 57% remains the same, and in a true 20 hypothetical negotiation absent the entrance of the other patent negotiations, there would be no 21 need to apportion the value by feature. However, the sum of the rates cannot be equal to more 22 than the fractional value of the patents as a whole. Further, if Impinj fails to convince the jury of 23 the underlying theory then a failure of proof will exist and lost profit damages will not issue. 24 NXP’s argument that Kindler should be excluded from testifying as to the value that each 25 patent contributes to the patented features fails to persuade. First, Kindler is entitled to rely on 26 Oliver, a technical expert, because this is a technical input. Second, and to that end, Ms. Kindler 27 does not intend to testify about these apportionments themselves. Third, there is no prohibition 1 13-CV-03999-BLF, 2015 WL 4272870, at *5 (N.D. Cal. July 14, 2015) (reasoning that patent 2 damages experts frequently rely on the technical expertise of others, and looking “to whether the 3 factual basis for [the contested expert’s] opinion was properly disclosed and available for cross- 4 examination) (internal citations omitted). 5 Relatedly, NXP’s final argument is that Kindler’s testimony should be excluded because 6 she relies on Oliver for her technical knowledge of the extent to which each asserted patent 7 contributes to the protected features is duplicitous. NXP cannot have it both ways. NXP 8 essentially criticizes Kindler for opining on technical elements when she is not a technical expert 9 while also lambasting her for relying on a technical expert for technical expertise. Nothing about 10 Kindler’s reliance on Oliver is outside the bounds of “facts or data in the case that the expert has 11 been made aware of[.]” Fed. R. Evid. 703. Moreover, “if experts in the particular field would 12 reasonably rely on those kinds of facts or data in forming an opinion on the subject, they need not 13 be admissible for the opinion to admitted.” Id. 14 For the foregoing reasons, the Court DENIES NXP’s motion to exclude Kindler. 15 B. Motion to Exclude Haas (Dkt. Nos. 238/242) 16 Plaintiff seeks to exclude opinions of Mr. Haas reflected in paragraphs 15 (fourth bullet), 17 16, 84, 90-100, 106-07, 111, 146, 160, 173-74, 183, 188-89, and 191. 18 Impinj seeks a broad exclusion of Mr. Haas’ opinions because he allegedly omits 19 consideration of Impinj’s anticipated loss of sales and profits, and defendant’s anticipated gain of 20 sales and profits, from the hypothetical negotiation, which Impinj believes are both relevant 21 factors. Further, Impinj seeks exclusion because Mr. Haas assumes that the parties would have 22 known that Impinj would recover lost profits for the sales not subject to a reasonable royalty, and 23 as a result, Haas claims that Ms. Kindler is “double counting” in her analysis. Said differently, 24 Impinj claims that Mr. Haas’ opinion suggests that by “simply seeking lost profits,” Impinj is 25 precluded from relying on the fact that it would have anticipated losing sales at the hypothetical 26 negotiation, which is the key consideration that would have driven the negotiation, and in effect, 27 results in “double counting.” NXP claims the issue is one of semantics. 1 in this matter should be $0” in paragraph 63 of the Haas Report because this opinion is not based 2 on an affirmative analysis and evaluation of the relevant legal factors. 3 The dispute centers on Ms. Kindler’s opinions that Impinj is entitled to (i) lost profits on 4 57% of defendant’s infringing sales, and her separate (ii) determination that royalty rates that 5 should be applied to any of defendant’s infringing sales for which Impinj is not awarded lost 6 profits—whether that be 43%, 100%, or anywhere in between. 7 The motion is granted wherein Mr. Haas assumes that the hypothetical negotiation would 8 assume the results of the lost profits analysis, e.g. ¶¶ 91-93. Mr. Haas’ opinion that Impinj’s 9 request for lost profits should affect the reasonable royalty analysis is excluded. Further, his 10 references to “double counting” are similarly excluded, e.g. ¶¶ 15, 92, 93, 111, 160, 183. The request 11 to otherwise exclude Mr. Haas’ critique of Ms. Kindler’s analysis is denied except that because 12 Mr. Haas did not conduct an affirmative analysis resulting in lost profits of zero, he may not so 13 opine. He certainly can opine regarding his view on Ms. Kindler’s failure of proof, but that is 14 different than claiming an analysis led to zero dollars in lost profits. 15 The request to seal at Docket No. 242 is granted as for the name of defendant’s customer, 16 defendant’s profit margins, as both confidential and not material to the motion and denied as to 17 Mr. Haas’ deposition testimony. 18 The request to seal at Docket 269-8 is denied. Evidence regarding damages does not 19 warrant sealing. 20 The request to seal at Docket No. 297 is denied. The deposition testimony is not confidential 21 information warranting sealing. 22 Because plaintiff’s motion was overbroad, and to ensure there is no dispute regarding the scope 23 of this ruling, the parties shall meet and confer as to the affected paragraphs of Mr. Haas’ report and 24 shall identify any disagreements no later than Friday, July 7, 2023 at 12:00 p.m. noon. The parties 25 shall avoid raising further disputes with the Court regarding semantics. 26 III. Pending Administrative Motions to Seal 27 Pending before the Court are several administrative motions to seal relating to the parties’ 1 352, 356, 363.) As indicated herein, many of the requests are denied. Not only are the 2 || designations related to anticipated testimony, but in general it does not relate to confidential 3 information that should be sealed during trial. The Court is not inclined to seal any information 4 || which will be made public during the trial. 5 To address these motions efficiently and expeditiously, and to clean up the Court’s docket 6 || in the meantime, the Court orders the parties to review all pending motions, meet and confer, and 7 to file appropriate withdrawals of the motions. For any item which remains an appropriate request 8 || for sealing, the parties shall file a joint chart summarizing that which remains outstanding, as 9 depicted below: 10 Requesting Party and | Description of Pages & Lines Court’s Ruling Sealing Request (Dkt. | Document Sought to 11 No. be Sealed for Administrative (Dkt. No. where a 12 Motion) unredacted document is located) © = 13 The parties shall file the requested chart by Monday, July 17, 2023. The parties shall also

submit a Word version of the chart to the chambers proposed order email address:

15 ygrpo@cand.uscourts/gov, Thereafter, the Court will issue an omnibus order addressing the

a 16 sealing motions. Any request which is denied shall be filed on the public document within five (5)

M7 business days of the order denying the request, including this one. O 18 This terminates Docket Nos. 238, 242, 263, 269-8, 297. 19 IT IS SO ORDERED. 20 ll Dated: 7/3/2023 21 22 VONNE GONZALEZ ROGERS UNITED STATES DISTRICT COURT JUDGE 23 24 25 26 27 28