Case: 23-1101    Document: 18     Page: 1   Filed: 06/03/2024




   United States Court of Appeals
       for the Federal Circuit
                  ______________________

                   ECOFACTOR, INC.,
                    Plaintiff-Appellee

                             v.

                     GOOGLE LLC,
                   Defendant-Appellant
                  ______________________

                        2023-1101
                  ______________________

    Appeal from the United States District Court for the
 Western District of Texas in No. 6:20-cv-00075-ADA, Judge
 Alan D. Albright.
                   ______________________

                  Decided: June 03, 2024
                  ______________________

      BRIAN DAVID LEDAHL, Russ August & Kabat, Los Ange-
 les, CA, argued for plaintiff-appellee. Also represented by
 MINNA CHAN, KRISTOPHER DAVIS, MARC A. FENSTER, REZA
 MIRZAIE, JAMES PICKENS.

     ROBERT A. VAN NEST, Keker, Van Nest & Peters LLP,
 San Francisco, CA, argued for defendant-appellant. Also
 represented by KRISTIN ELIZABETH HUCEK, LEO L. LAM,
 ROBERT ADAM LAURIDSEN, EUGENE M. PAIGE.
                 ______________________

     Before LOURIE, PROST, and REYNA, Circuit Judges.
Case: 23-1101    Document: 18      Page: 2    Filed: 06/03/2024




 2                             ECOFACTOR, INC. v. GOOGLE LLC




     Opinion for the court filed by Circuit Judge REYNA.
  Opinion dissenting-in-part filed by Circuit Judge PROST.
 REYNA, Circuit Judge.
     EcoFactor sued Google in the Western District of Texas
 alleging patent infringement of 

.
 After discovery and resolution of various motions, the case
 was heard by a jury. The jury found that Google infringed
 the asserted claim 5 of the ’327 patent and awarded dam-
 ages to EcoFactor. Google appeals three of the district
 court’s orders: the denial of Google’s motion for summary
 judgment that claim 5 of the ’327 patent was invalid under
 

; the denial of Google’s motion for judgment
 as a matter of law of non-infringement of the ’327 patent;
 and the denial of Google’s motion for a new trial on dam-
 ages. For the following reasons, we affirm.
                         BACKGROUND
                 A. 

.S. Patent No. 8,738,327 (“’327 patent”) relates gen-
 erally to the operation of smart thermostats in computer-
 networked heating and cooling systems (“HVAC systems”).
 The primary recited purpose of the patent is to reduce
 strain on the electricity grid during a period of expected
 high demand through adjustments to the user’s thermostat
 settings that reduce the electricity consumed by the user’s
 HVAC system. ’327 patent at 1:21–27, 9:46–54. Claim 1 of
 the ’327 patent recites a system “for controlling the opera-
 tional status of an HVAC system” where “at least one ther-
 mostat [is] associated with a structure that receives
 temperature measurements from inside the structure.” 

 at 9:26–31. Claim 1 includes an “estimation” limitation
 where “one or more servers receive inside temperatures
 from the thermostat and compare[] the inside tempera-
 tures of the structure and the outside temperatures over
 time to derive an estimation for the rate of change in inside
 temperature of the structure in response to outside
Case: 23-1101    Document: 18     Page: 3    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                             3



 temperature.” 

 at 9:38–45 (emphasis added). Claim 5
 adds that “the estimation [limitation in claim 1] is a pre-
 diction about the future rate of change in temperature in-
 side the structure.” 

 at 9:65–67.
                     B. Procedural History
      EcoFactor, owner of the ’327 patent, sued Google for
 patent infringement over Google’s smart thermostat prod-
 ucts, particularly several Nest thermostats. 1 After discov-
 ery, Google moved for summary judgment that certain
 claims of the ’327 patent (including claim 5) were invalid
 because they were directed to patent ineligible subject mat-
 ter, an abstract idea, under 

. The district
 court denied the motion. J.A. 5046.
     The district court also denied Google’s Daubert motion
 to exclude the opinion of EcoFactor’s damages expert, Mr.
 Kennedy, rejecting Google’s argument that Mr. Kennedy’s
 opinion was unreliable and therefore prejudicial.
 J.A. 2254.
     Following a six-day jury trial, the jury found that
 Google infringed claim 5 of the ’327 patent and awarded
 EcoFactor damages. J.A. 45–49. Google renewed its mo-
 tion for judgment as a matter of law (“JMOL”) of non-in-
 fringement of the ’327 patent, arguing that the accused
 products do not measure, but rather, estimate the temper-
 ature inside the structure and therefore cannot infringe.
 Google also moved for a new trial on damages, arguing that
 the opinion of EcoFactor’s damages expert, Mr. Kennedy,
 was speculative and unreliable such that it should have
 been excluded from trial. The district court denied both
 motions from the bench. J.A. 6662; J.A. 6688.




    1    Google acquired Nest Labs, Inc. prior to the under-
 lying lawsuit.
Case: 23-1101    Document: 18        Page: 4   Filed: 06/03/2024




 4                                ECOFACTOR, INC. v. GOOGLE LLC




    Google appeals.          We     have   jurisdiction   under
 

(a)(1).
                           DISCUSSION
     Google raises three issues on appeal. 2 First, Google
 contends the district court erred in denying Google’s mo-
 tion for summary judgment that claim 5 of the ’327 patent
 was directed to patent ineligible subject matter under
 § 101. Second, Google asserts the district court erred in
 denying Google’s JMOL motion for non-infringement of the
 ’327 patent. Third, Google contends the district court erred
 in denying Google’s motion for a new trial on damages be-
 cause Mr. Kennedy’s damages opinion was based on unre-
 liable methodology. We address each issue in turn.
                      I. Patent Eligibility
     Google appeals the district court’s denial of summary
 judgment that claim 5 of the ’327 patent was patent ineli-
 gible under § 101.
     Section 101 of the Patent Act provides that: “Whoever
 invents or discovers any new and useful process, machine,
 manufacture, or composition of matter, or any new and
 useful improvement thereof, may obtain a patent therefor,
 subject to the conditions and requirements of this title.”
 

. The Supreme Court has articulated a two-
 step test, commonly referred to as the “Alice” test, for ex-
 amining whether a patent claims patent-ineligible subject
 matter. Alice Corp. Pty. Ltd. v. CLS Bank Intern., 

    This appeal was originally consolidated and in-
 cluded an original appeal by EcoFactor and cross-appeal by
 Google. The consolidated appeal contained other patents
 and issues. Prior to oral argument, the parties stipulated
 to the dismissal of the original appeal by EcoFactor, Appeal
 No. 22–1974, leaving only Google’s cross-appeal involving
 the ’327 patent. See Appeal No. 22–1974, ECF No. 59 at 7.
Case: 23-1101     Document: 18     Page: 5    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                               5



 208, 217–18 (2014). At Alice step one, we review whether
 a claim is directed to a patent-ineligible concept, such as an
 abstract idea. 

 At Alice step two, we review whether the
 claim recites elements sufficient to transform the abstract
 idea into a patent-eligible application. 

 at 217–18, 221.
     Prior to trial, Google filed a motion for summary judg-
 ment, arguing that claim 5 of the ’327 patent (among oth-
 ers) was invalid as directed to patent ineligible subject
 matter under § 101. The district court reviewed the mo-
 tion, relying on the Alice inquiry. The district court denied
 the motion and submitted step two of the Alice inquiry to
 the jury. J.A. 5046.
     At trial, the verdict form asked whether Google met its
 burden to prove by clear and convincing evidence that the
 elements of claim 5 of the ’327 patent, when taken individ-
 ually and as an ordered combination, involved activities or
 technology that were well-understood, routine, or conven-
 tional to a skilled artisan at the time of the invention.
 J.A. 47. After hearing testimony and receiving evidence
 from both parties, the jury answered “no” for claim 5 of the
 ’327 patent. J.A. 47. Google filed a post-trial JMOL motion
 repeating its § 101 arguments, which the district court de-
 nied.
     Google now appeals the district court’s denial of sum-
 mary judgment regarding patent ineligibility of claim 5 of
 the ’327 patent, but we have held that a district court’s de-
 nial of summary judgment is not appealable after a trial on
 the merits. Syngenta Crop Prot., LLC v. Willowood, LLC,
 

, 1364 n.7 (Fed. Cir. 2019) (citing Ortiz v. Jor-
 dan, 

, 183–84 (2011)); see also 10 Wright and
 Miller, Federal Practice and Procedure § 2715 (4th ed.) (ex-
 plaining a denial from summary judgment is an order
 “from which no immediate appeal is available”). We have
 explained that an order denying summary judgment is “not
 a judgment” and “does not foreclose trial on the issues on
 which summary judgment was sought;” rather, it is
Case: 23-1101    Document: 18     Page: 6    Filed: 06/03/2024




 6                             ECOFACTOR, INC. v. GOOGLE LLC




 “merely a judge’s determination that genuine issues of ma-
 terial fact exist.” Glaros v. H.H. Robertson Co., 

 (Fed. Cir. 1986) (reasoning that a denial of sum-
 mary judgment “does not settle or even tentatively decide
 anything about the merits of the claim” (citation omitted)).
 Denial of summary judgment decides only one thing—that
 the case should go to trial. 

     At trial, the jury heard testimony from various wit-
 nesses on whether the elements of claim 5 were well-un-
 derstood, routine, or conventional. See, e.g., J.A. 5345–
 5346 (209:20–210:6); J.A. 6373–6374 (1237:15–1238:19);
 J.A. 6415–6416 (1279:1–1280:20); J.A. 6449–6451
 (1313:17–1315:11). Google, however, appeals the order
 denying summary judgment but not the jury verdict of in-
 eligibility. As the Supreme Court has explained, “the full
 record developed in court supersedes the record existing at
 the time of the summary-judgment motion.” Ortiz, 

. Because trial on the merits of the § 101 issue
 was held, the court’s denial of summary judgment is not
 appealable.
                      II. Infringement
     For infringement, the only limitation at issue is
 claim 1’s recitation of a system for controlling the HVAC
 system that includes a thermostat “that receives tempera-
 ture measurements from inside the structure.” ’327 patent
 at 9:26–31. Google alleges that because the accused ther-
 mostat products are designed to be completely enclosed in
 metal, plastic, and/or glass housings, they cannot directly
 measure the surrounding ambient temperature “inside the
 structure” like other thermostats. 3 Appellant Br. 41–44.



     3   The parties agree that the term “ambient tempera-
 ture” refers to the temperature surrounding a particular
 thermostat, i.e., the temperature of the room or structure
Case: 23-1101    Document: 18      Page: 7    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              7



 Google argues that its thermostats can only derive an esti-
 mate of the ambient temperature by measuring only the
 temperature within the thermostat housing itself, which is
 not “inside the structure.” See id. As a result, Google ar-
 gues that the jury’s verdict of infringement is unsupported
 by substantial evidence.
     We review the disposition of motions for JMOL under
 the law of the regional circuit, here the Fifth Circuit. See
 Energy Transp. Grp. Inc. v. William Demant Holding A/S,
 

 (Fed. Cir. 2012). The Fifth Circuit re-
 views de novo the grant or denial of a JMOL motion. Clear-
 Value, Inc. v. Pearl River Polymers, Inc., 

 (Fed. Cir. 2012). Under Fifth Circuit law, a jury’s ver-
 dict is upheld if supported by substantial evidence. Med.
 Care Am., Inc. v. Nat’l Union Fire Ins. Co., 

 (5th Cir. 2003).
      We conclude that the jury’s infringement verdict is sup-
 ported by substantial evidence. EcoFactor’s infringement
 expert testified that the accused thermostat products meet
 the claimed limitation because the thermostats measured
 temperature of the structure and not just the temperature
 within the thermostat housing.              J.A. 5462–63
 (326:20–327:6). EcoFactor’s expert supported his conclu-
 sion with several forms of evidence. EcoFactor’s expert re-
 lied on website guides maintained by Google for the benefit
 of software engineers who develop applications for use with
 Nest thermostats. One website page states that the Nest
 thermostats measure the “[a]mbient temperature,” defined
 as the “temperature measured near the thermostat”—not
 just within the thermostat. J.A. 10429 (emphasis added).
 Another website page explains that the temperature sen-
 sors of certain Nest products measure ambient room tem-
 perature. J.A. 10888. EcoFactor’s expert testified that this


 in which the thermostat is placed. Appellant Br. 13, 43;
 Appellee Br. 1, 9–10.
Case: 23-1101    Document: 18     Page: 8    Filed: 06/03/2024




 8                             ECOFACTOR, INC. v. GOOGLE LLC




 evidence shows that the Nest thermostat “devices have
 temperature measurements near the thermostat” that
 measures the “inside temperature” of the structure and are
 not limited to measuring temperature inside the thermo-
 stat housing. J.A. 5453–55 (317:17–319:14). EcoFactor’s
 expert cited Google’s source code for the accused products
 and demonstrated where it described the function of the
 accused products to measure the surrounding temperature.
 J.A. 5456–60 (320:18–324:14).
     Google’s experts conceded the substance of EcoFactor’s
 evidence on cross-examination. Google’s non-infringement
 expert agreed that, according to Google’s website pages, the
 current ambient temperature in the room is measured by
 the Nest thermostat’s internal sensors. J.A. 6154–55
 (1018:9–1020:13). Another Google expert witness agreed
 that the accused thermostat products contain temperature
 sensors that measure the temperature inside customer
 homes. J.A. 5945–46 (809:4–809:17).
     The expert testimony from both parties, documentary
 evidence, and source code information demonstrating that
 the accused products measure temperature of the sur-
 rounding structure (and not just the housing) is substantial
 evidence. See In re Mouttet, 

 (Fed. Cir.
 2012) (explaining substantial evidence is “relevant evi-
 dence as a reasonable mind might accept as adequate to
 support a conclusion”).
    In conclusion, the jury’s infringement verdict that the
 accused Nest thermostat products satisfy the claim lan-
 guage of “receives temperature measurements from inside
 the structure” is supported by substantial evidence.
                       III. Damages
     Google argues that the district court abused its discre-
 tion in denying its Rule 59 motion for a new trial on dam-
 ages. Appellant Br. 30. According to Google, a new trial
 on damages was warranted because the initial trial was
Case: 23-1101     Document: 18      Page: 9    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                                9



 unfair or marred by prejudicial error. Id. at 25; J.A. 6689
 (91:4–8). The alleged error was the district court’s eviden-
 tiary ruling that the opinion of EcoFactor’s damages ex-
 pert, Mr. Kennedy, was admissible. Appellant Br. 25, 30.
 Google argues that Mr. Kennedy’s damages opinion should
 have been excluded from trial because it lacked any relia-
 ble methodology or underlying calculations. Id. at 30.
 Google also argues that Mr. Kennedy’s opinion should have
 been excluded for lack of comparability and apportionment.
 Id. at 34. We address each argument in turn.
     We review the denial of a motion for a new trial under
 regional circuit law. Wordtech Sys., Inc. v. Integrated Net-
 works Sols., Inc., 

 (Fed. Cir. 2010). The
 Fifth Circuit reviews a district court’s denial of a motion for
 a new trial for an abuse of discretion. Fornesa v. Fifth
 Third Mortg. Co., 

 (5th Cir. 2018). A new
 trial may be granted if the district court finds that “the ver-
 dict is against the weight of the evidence, the damages
 awarded are excessive, the trial was unfair, or prejudicial
 error was committed.” Seidman v. Am. Airlines, Inc., 

 (5th Cir. 1991) (citation omitted); Fed. R.
 Civ. P. 59. “Courts do not grant new trials unless it is rea-
 sonably clear that prejudicial error has crept into the rec-
 ord or that substantial justice has not been done, and the
 burden of showing harmful error rests on the party seeking
 the new trial.” Jordan v. Maxfield & Oberton Holdings,
 L.L.C., 

 (5th Cir. 2020) (citation omitted).
      Here, Mr. Kennedy used the hypothetical negotiation
 approach for calculating reasonable royalty damages under
 

. This approach “necessarily involves an el-
 ement of approximation and uncertainty.” Lucent Techs.,
 Inc. v. Gateway, Inc., 

 (Fed. Cir. 2009)
 (citation omitted); see also generally 2 Janice M. Mueller,
 Mueller on Patent Enforcement § 20.04(a) at 869–70 (rev.
 ed. 2019). According to Mr. Kennedy, EcoFactor would
 have entered the hypothetical negotiation with the
Case: 23-1101    Document: 18        Page: 10   Filed: 06/03/2024




 10                             ECOFACTOR, INC. v. GOOGLE LLC




 expectation of receiving a royalty in the amount of $X 4 per
 unit and would have requested that from Google.
 J.A. 1277. Based on this $X rate, Mr. Kennedy calculated
 his proposed damages amount of $Y 5.            J.A. 5740
 (604:14–17).
                                A.
     Google argues that Mr. Kennedy’s damages model was
 speculative and conclusory. Appellant Br. 31. Specifically,
 Google argues that Mr. Kennedy’s proposed $X royalty rate
 was “plucked . . . out of nowhere.” Id. at 34 (citation omit-
 ted). We disagree.
     “[W]hile all [damages] approximations involve some
 degree of uncertainty, the admissibility inquiry centers on
 whether the methodology employed is reliable.” Summit 6,
 LLC v. Samsung Elecs. Co., 

 (Fed. Cir.
 2015). This includes whether a damages expert’s testi-
 mony is tied to the particular facts of the case. Virnetx, Inc.
 v. Cisco Sys., Inc., 

, 1333–34 (Fed. Cir. 2014);
 Whitserve, LLC v. Comput. Packages, Inc., 

 (Fed. Cir. 2012); Lucent Techs., 

. Testi-
 mony is inadmissible when it is based only on speculation
 or guesswork, such that the jury is left to fill in the gaps
 when calculating a damages award. Whitserve, 694 F.3d at
 30–33 (holding that testimony was conclusory and specula-
 tive when expert did not explain how lump sum amounts
 could be converted to a reasonable royalty rate); Wordtech,
 

 (holding that expert’s reliance on two




      4  The amount of the per-unit royalty rate is confiden-
 tial business information subject to a protective order, and
 as such, is not recited in this opinion.
      5  The amount of EcoFactor’s proposed damages
 award is confidential business information subject to a pro-
 tective order, and as such, is not recited in this opinion.
Case: 23-1101    Document: 18      Page: 11     Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                               11



 lump licenses was inappropriate when neither explained
 how the lump-sum amounts were calculated).
     Far from plucking the $X royalty rate from nowhere,
 Mr. Kennedy based this rate on the following admissible
 evidence: three license agreements and the testimony of
 EcoFactor’s CEO, Mr. Habib. Turning first to the agree-
 ments, Mr. Kennedy relied on three license agreements
 EcoFactor entered into with third-party smart thermostat
 manufacturers—the Schneider and Daikin licenses in
 2020, and the Johnson license in 2021. J.A. 10389–399;
 J.A. 10400–410; J.A. 10411–419. Each of these agreements
 included the same $X royalty rate at issue here. Each li-
 cense agreement provided in a whereas clause that the li-
 censee would pay EcoFactor a lump sum amount “set forth
 in this Agreement based on what EcoFactor believes is a
 reasonable royalty calculation of [$X] per-unit for . . . esti-
 mated past and [] projected future sales of products ac-
 cused of infringement in the Litigation.” J.A. 10389
 (emphasis added); J.A 10400; J.A. 10411. Thus, as Mr.
 Kennedy testified at trial, the $X royalty rate was “specifi-
 cally spelled out in the license agreement[s].” J.A. 5764
 (628:2–3).
      Mr. Kennedy then relied on the testimony of EcoFac-
 tor’s CEO, Mr. Habib, who signed the three license agree-
 ments on behalf of EcoFactor. J.A. 5794 (658:17–18);
 J.A. 5666 (530:23–25); J.A. 5669 (533:6–8). Mr. Habib tes-
 tified that he had seven years in the industry, an under-
 standing of the market, and “what is reasonable for the
 technologies that [EcoFactor] ha[s].”           J.A. 5670
 (534:22–25). He then testified that the lump sums con-
 tained in each of the three license agreements were based
 on the $X royalty rate. J.A. 5672, 536:17–18 (“[M]y under-
 standing was that all of it is based on [$X] per infringing
 unit.”). He testified that while he was shielded from the
 licensees’ confidential sales information, he understood
 that EcoFactor calculated each of the three licenses’ lump
 sums using the $X royalty rate and the past and future
Case: 23-1101    Document: 18     Page: 12    Filed: 06/03/2024




 12                             ECOFACTOR, INC. v. GOOGLE LLC




 projected sales for each licensee. J.A. 5798 (662:2–5); J.A.
 5670 (534:4–10). He also testified that despite being
 shielded from the licensees’ confidential sales numbers, he
 believed, based on his understanding of the market, that
 the lump sums reasonably reflected the licensees’ sales.
 J.A. 5672 (536:14–24). According to Mr. Habib, there were
 “large players” and high barriers to entry in the smart ther-
 mostat and smart HVAC control industry, and the licen-
 sees were “relatively new or more recent.” J.A. 5672
 (536:19–24). Thus, “it ma[de] sense that their sale num-
 bers would be low since they’d recently started.” J.A. 5672
 (536:23–24). He testified that the $X royalty rate in each
 of the three license agreements was accepted by the par-
 ties. J.A. 5671 (535:5–11) (“So, you know, if three compa-
 nies were willing to accept it, then yeah. That further
 made it clear to me that it was a reasonable royalty rate
 that was being accepted by counterparties.”).
     Finally, in support of Mr. Kennedy’s proposed $X roy-
 alty rate, EcoFactor introduced at trial an email chain be-
 tween EcoFactor and Johnson concerning the $X royalty
 rate. J.A. 10797–99; J.A. 6278 (1142:3–10). In the chain,
 which was dated a few months before the parties signed the
 license agreement, the parties discuss the $X royalty rate.
 J.A. 10797–99. Johnson notes that “[w]e are applying the
 [$X rate] to the time period” identified by EcoFactor.
 J.A. 10798.
     In light of the three license agreements, Mr. Habib’s
 testimony, and the EcoFactor-Johnson email chain, we de-
 termine that Mr. Kennedy’s damages opinion concerning
 the $X royalty rate was sufficiently tied to the facts of the
 case and thus admissible. See Finjan, Inc. v. Secure Com-
 puting Corp., 

 (Fed. Cir. 2010); C & F
 Packing Co., v. IBP, Inc., 

, 1304–05 (Fed. Cir.
 2000). And based on this context, the “jury was entitled to
 hear the expert testimony” from Mr. Kennedy concerning
 the $X royalty rate and “decide for itself what to accept or
 reject.” Pavo Sols. LLC v. Kingston Tech. Co., 35 F.4th
Case: 23-1101    Document: 18     Page: 13    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                             13



 1367, 1379 (Fed. Cir. 2022) (citation omitted). That is ex-
 actly what the jury did in this case. The jury heard Mr.
 Kennedy’s testimony and Google’s extensive cross-exami-
 nation concerning Mr. Kennedy’s understanding of the
 three license agreements, his reliance on Mr. Habib’s testi-
 mony, and testimony concerning the emails between Eco-
 Factor and Johnson about the $X royalty rate. J.A.
 5793–5812 (657:4–676:2); J.A. 5794 (658:17–18); J.A. 5667
 (531:8–25); J.A. 5668–5670 (532:3–534:3); J.A. 6278–6280
 (1142:6–1144:19). Ultimately, the jury returned a verdict
 of $20,019,300, which represents significantly less than
 Mr. Kennedy’s proposed damages amount of $Y that would
 have resulted from applying the $X royalty rate to Google’s
 past sales.
      Google argues that Mr. Kennedy’s testimony is unreli-
 able because there is no evidence that the parties to the
 three license agreements actually applied the $X royalty
 rate. To the contrary. First, the three admissible license
 agreements each disclose that EcoFactor believed that the
 lump sums in each license was “based on” the $X royalty
 rate. Additionally, in its whereas clause, the Schneider li-
 cense agreement, unlike the Johnson and Daikin agree-
 ments, states that “nothing in this clause should be
 interpreted as agreement by Schneider that [$X] per unit
 is a reasonable royalty.” J.A. 10400. This clause, included
 by Schneider, speaks to its belief that $X may not have been
 reasonable but it does not speak to whether $X was actu-
 ally applied in arriving at the lump sum. Arguably, this
 provision, when read in context, could also mean that the
 $X royalty rate was applied by EcoFactor and Schneider.
 If Schneider did not believe that the $X royalty rate was
 actually being applied, it could have said such in the agree-
 ment. But Schneider did not. Finally, as noted above,
 Johnson noted in an email chain with EcoFactor that it was
 “applying” the $X royalty rate. How much weight should
 be given to the provisions in the license agreements, includ-
 ing whether they are “self-serving” as Google claims, and
Case: 23-1101    Document: 18        Page: 14   Filed: 06/03/2024




 14                             ECOFACTOR, INC. v. GOOGLE LLC




 the EcoFactor-Johnson email is a question for the jury. See
 Pavo, 35 F.4th at 1379; Finjan, 

; C & F
 Packing, 

. 6
     To conclude, we determine that Mr. Kennedy’s opinion
 concerning the $X royalty rate was sufficiently reliable for
 admissibility purposes. For this reason, we hold that the
 district court did not abuse its discretion in denying
 Google’s motion for a new trial on damages.
                                B.
     Google argues that Mr. Kennedy’s damages testimony
 should have also been excluded from trial for a lack of com-
 parability and apportionment. Appellant Br. 34. Google
 does not dispute the technical comparability between the
 three licenses and Mr. Kennedy’s hypothetically negotiated
 agreement. Nor could it. Mr. Kennedy relied on the testi-
 mony of EcoFactor’s technical expert, Mr. De la Iglesia, for
 his opinion that the three license agreements were



      6  The dissent relies on a statement in the body of the
 of the Schneider and Daikin license agreements to support
 its position that the parties did not apply the $X royalty
 rate contained in these two license agreements’ whereas
 clauses. See Dissent 3–6. This statement provides that the
 agreed to lump sum “does not reflect or constitute a roy-
 alty.” J.A. 10391; J.A. 10402. That the lump sum amount
 is not a royalty does not mean the parties did not use the
 $X royalty rate discussed in the agreements to arrive at the
 lump sum amount. But even if we were to set aside these
 two license agreements, the Johnson license agreement
 alone would suffice. As Google’s own expert agreed at trial,
 “just one” license agreement can be sufficient to support a
 damages opinion. J.A. 6269 (1133:10–14). This assertion
 comports with our damages precedent, which does not de-
 mand “absolute precision” but may involve some degree of
 approximation and uncertainty. Virnetx, 

.
Case: 23-1101    Document: 18      Page: 15    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              15



 technically comparable to the hypothetically negotiated li-
 cense. J.A. 5578–5583 (442:22–447:2); J.A. 5763
 (627:7–23); J.A. 5768 (632:7–19). Google’s experts did not
 rebut Mr. De la Iglesia’s opinion on this issue at trial. J.A.
 6268–6270 (1132:4–1134:5).
     Rather, Google challenges Mr. Kennedy’s economic
 comparability analysis of the three licenses and the hypo-
 thetically negotiated agreement. Appellant Br. 36–37; Re-
 ply Br. 9.     According to Google, the three license
 agreements were for EcoFactor’s entire patent portfolio
 and Mr. Kennedy failed to account for the value of the ’327
 patent within that portfolio. Appellant Br. 36. We disa-
 gree.
     Damages owed to the patentee must reflect the value
 of only the patented improvement—called apportionment.
 Omega Pats., LLC v. CalAmp Corp., 

 (Fed. Cir. 2021). If a sufficiently comparable license is used
 for determining the appropriate reasonable royalty rate,
 further apportionment may not be required because the
 comparable license has built-in apportionment. 

. “Built-in apportionment effectively assumes that the
 negotiators of a comparable license settled on a royalty rate
 and royalty base combination embodying the value of the
 asserted patent.” 

 (citation omitted). “For built-in ap-
 portionment to apply, the license must be sufficiently com-
 parable in that principles of apportionment were effectively
 baked into the purportedly comparable license.” 

 (cita-
 tion omitted). Part of this comparability analysis requires
 an expert to account “for differences in the technologies and
 economic circumstances of the contracting parties” to the
 past licenses and to the hypothetical negotiation at issue.
 Finjan, 626 F.3d at 1211–12.
     The degree of comparability of license agreements is a
 “factual issue[] best addressed by cross examination and
 not by exclusion.” ActiveVideo Networks, Inc. v. Verizon
 Commc’ns, Inc., 

 (Fed. Cir. 2012); Bio-
Case: 23-1101    Document: 18      Page: 16     Filed: 06/03/2024




 16                             ECOFACTOR, INC. v. GOOGLE LLC




 Rad Lab’ys, Inc. v. 10X Genomics Inc., 

, 1373
 (Fed. Cir. 2020); Ericsson, Inc. v. D-Link Sys., Inc., 

 (Fed. Cir. 2014); Finjan, 

. For
 example, in Bio-Rad, we concluded that there was no abuse
 of discretion in allowing an expert to testify about three li-
 censes, even though one of the licenses was ultimately not
 proven to be technically comparable to the hypothetically
 negotiated license. 967 F.3d at 1374 (holding that the “‘de-
 gree of comparability’ was appropriately left for the jury to
 decide”).
      Here, Mr. Kennedy sufficiently showed, for purposes of
 admissibility, that the three license agreements were eco-
 nomically comparable to the hypothetically negotiated
 agreement. Mr. Kennedy acknowledged that, based on Mr.
 De la Iglesia’s unrebutted testimony, the Schneider and
 Daikin licenses list seven technically comparable asserted
 patents, including the ’327 patent at issue in the hypothet-
 ically negotiated agreement. See J.A. 10398; J.A. 10409.
 He also noted that the Johnson license did not list the ’327
 patent as an asserted patent but listed four others that cov-
 ered the same interrelated smart thermostat technologies.
 J.A. 10411; J.A. 1276. Finally, Mr. Kennedy acknowledged
 that the three licenses also covered patents in EcoFactor’s
 portfolio that were not asserted in the underlying litigation
 facing Johnson, Schneider, and Daikin. J.A. 10398;
 J.A. 10409; J.A. 10411; J.A. 1275–76.
     Mr. Kennedy accounted for such differences. Mr. Ken-
 nedy testified that in arriving at the $X royalty rate in a
 hypothetical negotiation, Google would argue that the
 three license agreements included EcoFactor’s portfolio,
 not just the ’327 patent, and thus the $X royalty rate
 should be decreased. J.A. 5767 (631:19–23). Mr. Kennedy
 then provided that the three license agreements reflect a
 settlement and thus the $X royalty rate reflects a risk that
 that EcoFactor’s patents would be found not infringed or
 invalid. J.A. 1276. According to Mr. Kennedy, this consid-
 eration would not be present at the hypothetical
Case: 23-1101    Document: 18     Page: 17   Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                            17



 negotiation between EcoFactor and Google, since the as-
 sumption is that the ’327 patent was infringed and valid.
 J.A. 1276. As a result, this point would place upward pres-
 sure on the negotiated rate.
      The three licenses aside, Mr. Kennedy separately
 grounded his apportionment opinion on underlying inter-
 nal profit and survey data from Google. Mr. Kennedy tes-
 tified that, based on underlying customer surveys from
 Google and based on EcoFactor’s technical expert’s testi-
 mony, the infringed technology at issue in this case at-
 tributed to Z% 7 of the profits for the infringed products.
 J.A. 5755–5758 (619:16–622:13); J.A. 5775–5777
 (639:22–641:7). Based on this data, Mr. Kennedy calcu-
 lated the amount of profit per unit that could be attributed
 to the ’327 patent, which was more than double the $X roy-
 alty rate. J.A. 5755–5758 (619:16–622:13). According to
 Mr. Kennedy, this would also place upward pressure on the
 negotiated rate at the hypothetical negotiation. J.A. 5780
 (644:6–7) (“And that’s EcoFactor’s response, saying it
 should actually be a lot higher.”). Mr. Kennedy thus con-
 cluded that the $X royalty rate “would be a very reasonable
 and conservative first offer.” J.A. 5779 (643: 17–18). This
 testimony is additional evidence for the jury to consider
 and weigh when calculating a damages award. C & F Pack-
 ing, 

; ResQNet.com, Inc., v. Lansa, Inc.,
 

 (Fed. Cir. 2010) (“At all times, the dam-
 ages inquiry must concentrate on compensation for the eco-
 nomic harm caused by infringement of the claimed
 invention.”).
     Based on this evidence, we conclude that the district
 court did not abuse its discretion in declining to grant a
 new trial on damages. Mr. Kennedy’s damages opinion


     7    The percentage amount is confidential business in-
 formation subject to a protective order, and as such, is not
 recited in this opinion.
Case: 23-1101    Document: 18     Page: 18    Filed: 06/03/2024




 18                             ECOFACTOR, INC. v. GOOGLE LLC




 relied on sufficiently comparable licenses and his opinion
 sufficiently apportioned the value of the ’327 patent for the
 issue to be presented to the jury.
     Supporting our conclusion is ActiveVideo. There, the
 damages expert relied on an agreement that included the
 patents at issue and other software services without any
 alleged attempt to “disaggregate the value of the patent li-
 cense from the value of the services.” 694 F.3d at 1333 (ci-
 tation omitted). We held that there was no error in failing
 to exclude the expert’s testimony because the degree of
 comparability and “any failure on the part of [the] expert
 to control for certain variables are factual issues best ad-
 dressed by cross examination and not by exclusion.” Id.
 Here, Mr. Kennedy went further than the ActiveVideo ex-
 pert by sufficiently accounting for the economic differences
 between the patents in the three license agreements (as-
 serted and non-asserted) and the hypothetically negotiated
 agreement. And like in ActiveVideo, if there were any fail-
 ures to control for certain variables in comparability, these
 factual issues were for the jury to decide. Id.
     Contrary to Google’s position, our case law does not
 compel a contrary result. In Omega, we remanded for a
 new trial where the expert “merely identified . . . differ-
 ences” between the patents in the licenses and the patents
 in the hypothetical negotiation and did not distinguish
 such facts. 13 F.4th at 1380–81. In Apple, two of the three
 license agreements relied on by the expert did not list the
 subject patent all, and the third license listed the subject
 patent as a non-asserted patent in a long list of “hundreds
 of Non-Asserted patents.” Apple Inc. v. Wi-LAN Inc., 

 (Fed. Cir. 2022). We determined that there
 was no record evidence supporting the expert’s assumption
 that the subject patent was a “key patent” in these three
 licenses. 

     Unlike in Omega and Apple, here we have two of the
 three licenses at issue explicitly listing the ’327 patent as
Case: 23-1101    Document: 18      Page: 19    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              19



 an “asserted patent.” J.A. 10398; J.A. 10409. Additionally,
 Mr. Kennedy addressed and distinguished the remaining
 patents discussed in the license agreements. He testified
 that at the hypothetical negotiation, Google would empha-
 size the downward pressure that these patents would have
 on the $X royalty rate. Mr. Kennedy then testified that
 EcoFactor would note upward pressure on the $X royalty
 rate by assuming that the ’327 patent was valid and in-
 fringed. And, unlike in these cases, Mr. Kennedy sepa-
 rately rooted his apportionment analysis on underlying
 internal profit and survey data from Google. As previously
 noted, based on this data, Mr. Kennedy was able to deter-
 mine that the $X royalty rate was a conservative amount
 attributable to the ’327 patent.
     Google loses sight of the issue on appeal and the appli-
 cable standard of review. Our focus is on the admissibility
 of Mr. Kennedy’s damages testimony, and we assess the
 district court’s determination of this issue under the highly
 deferential abuse of discretion standard. “Credibility de-
 terminations, the weighing of the evidence, and the draw-
 ing of legitimate inferences from the facts” are jury
 functions, not those of a trial judge, and certainly not of an
 appellate judge. Reeves v. Sanderson Plumbing Prods.,
 Inc., 

 (2000) (citation omitted). If the
 standard for admissibility is raised too high, then the trial
 judge no longer acts as a gatekeeper but assumes the role
 of the jury.
      Based on the record before us, we conclude that the dis-
 trict court did not abuse its discretion when it denied
 Google’s motion for a new trial. Google has not shown that
 the district court’s decision to admit Mr. Kennedy’s dam-
 ages opinion resulted in prejudicial error or a substantial
 injustice requiring a new trial on damages.
                         CONCLUSION
     We have considered Google’s remaining arguments and
 find them unpersuasive. We reject Google’s attempt to
Case: 23-1101    Document: 18     Page: 20    Filed: 06/03/2024




 20                             ECOFACTOR, INC. v. GOOGLE LLC




 appeal the district court’s denial of summary judgment.
 We affirm the district court’s post-trial denials of Google’s
 motion for JMOL of non-infringement and Google’s motion
 for a new trial on damages.
                        AFFIRMED
                            COSTS
 No costs.
Case: 23-1101    Document: 18     Page: 21   Filed: 06/03/2024




    United States Court of Appeals
        for the Federal Circuit
                  ______________________

                   ECOFACTOR, INC.,
                    Plaintiff-Appellee

                             v.

                     GOOGLE LLC,
                   Defendant-Appellant
                  ______________________

                        2023-1101
                  ______________________

    Appeal from the United States District Court for the
 Western District of Texas in No. 6:20-cv-00075-ADA, Judge
 Alan D. Albright.
                  ______________________

 PROST, Circuit Judge, dissenting-in-part.
     In recent years, our court has made some progress in
 clarifying important questions related to damages for pa-
 tent infringement. Such clarifications relate to deriving a
 reasonable royalty from a lump-sum license and requiring
 the patentee to confine its damages to the value of the pa-
 tented technology. Unfortunately, the majority opinion
 here at best muddles our precedent and at worst contra-
 dicts it. I therefore respectfully dissent from the decision
Case: 23-1101     Document: 18      Page: 22     Filed: 06/03/2024




 2                               ECOFACTOR, INC. v. GOOGLE LLC




 to affirm the district court’s denial of Google’s motion for a
 new trial. 1
     Google argues that (1) Mr. Kennedy, EcoFactor’s dam-
 ages expert, calculated an $X royalty rate 2 from the Schnei-
 der, Daikin, and Johnson lump-sum licenses in an
 unreliable way; and (2) the $X rate in any event did not
 reflect the value of the ’327 patent (as distinct from that of
 other patents covered by those licenses). Google is right on
 both counts. The district court therefore, in my view,
 abused its discretion by not granting a new damages trial
 given Mr. Kennedy’s flawed testimony. 3
                                I
      Mr. Kennedy’s $X rate rests on EcoFactor’s self-serv-
 ing, unilateral “recitals” of its “beliefs” in the license agree-
 ments. These recitals are not only directly refuted by two
 of those same agreements; they also have no other support
 (e.g., sales data or other background testimony) to back
 them up. Our law does not allow damages to be so easily
 manufactured.
     When deriving reasonable royalties from lump-sum li-
 censes, we have emphasized that “lump sum payments . . .
 should not support running royalty rates without testi-
 mony explaining how they apply to the facts of the case.”


     1    I join the other aspects of the majority’s decision.
     2    Because the specific per-unit royalty rate that Mr.
 Kennedy uses has been designated confidential, I use $X to
 refer to his rate.
     3    When reviewing a district court’s exercise of discre-
 tion on a critical, often-complicated evidentiary decision
 such as a damages-expert Daubert, it usually helps to see
 the court’s explanation for its decision. Here, at both the
 Daubert stage and in the context of Google’s new-trial mo-
 tion, the district court gave no explanation. J.A. 2254,
 6687–89.
Case: 23-1101    Document: 18     Page: 23    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              3



 Whitserve, LLC v. Comput. Packages, Inc., 

 (Fed. Cir. 2012); see also Lucent Techs., Inc. v. Gateway,
 Inc., 

 (Fed. Cir. 2009) (requiring that
 “some basis for comparison must exist in the evidence pre-
 sented to the jury”). We have vacated damages awards
 where the derivation of a reasonable royalty from a lump
 sum was “incompatible with the . . . agreement as a whole,”
 MLC Intell. Prop., LLC v. Micron Tech., Inc., 

 (Fed. Cir. 2021), where there was no testimony ex-
 plaining how lump-sum “payments could be converted to a
 royalty rate,” Whitserve, 

, and where “[n]ei-
 ther license describe[d] how the parties calculated each
 lump sum,” Wordtech Sys., Inc. v. Integrated Network Sols.,
 Inc., 

 (Fed. Cir. 2010). Mr. Kennedy’s
 $X rate repeats the same fatal errors we identified in MLC,
 Whitserve, and Wordtech. As in those cases, the licenses
 here are for a lump-sum amount with no record evidence
 supporting a calculation of a royalty rate.
      Consider what these licenses do (and do not) say.
 Starting with the Schneider license, one preliminary re-
 cital states: “WHEREAS EcoFactor represents that it has
 agreed to the payment set forth in this Agreement based
 on what [it] believes is a reasonable royalty calculation of
 [$X] per-unit for what it has estimated is past and pro-
 jected future sales of products accused of infringement in
 this Litigation.” J.A. 10400 (emphasis added). Yet the
 body of the license (i.e., its substantive and agreed upon
 terms and conditions)—which, unlike the recitals, reflects
 the view of both parties—says that its lump-sum payment
 “is not based upon sales and does not reflect or constitute a
 royalty.” J.A. 10402 (emphasis added).
     The Daikin and Johnson licenses both contain nearly
 identical preliminary recitals about the $X rate.
 J.A. 10389; J.A. 10411. As in the Schneider license, the
 body of the Daikin license—which, again, reflects the view
 of both parties—says that its lump-sum payment “is not
 based upon sales and does not reflect or constitute a
Case: 23-1101     Document: 18      Page: 24    Filed: 06/03/2024




 4                               ECOFACTOR, INC. v. GOOGLE LLC




 royalty.” J.A. 10391 (emphasis added). And while the
 Johnson license lacks a similar refutation of the recital’s
 “belief” (a belief that, again, was EcoFactor’s alone), it still
 offers nothing more than the recital itself to support any
 royalty rate.
     These recitals became Mr. Kennedy’s $X rate.
 J.A. 5764 (628:10–17); J.A. 5769 (633:9–18); J.A. 5772–73
 (636:22–637:4). Mr. Kennedy also relied on the testimony
 of EcoFactor’s CEO, Mr. Habib, as support for this rate.
 J.A. 5794 (658:17–18). But Mr. Habib’s testimony does not
 establish anything beyond his unsupported “understand-
 ing” that these licenses used the $X royalty rate. See
 J.A. 5668–69 (532:23–533:2). When asked during direct
 examination about the basis for his “understanding,” Mr.
 Habib testified that he was not allowed to see any underly-
 ing financial information or sales data. J.A. 5670 (534:4–
 14). Mr. Habib also explained his basis for believing that
 the $X rate was a reasonable rate based on his understand-
 ing of the market, J.A. 5672 (536:14–24), but his market-
 based testimony provided no explanation for converting
 from the lump-sum payments in these licenses to any roy-
 alty rate, let alone the $X royalty rate. 4
     On this record, it’s impossible to establish that these
 lump-sum payments were calculated using any royalty
 rate, let alone the specific $X rate. The self-serving recitals
 reflect only EcoFactor’s transparent attempt to


     4   The only other basis Mr. Habib offers is that the $X
 royalty rate was EcoFactor’s baseline policy for licensing,
 regardless of the number of patents. J.A. 5671 (535:12–
 24). Understandably, neither the majority nor EcoFactor
 rely on the baseline policy as a valid basis for Mr. Habib’s
 understanding that the $X rate was used. See Omega
 Pats., LLC v. CalAmp Corp., 

, 1379–80 (Fed.
 Cir. 2021) (concluding that a comparable license analysis
 using a baseline royalty rate policy is unreliable).
Case: 23-1101    Document: 18     Page: 25    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              5



 manufacture a royalty rate using its “belief.” EcoFactor
 even had to refute its “belief” by agreeing, in the Schneider
 and Daikin licenses, that the lump-sum payment is not a
 royalty. Mr. Kennedy’s assertion that the Schneider and
 Daikin licenses used the $X rate is “incompatible with the
 . . . agreement[s] as a whole.” MLC Intell. Prop., 

. And the Johnson license does not “describe how
 the parties calculated [the] lump sum.” Wordtech Sys., 

.
     Mr. Kennedy cited nothing else showing that the $X
 rate was actually used. He cited no documents, records,
 sales data, or testimony showing any calculation of the
 lump-sum payments or otherwise establishing that these
 licenses used the $X rate. Mr. Habib’s testimony, relying
 on no underlying data, likewise offers no support. EcoFac-
 tor offered no testimony explaining how the lump-sum
 “payments could be converted to a royalty rate.” Whitserve,
 

.
     At bottom, all we have are the recitals of one party’s
 “beliefs” contradicted by mutually agreed upon contractual
 language by both parties. That’s not enough under our law.
      None of the majority’s responses on this issue with-
 stand scrutiny. The majority first insists that the Schnei-
 der and Daikin licenses do not disclaim the $X royalty rate.
 It reasons, “[t]hat the lump sum amount is not a royalty
 does not mean [that] the parties did not use the $X royalty
 rate discussed in the agreements to arrive at the lump sum
 amount.” Maj. 14 n.6 (emphasis in original). If the major-
 ity’s point is that a lump sum is not itself a royalty, then
 fair enough—no one disputes that truism. The issue here,
 however, is whether the lump sum in these licenses reflects
 the application of the $X royalty rate (or, in the majority’s
 words, whether it was used “to arrive at the lump sum
 amount”). And the majority cannot credibly claim that was
 the case when the licenses themselves say that each lump-
 sum payment “is not based upon sales and does not reflect
Case: 23-1101    Document: 18       Page: 26   Filed: 06/03/2024




 6                              ECOFACTOR, INC. v. GOOGLE LLC




 or constitute a royalty.” J.A. 10391, 10402 (emphasis
 added). I’m not sure how these licenses could more clearly
 establish that the lump sums were not calculated using
 royalties.
     The majority next asserts that “the Johnson license
 agreement alone would suffice.” Maj. 14 n.6. But the John-
 son license contains no language describing how its lump-
 sum payment was calculated, see Wordtech Sys., 

, and Mr. Kennedy offered no other basis from
 which he could conclude that the Johnson license used the
 $X rate.
     The majority also cites an email chain that purportedly
 concerns the use of the $X rate in the Johnson license. But
 Mr. Kennedy never discussed this email; EcoFactor intro-
 duced this email during the cross-examination of Google’s
 expert. J.A. 6278 (1142:3–10). The question here is not
 whether any document in the record supports the jury’s
 damages award. We are instead asking whether Mr. Ken-
 nedy’s testimony was so unreliable that it requires a new
 trial. I can’t see how an email Mr. Kennedy never ad-
 dressed supports the reliability of his analysis.
     In the end, Mr. Kennedy conjured the $X rate from
 nothing, and the majority’s treatment of his analysis can-
 not be squared with our law or the facts.
                               II
     Even if these licenses used the $X rate, Mr. Kennedy’s
 analysis has another significant problem: the $X rate does
 not reflect the ’327 patent’s value; rather, it includes the
 value of other patents. Our law is clear that this basic fail-
 ure requires a new trial.
     “When relying on comparable licenses to prove a rea-
 sonable royalty, we require a party to account for differ-
 ences in the technologies and economic circumstances of
 the contracting parties.” Apple Inc. v. Wi-LAN Inc., 

 (Fed. Cir. 2022) (cleaned up). This
Case: 23-1101    Document: 18     Page: 27    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              7



 accounting, although not overly rigid and involving approx-
 imation, requires apportioning to just the value of patent(s)
 in the hypothetical negotiation. Omega Pats., LLC v.
 CalAmp Corp., 

, 1380–81 (Fed. Cir. 2021).
 This apportionment must be tied to the facts of each case.
 

 at 1379–80.
     The licenses here have a broad scope. Each license is
 to “all patents and patent applications (along with patents
 issuing thereon) . . . that are now, or ever come to be, as-
 signed to, owned by, or controlled by EcoFactor.”
 J.A. 10390; J.A. 10401; J.A. 10412. Because each license
 reflects a settlement, they also list the patents asserted in
 each underlying litigation. The Schneider and Daikin li-
 censes list seven asserted patents, including the ’327 pa-
 tent. J.A. 10398, 10409. The Johnson license lists four
 asserted patents; none are the ’327 patent. J.A. 10411.
     When calculating the ’327 patent’s value, Mr. Kennedy
 relied on EcoFactor’s technical expert, Mr. de la Iglesia,
 who compared the asserted patents in each license to the
 ’327 patent and concluded that the asserted patents and
 the ’327 patent were technologically comparable. J.A.
 5578–82 (442:14–446:10). But EcoFactor’s technical expert
 didn’t discuss the remaining patents in each license—the
 non-asserted patents in EcoFactor’s portfolio. 5 Rather, Mr.
 Kennedy explained that since, “in the real world,” “the rest
 of the patents are thrown in usually either for nothing or
 very little additional value,” the presence of these non-as-
 serted patents would place “downward pressure on the roy-
 alty rate” in a hypothetical negotiation over the ’327




     5    Although we don’t know the exact size of EcoFac-
 tor’s portfolio, at least one document suggests that EcoFac-
 tor’s portfolio is over three times larger than the seven
 asserted patents in the Schneider and Daikin licenses.
Case: 23-1101    Document: 18     Page: 28    Filed: 06/03/2024




 8                              ECOFACTOR, INC. v. GOOGLE LLC




 patent. J.A. 5767–68 (631:22–632:2); see also J.A. 5771–72
 (635:19–636:3).
      Mr. Kennedy’s opinion is “untethered to the facts of
 this case.” Apple, 

. His generic testimony
 about “the real world,” general industry practice, and
 “downward pressure” does not account for the impact of
 EcoFactor’s specific non-asserted patents, and we don’t
 know whether any non-asserted patent in EcoFactor’s port-
 folio covers the same technological areas as the asserted
 patents. Mr. Kennedy did not ask the necessary question
 under our law—what effect the specific non-asserted pa-
 tents in EcoFactor’s portfolio would have on the hypothet-
 ical negotiation. 6 Even worse, other evidence in the record
 indicates that the specific non-asserted patents were not
 considered at all. Mr. Habib testified that the remaining
 patents in EcoFactor’s portfolio had no effect on the rate
 used in these licenses because the $X rate was EcoFactor’s
 baseline policy. J.A. 5671 (535:12–24).
     In the end, Mr. Kennedy’s circumstance-agnostic anal-
 ysis is insufficient under our law. Apple, 

 (vacating a verdict where the same expert as in this case,
 Mr. Kennedy, testified that excluding the non-asserted pa-
 tents in a portfolio license would reduce the royalty rate by
 25 percent as a matter of industry practice); Omega Pats.,
 

 (vacating a verdict where the proposed
 $5.00 royalty rate was the same for any number of patents);
 MLC Intell. Prop., 

 (vacating a verdict




     6  It would not be difficult for EcoFactor to offer an
 answer. For example, Mr. de la Iglesia could have deter-
 mined that the non-asserted patents in the Schneider, Dai-
 kin, and Johnson licenses have no technological overlap
 with the ’327 patent and concluded that the non-asserted
 patents added only nominal value to the license.
Case: 23-1101    Document: 18     Page: 29    Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                              9



 where no evidence or explanation supported using the roy-
 alty rate in a forty-one-patent license for a single patent).
      The majority, in excusing Mr. Kennedy’s failure to
 properly apportion, ignores our law’s requirements. It
 merely states that Mr. Kennedy “acknowledged that the
 three licenses also covered patents not in EcoFactor’s port-
 folio.” Maj. 16. But the majority ignores the key fail-
 ure—Mr. Kennedy failed to account for the impact of the
 specific remaining patents in EcoFactor’s portfolio, other
 than by referencing a generic “downward pressure.” Nei-
 ther Mr. Kennedy nor Mr. de la Iglesia tied this “downward
 pressure” to the specific non-asserted patents, and Mr.
 Habib affirmatively testified that EcoFactor’s practice was
 to use the same $X rate regardless of the number of pa-
 tents. In these circumstances, I “fail to see how this pa-
 tent/claim-independent       approach      accounts      for
 apportionment.” Omega Pats., 

.
      The majority also relies on ActiveVideo Networks, Inc.
 v. Verizon Communications, Inc., 

 (Fed. Cir.
 2012), for its conclusion. But in ActiveVideo, we concluded
 that “disagreements . . . with the conclusions reached by
 ActiveVideo’s experts and the factual assumptions and con-
 siderations underlying those conclusions” were “factual is-
 sues best addressed by cross examination and not
 exclusion.” 

. Our conclusion presupposed that
 “the methodology is sound” and that “the evidence relied
 upon [is] sufficiently related to the case at hand.” i4i Ltd.
 P’Ship v. Microsoft Corp., 

 (Fed. Cir.
 2010), aff’d, 

 (2011). As I explained above,
 that’s far from the case here. And to the extent the major-
 ity relies on one party’s characterization in ActiveVideo of
 the expert’s opinion as failing to “disaggregate the value of
 the patent license from the value of the services,” Maj. 18,
 we did not adopt that characterization of the expert’s opin-
 ion, see ActiveVideo, 694 F.3d at 1333. Additionally, any
 suggestion that disaggregating the value of the patented
 technology from the overall value of a license is not
Case: 23-1101    Document: 18      Page: 30    Filed: 06/03/2024




 10                             ECOFACTOR, INC. v. GOOGLE LLC




 required is flatly inconsistent with our law and 140 years
 of Supreme Court precedent. See VirnetX, Inc. v. Cisco
 Sys., Inc., 

 (Fed. Cir. 2014) (“[A] pa-
 tentee ‘must in every case give evidence tending to sepa-
 rate or apportion the defendant’s profits and the patentee’s
 damages between the patented feature and the unpatented
 features.’” (quoting Garretson v. Clark, 

 (1884))).
     To establish a purportedly independent basis for Mr.
 Kennedy’s conclusion, the majority asserts that he demon-
 strated the reasonableness of his $X rate by analyzing
 Google’s profits. Maj. 17. Mr. Kennedy determined what
 profits on Google’s accused Nest thermostats came from
 features that he maintained were attributable to the ’327
 patent. He then split the profits between Google and Eco-
 Factor, using the $X rate as a reasonable basis. See J.A.
 5778 (642:13–17). His methodology isn’t reasonable. Ra-
 ther than offering a cross-check, Mr. Kennedy’s circular
 profit analysis begins and ends with the same “fundamen-
 tally flawed premise”—the $X rate. Uniloc USA, Inc. v.
 Microsoft Corp., 

 (Fed. Cir. 2011).
     Ultimately, the majority’s real concern is that, “[i]f the
 standard for admissibility is raised too high, then the trial
 judge no longer acts as the gatekeeper but assumes the role
 of the jury.” Maj. 19. But we must pay close attention to
 the reliability of the methodology underlying expert testi-
 mony to ensure that the jury can fulfill its proper role as
 the factfinder. See Gen. Elec. Co. v. Joiner, 

 (1997) (emphasizing the district court’s “gatekeeper”
 role in “screening” expert testimony). As the Supreme
 Court has recognized, “the expert’s testimony often will
 rest upon an experience confessedly foreign in kind to the
 jury’s own.” Kumho Tire Co. v. Carmichael, 

 (1999) (cleaned up). Thus, an “effort to assure that the
 specialized testimony is reliable and relevant can help the
 jury evaluate that foreign experience.” 

 Our damages
 law ensures that an expert asks the right questions. Many
Case: 23-1101    Document: 18       Page: 31   Filed: 06/03/2024




 ECOFACTOR, INC. v. GOOGLE LLC                             11



 admissible answers to these questions are possible, and it
 is those answers that are subject to the crucible of cross-
 examination. Mr. Kennedy failed to ask the right ques-
 tions at multiple junctures. The majority’s decision to over-
 look the prejudicial impact of his unreliable testimony
 abdicates its responsibility as a gatekeeper and contradicts
 our precedent.
                              III
      Mr. Kennedy’s analysis is unreliable. His $X rate has
 no basis in the record, and his $X rate does not reflect the
 ’327 patent’s value alone but instead includes the value of
 other patents. Mr. Kennedy’s testimony did not meet the
 baseline standards of admissibility, and therefore the dis-
 trict court abused its discretion by not granting a new trial
 on damages. The majority’s conclusion otherwise departs
 from our law. I respectfully dissent.