TSI's Motion For Judgment Nov As Filed
TSI's Motion For Judgment Nov As Filed
TSI's Motion For Judgment Nov As Filed
7/11/2018 2:20 PM
Donna Kay McKinney
Bexar County District Clerk
Accepted By: Cynthia Gonzales
INTRODUCTION .......................................................................................................................................1
SUMMARY OF ARGUMENT.......................................................................................................................3
ARGUMENT ...........................................................................................................................................14
II. The jury’s massive damages award is contrary to law and cannot stand.....................52
III. Title Source is entitled to JNOV on its claim that HouseCanary breached
Amendment One ..........................................................................................................80
PRAYER ................................................................................................................................................83
Page(s)
Cases
Akin, Gump, Strauss, Hauer & Feld, LLP v. Nat’l Dev. & Res. Corp.,
299 S.W.3d 106 (Tex. 2009)....................................................................................................14
Atlas Copco Tools, Inc. v. Air Power Tool & Hoist, Inc.,
131 S.W.3d 203 (Tex. App.—Fort Worth 2004, pet. denied) ...........................................69, 70
Haase v. Glazner,
62 S.W.3d 795 (Tex. 2001)......................................................................................................46
Hancock v. Variyam,
400 S.W.3d 59 (Tex. 2013)......................................................................................................30
Lakeway Reg’l Med. Ctr., LLC v. Lake Travis Transitional LTCH, LLC,
No. 03-15-00025-CV, 2017 WL 672451 (Tex. App.—Austin Feb. 17, 2017,
pet. denied) ....................................................................................................................... passim
Lesikar v. Rappeport,
33 S.W.3d 282 (Tex. App.—Texarkana 2000, pet. denied) ....................................................29
Rusty’s Weigh Scales & Service, Inc. v. North Texas Scales, Inc.,
314 S.W.3d 105 (Tex. App.—El Paso 2010, no pet.) ........................................................55, 77
Smith v. Nelson,
53 S.W.3d 792 (Tex. App.—Austin 2001, pet. denied)...........................................................74
Smith v. O’Donnell,
288 S.W.3d 417 (Tex. 2009)....................................................................................................76
United Way of San Antonio, Inc. v. Helping Hands Lifeline Found., Inc.,
949 S.W.2d 707 (Tex. App.—San Antonio 1997, writ denied)...............................................70
Constitutional Provisions
Statutes
Rules
Other Authorities
INTRODUCTION
Juries play a foundational role in our judicial system. But the opportunity for review and
relief through a JNOV exists because juries are not infallible, trials can be infected with
misleading and confusing tactics, and proficient storytelling and appeals to emotion can
overshadow evidence and fact. That is why courts bear the ultimate responsibility to ensure any
The punitive damages award in this case is the largest in the history of Bexar County.
Yet the case began with Title Source filing suit for breach of a $5-million contract after
HouseCanary failed to deliver what it had promised—a mobile application that appraisers could
use to complete appraisals in the field—and instead produced, months late, a non-functioning
application that still had “serious bugs” and failed to include necessary components.
HouseCanary, however, twisted a suit about its failure to deliver a product into nearly a billion-
dollar misappropriation verdict in its favor, leading the jury astray with conspiracy theories,
heated diatribes on corporate ethics, and legally irrelevant “evidence.” But verdicts cannot be
based on assertions and arguments of counsel. They cannot be based on speculation and
conjecture. And they cannot be based on emotion and caprice. Because the verdict here has no
The Texas Supreme Court, the Fourth Court of Appeals, and other appellate courts have
ruled that the JNOV tool was created for cases just like this one—and Title Source respectfully
requests that the Court accept the invitation extended by those courts to correct a jury’s mistake
v. Pitzner, 106 S.W.3d 724, 727-29 (Tex. 2003) (reversing and rendering judgment for defendant
because “there [was] legally insufficient evidence to support” the verdict, and the plaintiff’s
Petroleum Chems. Co. v. Waldrop, 75 S.W.3d 549, 554-55 (Tex. App.—San Antonio 2002, no
pet.) (concluding the “trial court erred in denying [defendant’s] motion for JNOV” and reversing
and rendering for defendant, because plaintiff’s “evidence is legally insufficient” to support the
verdict); Aquila Sw. Pipeline, Inc. v. Harmony Expl., Inc., 48 S.W.3d 225, 246 (Tex. App.—San
Antonio 2001, pet. denied) (affirming JNOV where “the underlying factual basis” of an expert’s
damages opinion was “merely speculative”); Hunter Bldgs. & Mfg., L.P. v. MBI Glob., L.L.C.,
436 S.W.3d 9, 22 (Tex. App.—Houston [14th Dist.] 2014, pet. denied) (reversing trade secrets
verdict and rendering judgment for defendants, where “[a]fter reviewing all of the trial evidence
under the applicable standard of review, we conclude this evidence is legally insufficient to
support a finding that the [defendants’] misappropriation of trade secrets proximately caused [the
claimant] to sustain lost-profits damages in the past”); see also Lakeway Reg’l Med. Ctr., LLC v.
Lake Travis Transitional LTCH, LLC, No. 03-15-00025-CV, 2017 WL 672451, at *14 (Tex.
App.—Austin Feb. 17, 2017, pet. denied) (rejecting trade secrets claim as a matter of law where
claimant’s proffered evidence “did not show a sufficient causal link” between any
misappropriation and damages, and “[w]ithout that causal link, [claimant] cannot succeed on a
This motion boils down to a simple request. Title Source asks the Court to look
systematically at the record evidence for each of HouseCanary’s claims and to apply the law to
each. This review of the evidence will demonstrate that HouseCanary’s claims fail on the
Because HouseCanary’s claims crumble under the required scrutiny, this Court must grant
SUMMARY OF ARGUMENT
Although dispassionate review reveals many reasons why the verdict on HouseCanary’s
trade secret claims runs badly afoul of governing law, two are fundamental and independently
First, the heart of HouseCanary’s theory of liability, upon which all its claims depend, is
misappropriation—requiring HouseCanary to establish that it owned bona fide trade secrets, that
Title Source acquired those trade secrets, and that Title Source used them in its automated
valuation model (“AVM”), known as MyAVM—the only model Title Source actually used in its
business. But by HouseCanary’s own repeated admissions, there is no evidence to support these
(Sicklick 3/7/18 AM 31),1 that make up its purported “secret sauce,” (Sicklick 3/9/18 AM 73)—
to Title Source. Instead, HouseCanary’s expert testified that all it provided to Title Source was
the model outputs—“They didn’t have any of the analytics inside of it.” (Rhyne 3/6/18 AM
119). The testimony did not stop there: HouseCanary executives repeatedly testified that they
never gave Title Source any set of algorithms, and that HouseCanary does not “provide
information about how its models operate to its customers.” (Stroud 2/26/18 PM 183). As a
1
Citations of the daily trial transcripts filed concurrently with this motion are referred to by
the speaker, followed by the date and time of the daily edition, and then the page number;
thus, “Sicklick 3/7/18 AM 31.”
logic, it could not replicate what it never had in the first place.
HouseCanary’s own expert testified that Title Source did not use any of HouseCanary’s
trade secrets—that there was no evidence of “any fingerprints, any clues, any reference to any
analyzing the methodology and source code for MyAVM and comparing it to HouseCanary’s
models, HouseCanary’s expert further admitted that MyAVM was not “related to the
HouseCanary AVM.” (Rhyne 3/6/18 AM 119). Faced with this fact, HouseCanary alternatively
asserted that Title Source used a “magic machine” to reverse engineer HouseCanary’s purported
trade secrets (i.e., by allegedly using HouseCanary output data to “train[]” a machine learning
HouseCanary’s theory again was belied by its own expert, who admitted that “Title Source did
not use any HouseCanary data to train MyAVM.” (Rhyne 3/6/18 PM 40; see also id. at 39-40
(admitting that he “know[s] that Title Source did not use HouseCanary data to train its MyAVM
product”)).
Stripped of its main theories—and faced with evidence conclusively establishing the
opposite—HouseCanary’s allegations rest on nothing more than red herrings, bare speculation
that misled the jury, and arguments that Title Source misappropriated generalized industry
concepts that, as a matter of black-letter law, are not trade secrets at all.
a former Title Source employee named Ryan Yang—but those baseless allegations were in fact
about the development of a different, simplistic valuation model that Title Source never moved
forward with, never used, and that had nothing to do with MyAVM. Because HouseCanary’s
actually used in its business), this whole line of accusations was nothing more than misdirection
Similarly, HouseCanary made great hay out of an internal Title Source presentation
suggesting that the “logic” for Title Source’s AVM shared some of the same concepts as
HouseCanary’s model. Putting aside the fact that MyAVM ultimately used a different
methodology, these modeling concepts were standard and basic, long published in literature, and
commonly known by practitioners. Indeed, the record in this case contains an industry journal
article, published in 2003, that outlined the concepts and methods at issue. PX359.
HouseCanary did not invent these generally known industry concepts, and as a matter of law, it
cannot claim them as its trade secrets. HouseCanary’s sensationalized allegations may have
worked to mislead the jury, but a sober review will leave no doubt that HouseCanary’s purported
In short, when the allegations in this case are reviewed within the confines of the law,
there is no evidence that Title Source acquired HouseCanary’s trade secrets, either directly or by
reverse engineering; there is no evidence that Title Source used them in MyAVM; and what was
alleged to have been used—commonly known industry concepts—were not trade secrets at all.
Second, there is no evidence to support the $201.6 million damages award for the alleged
misappropriation—and the incredible claim of Walter Bratic, HouseCanary’s expert, that the
trade secrets would generate this much value over a mere two years of expected use is no
evidence. Not only was Bratic’s damages calculation entirely speculative and untethered to the
facts, it was starkly at odds with them, and with demonstrable market reality. Bratic did not offer
calculation. That calculation was based on the assumption that every allegedly misappropriated
trade secret would be somehow incorporated into one-page AVM “value reports” that Title
Source purportedly expected to generate for two years. (Bratic 3/9/18 PM 126-28; see also id. at
150-51). But Bratic conjured numbers out of thin air, and committed other serious errors, in an
effort to maximize damages over the short two-year period. (E.g., Bratic 3/12/18 PM 80).
As his primary measure of damages, Bratic purported to calculate the value a “reasonably
prudent investor” would have paid for the alleged trade secrets. But Bratic’s claim that Title
Source would value the allegedly misappropriated valuation technology at over $200 million for
two years of expected use flies in the face of the objective evidence and market reality. Multiple
AVMs are available for free. (See Bratic 3/12/18 AM 48-49). HouseCanary itself paid less than
$1 million per year to receive a license for unlimited use of two different AVMs from a third
party, Black Knight—along with all of Black Knights’ data. Id. at 27-29. Most importantly,
Bratic ignored the fact that the parties had already defined the value of the products at issue in
this case: the parties contracted for Title Source to use all of HouseCanary’s technology and
data—including the main and most important product, the never-delivered appraisal app—for $5
To calculate his outlandish figure, Bratic made a series of unfounded assumptions. Bratic
first assumed, contrary to record evidence, that a reasonably prudent investor would be willing to
pay for high-volume use of AVM value reports on a per-use basis. He then calculated the value
per alleged use at an outlandish figure––$11 per use––that has no basis in reality, or in the
evidence. The $11 figure is belied by its own source, which stated that $11 per such use would
have been “cost prohibitive.” DX36; (Bratic 3/9/18 PM 116). Demonstrating just how absurd
contain valuation model outputs, but up to eight pages of additional information) to retail
consumers for only a fraction of this price—with further discounts for large-volume purchasers.
In other words, HouseCanary’s damages calculation assumes that Title Source, as a “reasonably
prudent investor,” would pay several times what HouseCanary actually charges small retail
Bratic reached his $201.6 million valuation by asserting, in a conjecture this Court called
“very speculative” and “a stretch,” (Bratic Robinson 1/24/18 PM 89), that Title Source would
value the alleged trade secrets based on paying $11 per use for nearly 14 million AVM value
reports a year—with 95 percent of these unsubstantiated uses attributed to Quicken Loans Inc., a
separate corporate entity that is not a party to this case. (Bratic 3/9/18 PM 127, 160). But this
assertion was based on the supposed maximum capacity or bandwidth of HouseCanary’s AVM
transmission pipeline—50,000 uses per day—not any evidence of actual use. See DX499;
(Bratic 3/9/18 PM 104; Sicklick 3/7/18 AM 122-123). Bratic asserted, without any evidence but
in a transparent effort to maximize damages, that Title Source would use the maximum
technological capacity every single day. And since Bratic estimated that Title Source at most
executes 2,750 appraisals per day (660,000 appraisals per year divided by 240 business days
(Bratic 3/9/18 PM 94, 110)), he had to turn to implausible speculation about Quicken Loans to
Precisely because Quicken Loans is not a party to this case, there was no competent
evidence of how and to what extent Quicken Loans actually uses an AVM (or value report), nor
an opportunity for proper and informed consideration of that question—and Bratic could only
speculate. Based solely on an internal HouseCanary email suggesting that Quicken Loans’
Quicken Loans’ bankers would generate a value report for each lead “to start th[e] dialogue with
a perspective [sic] customer,” and do so 24 hours per day. (Bratic 3/9/18 PM 106-07). He then
tacked on—based on yet another erroneous assumption—an additional 5,000 uses per day to
respond to customer inquiries, including routine phone calls asking about the interest rate or
generic loan terms (thereby reaching, and indeed even exceeding, the maximum bandwidth of
50,000 total uses per day). Id. at 110; (Bratic 3/12/18 AM 53).
Bratic’s assumptions are both illogical and directly contrary to actual, unrebutted
testimony from Title Source that it expected Quicken Loans to use automated valuation
technology for only a small subset of prospective clients—only for refinance loans, and only for
those that actually reach the loan application stage. Bratic’s assumptions also defy common
sense and would require that every cold call from these bulk lists results in a successful contact
with someone seriously interested in a mortgage; that a value report is generated in every
instance, regardless of whether it makes any sense; and that this uninterrupted streak of success
occurs with 2,000 new customers every hour of the day, including in the middle of the night.
The heart of Bratic’s damages opinion on the alleged value to Title Source—accounting
for 95 percent of the damages award—thus runs contrary to unrebutted testimony and rests on
nothing more than bare speculation about use by a third-party over whom Title Source has no
control. But there is no need for this speculation. Implicitly recognizing that it cannot recover
damages here based on Quicken Loans’ supposed use, HouseCanary has filed a separate lawsuit
against Quicken Loans in Texas federal district court. Basic principles of due process require
that any misappropriation claims based upon Quicken Loans’ alleged use of HouseCanary trade
attributed to Quicken Loans’ supposed use must be eliminated from the verdict here.
In the end, Bratic’s speculative opinions on damages are no evidence at all. Instead, they
are entirely at odds with the objective evidence and demonstrable market reality. As the Texas
Supreme Court has held, “relying on imagination is not justified when objective evidence is
available,” and here the parties’ contract, as well as record evidence on the market price for
unlimited uses of AVM products, constitute “objective evidence that . . . bears directly on ‘the
amount that a person desiring to use the trade secret would be willing to pay for its use.’” Sw.
Energy Prod. Co. v. Berry-Helfand, 491 S.W.3d 699, 721 (Tex. 2016). But such objective
evidence was not used, because it would not support Bratic’s goal of conjuring a massive
damages award.
Texas courts routinely grant JNOV where “the underlying factual basis” of an expert’s
damages opinion is “merely speculative,” Aquila Sw. Pipeline, 48 S.W.3d at 246, or “based on
assumed facts that vary materially from the actual, undisputed facts,” Houston Mercantile Exch.
Corp. v. Dailey Petroleum Corp., 930 S.W.2d 242, 248 (Tex. App.—Houston [14th Dist.] 1996,
no writ). This Court must do the same. As there is no legally sufficient basis for the damages
* * *
Faced with record evidence that gives no quarter to its claims, HouseCanary will likely
argue that this case rests on matters of witness credibility and will emphasize and sensationalize
legally irrelevant evidence. Misdirection is the hallmark of magic; but rigorous application of
law and evidence is the foundation of justice. When this Court systematically reviews each
fails on dispositive legal and evidentiary questions. There is no evidence to support the verdict.
II. HouseCanary’s fraud and breach of contract claims fail along with its TUTSA
claims.
HouseCanary’s tag-along claims for fraud and breach of contract are a mere repackaging
of its misappropriations claims, and with no evidence of misappropriation, they fall together. In
any event, because all of HouseCanary’s claims rest on the same fundamental facts and supposed
injury, the “one satisfaction rule” means that HouseCanary can recover under only one of its
fraud claims are barred as a matter of law by TUTSA’s preemption provision, which plainly
“displaces conflicting tort, restitutionary, and other law of this state providing civil remedies for
misappropriation of a trade secret.” Tex. Civ. Prac. & Rem. Code Ann. § 134A.007(a). A claim
can withstand preemption only if it is “based on facts unrelated to the misappropriation of the
trade secret.” 360 Mortg. Grp., LLC v. Homebridge Fin. Servs., Inc., No. A-14-CA-00847-SS,
2016 WL 900577, at *6–7 (W.D. Tex. Mar. 2, 2016). Here, the fraud claim is based on exactly
the same allegations. See, e.g., Ex. A (HouseCanary’s 4th Am. Answer & Countercls.), ¶ 83
(alleging under “Count Three: Fraud and Fraudulent Inducement” that “TSI misappropriated
Alternatively, both the fraud and TUTSA claims fail because, if anything, this is only a
contract dispute. E.g., (HouseCanary closing 3/14/18 PM 42). Not surprisingly, given that this
HouseCanary’s TUTSA and fraud claims arise—if at all—out of that contract. Because Texas
law squarely prohibits plaintiffs from using “artful pleading to morph contract claims into fraud
can at most show a breach of contract—and the fraud claim fails for that additional, alternative
reason. Similarly, because HouseCanary “has not alleged facts establishing the breach of any
legal duty independent of the [parties’] agreement[s], its [TUTSA] claims sound only in
contract” and must be dismissed. Educ. Mgmt. Servs. v. Cadero, No. SA-14-CA-587, 2014 WL
However, HouseCanary’s breach of contract claims also fail as a matter of law because
Title Source breached the contract by failing to provide notice and an opportunity to cure in
response to HouseCanary’s failure to deliver a functioning app as required under the contract.
But the contract itself provided for immediate termination under that very circumstance. PX3,
In short, HouseCanary repackaged its allegations into multiple claims, but each iron in
the fire was made of the same fundamental facts and alleged injury. None of them can withstand
the heat. There is no evidence allowing them to stand individually and, being overlapping and
Finally, the jury based its damages award for the breach of contract and fraud claims on
Bratic’s legally insufficient opinion. In addition to the damages for the purported
misappropriation, the jury awarded $33.8 million in damages for the alleged breach of contract,
and the exact same amount for the alleged fraud, based on Bratic’s speculative measure of lost
profits. Ex. B (Jury Verdict Form), Qs. 13, 28. But the contract at issue limits damages for both
breach of contract and fraud to a cap of $2 million. See PX3 § 8. In any event, Bratic based his
party appraisers would use the never-completed HouseCanary app based solely upon Title
Source’s use. Bratic extended his unsupported opinion by adopting the factually and legally
infirm supposition that HouseCanary was entitled to three years of revenue under the contract
even though the contract was subject to termination after one year on purely discretionary
grounds. (Bratic 3/9/18 PM 164). Bratic’s opinions are speculative, contrary to law, and provide
In the end, if liability were to stand, and it manifestly should not, HouseCanary cannot
recover on its numerous overlapping theories and its wildly speculative damages figures.
Instead, what started out as a contract case should end as a contract case. If HouseCanary is to
recover anything, it is limited to damages caused by the alleged breach of the confidentiality
provisions of the contract (i.e., what HouseCanary calls misappropriation)—that is all. The
alleged breach did not destroy the value of the purported trade secrets, and HouseCanary did not
argue that it reduced their value in the market, as nothing has prevented HouseCanary from
marketing its technology to other users. As a result, the entirety of the damages in this action can
be nothing more than the value HouseCanary would have received under the contract—at most,
III. The punitive damages award violates Texas law and the U.S. Constitution.
a matter of law, the jury’s record-breaking punitive damages award necessarily fails. But the
jury’s award of nearly half a billion dollars in punitive damages cannot stand regardless. The
award violates Texas law because HouseCanary proffered no evidence—much less clear and
convincing evidence—that Title Source acted with any malice, as black-letter law requires. It
disproportionate. State Farm Mut. Auto. Ins. Co. v. Campbell, 538 U.S. 408, 416–17 (2003).
HouseCanary chose to turn this case into a referendum on “the perceived deficiencies” of
Title Source, its affiliates, and all of “Corporate America.” It asked the jury to “stand[ ] up and
. . . tell[ ] Corporate America, I’ve had enough of this” by putting this case “in the Wall Street
the jury, but it violates Texas law and the U.S. Constitution.
Title Source initiated this action because HouseCanary broke its contractual promise to
provide a mobile application that appraisers could use to complete appraisals in the field. Again,
HouseCanary’s own witnesses fatally undermined its legal position: they admitted HouseCanary
never delivered a completed app. E.g., (Poindexter 3/2/18 AM 82–83; Rhyne 3/6/18 PM 140).
HouseCanary failed to deliver a working product—much less one that could live up to its
promises.
HouseCanary tried to pin the blame for its breach on others—including Title Source. But
those excuses fail because there is no evidence to support them. Instead, the evidence
conclusively establishes that the app HouseCanary was contractually bound to deliver could not
be used to complete and submit appraisals—thereby thwarting the fundamental purpose of the
agreement—and there was no evidence that HouseCanary’s failure was excused. This Court
should therefore reverse and render judgment for Title Source on its breach of contract claim.
* * *
The stakes in this litigation have become enormous, not only for Title Source, but for this
critical area of commercial law. If bare assertion and insinuation of the kind proffered here can
afforded trade secret protection, the consequences for commercial relationships and innovation
throughout the economy will be profound. The matter before this Court is of great weight, but
the evidence and the law lead to one answer: the grant of judgment notwithstanding the verdict.
LEGAL STANDARD
A trial court should render JNOV when there is “no evidence of an essential element of
the [litigant’s] claims.” Gharda USA, Inc. v. Control Sols., Inc., 464 S.W.3d 338, 342 (Tex.
2015). No evidence supports a verdict where there is: “(a) a complete absence of evidence of a
vital fact; (b) the court is barred by rules of law or of evidence from giving weight to the only
evidence offered to prove a vital fact; (c) the evidence offered to prove a vital fact is no more
than a mere scintilla; (d) the evidence establishes conclusively the opposite of the vital fact.” Id.
at 347.
“Evidence does not exceed a scintilla” where it “do[es] no more than create a mere
surmise or suspicion that the fact exists.” Akin, Gump, Strauss, Hauer & Feld, LLP v. Nat’l Dev.
& Res. Corp., 299 S.W.3d 106, 115 (Tex. 2009) (citations omitted); see also Marathon Corp. v.
Pitzner, 106 S.W.3d 724, 728 (Tex. 2003) (“[A]n inference stacked only on other inferences is
not legally sufficient evidence.”). Although courts may not “substitut[e] their opinions on
credibility for those of the jurors,” “proper review also prevents jurors from substituting their
opinions for undisputed truth.” City of Keller v. Wilson, 168 S.W.3d 802, 816-17 (Tex. 2005).
ARGUMENT
HouseCanary’s claims are all grounded in the same core theory: that Title Source was
scheming all along to abscond with the alleged trade secrets for its own use. That conspiracy
The starting point for any judgment is the law. For each of the five alleged trade secrets,
HouseCanary bore the burden of satisfying each element of TUTSA misappropriation: (1)
HouseCanary owned a bona fide trade secret; (2) the trade secret was acquired by Title Source
through “improper means” or acquired and impermissibly used for a commercial benefit, and (3)
HouseCanary suffered harm as a result.2 These are the essential elements of a misappropriation
claim, and each element must be proven for each trade secret that was allegedly
misappropriated.
To qualify as a trade secret, the owner of technology must have taken “reasonable
measures . . . to keep the information secret,” and the information must derive “independent
economic value . . . from not being generally known” in addition to “not being readily
ascertainable.” Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(6). Thus, when “information [is]
generally known and readily available [it] is not protectable.” Trilogy Software, Inc. v. Callidus
Software, Inc., 143 S.W.3d 452, 467 (Tex. App.—Austin 2004, pet. denied). Further, it is black-
2
See Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(6) (defining “trade secret”);
§ 134A.002(3-a) (defining “owner”); St. Jude Med. S.C., Inc. v. Janssen-Counotte, No. A-14-
CA-877-SS, 2014 WL 7237411, at *15 (W.D. Tex. Dec. 17, 2014) (plaintiff “has not shown”
that it “actually owned those trade secrets”); Tex. Civ. Prac. & Rem. Code Ann.
§ 134A.002(3) (defining “misappropriation” as “(A) acquisition of a trade secret of
another . . . by improper means” or “(B) disclosure or use of a trade secret of another without
express or implied consent”); § 134A.002(1) (defining a “claimant” as “a party seeking to
recover damages”); Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *6 (There must be a
“‘direct causal link’ between the misconduct, the plaintiff’s injury, and the damages
awarded.”).
“imprecise,” that “weighs heavily against it being a trade secret” and calls into “question what
information [defendants] should be prohibited from using.” Glob. Water Grp., Inc. v. Atchley,
244 S.W.3d 924, 930 (Tex. App.—Dallas 2008, pet. denied) (JNOV for defendant was proper
For a trade secret to be acquired using “improper means,” the acquisition must have
maintain secrecy.” Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(2). Notably, a defendant
does not acquire a claimant’s trade secret by “improper means” if the claimant discloses the trade
secret willingly pursuant to a confidentiality agreement—even if the trade secret is later used in
violation of that agreement. See Educ. Mgmt. Services, LLC v. Tracey, 102 F. Supp. 3d 906, 914
(W.D. Tex. 2015); Demond v. Infiniti HR LLC, No. 3:17-CV-1322-D, 2017 WL 3835951, at *6
To prove the element of use, HouseCanary had to proffer evidence that “specific” trade
secrets “were misappropriated” and then put to commercial use by Title Source. Sw. Energy
Prod. Co., 491 S.W.3d at 722 (use means “commercial use by which the offending party seeks to
profit from the use of the secret’”) (quoting Glob. Water Grp., 244 S.W.3d at 930).
To prove harm, HouseCanary was required to establish a “direct causal link between the
misconduct, the plaintiff’s injury, and the damages awarded.” Lakeway Reg’l Med. Ctr., 2017
HouseCanary alleged, and the jury found, that Title Source misappropriated five
purported trade secrets: (1) HouseCanary AVMs, (2) Similarity Score, (3) Data Dictionary, (4)
rested its entire theory of misappropriation on Title Source’s development and use of MyAVM.
See, e.g., (HouseCanary Opening 1/31/18 AM 95 (claiming Title Source was “using our
confidential information, our data to train their model”); HouseCanary closing 3/14/18 PM 39-40
(claiming Title Source was attempting to “replicate” trade secrets with their “own
model”)). Thus, the dispositive legal question is whether each purported trade secret was
acquired and improperly used in the development of MyAVM. As the record shows, with
respect to each purported trade secret, the answer is no. (E.g., Rhyne 3/6/18 PM 87
(HouseCanary’s technical expert admits there is no evidence of “any fingerprints, any clues, any
Because its own expert prevented it from establishing that any trade secret was used in
MyAVM, HouseCanary resorted to lumping together disparate emails and communications that
were unrelated to MyAVM, misleading the jury into believing that this mishmash of unrelated
evidence somehow showed misappropriation. But a systematic, objective review of the elements
In the alternative, HouseCanary asserted that Title Source used a so-called “magic
HouseCanary data to “train” its own machine learning model to replicate HouseCanary’s
outcomes. (HouseCanary closing 3/14/18 PM 62). But this theory was again foreclosed by its
own expert, who admitted that “Title Source did not use any HouseCanary data to train
respect to all of the alleged trade secrets. Instead, the entire verdict fails as a matter of law if
even a single element is unsatisfied for just one trade secret, because HouseCanary’s damages
expert combined all the trade secrets together into one lump-sum calculation and “did not explain
which of the trade secrets contributed to what amount of [damages].” Tex. Adv. Optoelectronic
Sols., Inc. v. Renesas Elecs. Am., Inc., 888 F.3d 1322, 1335 (Fed. Cir. 2018) (vacating entire
award, under Texas law, when expert valued all trade secrets together and at least some of the
trade secret findings were invalid); Hunter Bldgs. & Mfg., 436 S.W.3d at 21 (where damages
opinion is based in part on conduct that does “not constitute misappropriation of trade secrets,”
entire opinion is no evidence of damages); see also Harris Cty. v. Smith, 96 S.W.3d 230, 234
(Tex. 2002) (reversing verdict based on “charge[] which mixed valid and invalid elements of
damages in a single broad-form submission” (citing Crown Life Ins. Co. v. Casteel, 22 S.W.3d
In light of these principles, each purported trade secret is analyzed based on the TUTSA
value that “a human appraiser” would assign to a home. (Rhyne 3/6/18 PM 102). HouseCanary
purported to have three AVMs: (1) a “Regression AVM,” (2) the “Cascade” AVM––a composite
of two AVMs it had licensed from a third-party, Black Knight––and (3) the “HouseCanary
AVM.” (See, e.g., HouseCanary closing 3/14/18 PM 18-19). HouseCanary failed to carry its
burden of proving each element of misappropriation for each AVM. Accordingly, judgment
must be rendered for Title Source on Jury Question Nos. 37(A) and 38(A).
20 years.” (Stroud 2/26/18 PM 180). As a result, the basic concepts behind various types of
AVMs are generally known in the industry and are not trade secrets. See Tex. Civ. Prac. & Rem.
Code Ann. § 134A.002(6)(B) (trade secrets must have “independent economic value . . . from
not being generally known”). Recognizing that general industry concepts cannot constitute trade
secrets, HouseCanary’s founder and CEO, Jeremy Sicklick, testified that the “secret sauce” in
each of the three AVMs is “[t]he HouseCanary algorithms, [that] put it all together.” (Sicklick
3/9/18 AM 73). It is this “set of algorithms” (Sicklick 3/7/18 AM 31)—the source code—that
“blend th[e] different analytics together,” providing “the right combination by market to get the
But even if HouseCanary may have a trade secret in some AVM algorithms, the record is
clear that Title Source never acquired or used any such trade secrets. Indeed, HouseCanary
admitted that it did not give any set of algorithms representing any of its AVM models to Title
Source. (See Rhyne 3/6/18 AM 119 (HouseCanary expert admits “[t]hey [Title Source] didn’t
have any of the analytics inside of” HouseCanary’s models); Stroud 2/26/18 PM 183 (testifying
that HouseCanary does not “provide information about how its models operate to its
customers”)). Rather, as demonstrated below, what was alleged to have been acquired and
a. Regression AVM.
not reveal to Title Source the analytics or algorithms behind its so-called Regression AVM. Title
Source, in fact, did not receive home valuation estimates from the Regression AVM. (See Stroud
2/16/18 PM 21 (testifying that all valuations provided were generated using the Cascade AVM or
HouseCanary AVM)).
a discussion with Jordan Petkovski, Title Source’s Vice-President and Chief Appraiser, in which
he was told that the appraiser app would use “a regression-based model” to help appraisers with
valuations in the field. (Stroud 2/26/18 PM 100).3 That is the sum total of what HouseCanary
revealed to Title Source about the methodology of its Regression AVM model. But regression
analysis is not a trade secret; it is a widely known form of statistical analysis dating to the
beginning of the 19th century. HouseCanary’s own expert, Vernon Rhyne, testified that Title
Source was “well aware of how to do a linear regression” before it met HouseCanary. (Rhyne
3/6/18 PM 64). As a result, HouseCanary cannot build a misappropriation claim—let alone one
worth hundreds of millions of dollars—on the allegation that it told Title Source it used common
No use of HouseCanary trade secrets. Nor did HouseCanary show any use by Title
HouseCanary’s own witness, Rhyne, testified that he performed an in-depth “investigation of the
Title Source source code” for MyAVM. (Rhyne 3/6/18 PM 85-86). But that investigation did
not reveal any use of HouseCanary’s trade secrets—Rhyne testified that “[he] didn’t see any
fingerprints, any clues, any reference to any HouseCanary technology.” Id. at 87 (emphasis
added). He testified that he “didn’t see any indication . . . of using HouseCanary technology”
3
HouseCanary elicited testimony from its expert that the “regression AVM” was “provided to
Title Source” in the never-delivered “appraiser app.” (Rhyne 3/6/18 PM 106-07 (reading
prior Stroud testimony that app would provide appraiser with “adjustments” calculated by the
regression AVM)). That is not evidence that Title Source ever acquired the algorithms that
make up the regression AVM. As any user of an app knows, a user only interacts with a user
interface, not the back-end algorithms that the user never sees. The app is not like a box with
physical AVMs inside that can be pulled out.
model that was even “related” to HouseCanary’s AVMs. (Rhyne 3/6/18 AM 119 (“Q: Was Title
Source able to develop anything related to the HouseCanary AVM[s]? A: No.”)). Rhyne’s own
testimony conclusively establishes that HouseCanary cannot satisfy the use element for its AVM
Additional evidence also demonstrated that the Regression AVM was never “used” by
Title Source. MyAVM is nothing like a regression AVM. MyAVM is a sophisticated machine
data.” (Petkovski 2/14/18 PM 109-10 (explaining that “machine learning” models “use actual
appraisal data to improve the accuracy of [their] predictions,” as the model “learns as it continues
method that is “totally different” from machine learning. (Rosenberg 3/1/18 AM 100; Rhyne
3/6/18 PM 101-02). Thus, any trade secret in HouseCanary’s Regression AVM is entirely
b. Cascade AVM.
As with the Regression AVM, HouseCanary never provided the Cascade AVM to Title
Source, and its own expert disavowed that it was ever used in MyAVM.
developed by combining two other AVMs that HouseCanary licensed from Black Knight, a third
party, with HouseCanary’s Home Pricing Index (HPI), which adjusts home prices across time.
(Stroud 2/26/18 PM 19). Under the terms of the parties’ contract, Title Source sent addresses to
HouseCanary, and HouseCanary ran those addresses through the Cascade AVM and sent value
reports back to Title Source. Id. at 19-21. In the words of HouseCanary’s expert, “all [Title
Source] had was the data [outputs]. They didn’t have any of the analytics inside of it.” (Rhyne
sensitive information, much less a trade secret. (E.g., Sicklick 3/7/18 AM 104).
In short, HouseCanary never provided Title Source with either the model or the inputs for
the Cascade AVM. In fact, Title Source had no knowledge that HouseCanary was even using
Black Knight’s AVMs as part of its model until after the start of litigation—HouseCanary hid the
fact that the Cascade AVM was not really its AVM at all. (See Stroud 2/26/18 PM 23-24). Nor
did Title Source have any information regarding the “weighting” of the model, which is the
aspect of the model that HouseCanary claimed was confidential. Id. at 209. HouseCanary’s
claim that Title Source misappropriated its trade secrets with respect to the Cascade AVM thus
fails. It is not possible to replicate what you never had, and never knew, in the first instance.
surprising that Title Source never used Cascade AVM when developing MyAVM. Nor can there
be any contention that MyAVM even resembles the Cascade AVM (i.e., combines two Black
Knight AVMs). As with the Regression AVM—and every HouseCanary trade secret—
HouseCanary’s expert, Rhyne, testified that he “didn’t see any indication . . . of using
evidence was introduced—and the record leaves no doubt that Title Source never acquired, and
never used, HouseCanary’s purported trade secrets with respect to the Cascade AVM.
c. HouseCanary AVM.
HouseCanary never provided Title Source with the model or analytics for its HouseCanary
AVM. Indeed, as with the Cascade AVM, HouseCanary’s own expert confirmed that
HouseCanary only disclosed some AVM model outputs to Title Source (i.e., the value estimates
3/6/18 AM 119).
HouseCanary’s own expert forecloses any claim of use of the “HouseCanary AVM.” See supra
pp. 20-22; (Rhyne 3/6/18 PM 87 (Rhyne “didn’t see any fingerprints, any clues, any reference to
Unable to show that Title Source acquired or used its formula or “secret sauce” for the
used some of the same modeling concepts as the HouseCanary AVM. See DX777 at 15. But the
structure discussed in this presentation was merely a proposal; it was never implemented and put
into use, as HouseCanary’s own expert, Rhyne, admitted. (Rhyne 3/6/18 AM 35). Moreover,
each of the basic concepts referenced in the PowerPoint has been publicly known, discussed, and
used in the real estate industry and by statistical modelers for decades. (E.g., Stroud 2/26/18 PM
195; Sicklick 3/7/18 PM 114-15; see also C. Wang 2/28/18 AM 141, 146-47). Thus, they are not
trade secrets.
the following basic concepts, each of which is generally known in the industry:
Title Source used to create MyAVM. HouseCanary cannot claim any proprietary
interest in the use of gradient boosting. HouseCanary did not invent GBM and
GBM code is available for free to everyone on the internet. (See C. Wang 2/28/18
own model did not even use GBM until after the parties’ relationship ended.
have replicated something that was not part of the HouseCanary AVM model in
sales—i.e., comparing the property being appraised to similar properties that have
already been sold––has been around since the early 2000s. (Stroud 2/26/18 PM
3/6/18 PM 43-44).
Time Series: Adjusting values for time is also not a trade secret, but a basic and
recommend and outline the use of time series adjustment in designing an AVM.
let alone showed, that MyAVM uses HouseCanary’s Home Pricing Index or HPI.
In sum, it is uncontested that Title Source never received HouseCanary’s actual model
formulas or analytics––its “secret sauce”—and generally known and published industry concepts
are not trade secrets. Baxter & Assocs., L.L.C. v. D & D Elevators, Inc., No. 05-16-00330-CV,
2017 WL 604043, at *9 (Tex. App.—Dallas Feb. 15, 2017, no pet.) (concept is not a “trade
Code Ann. § 134A.002(6)(B) (trade secrets must have “independent economic value . . . from
not being generally known”).4 Thus, HouseCanary’s misappropriation claim with respect to its
Failing to establish direct acquisition and use, HouseCanary alternatively asserted that
Title Source used a “magic machine” to take HouseCanary’s output data and reverse engineer its
purported trade secrets. (See HouseCanary closing 3/14/18 PM 22-23, 62). But HouseCanary
HouseCanary’s attorneys argued that Title Source reverse engineered its models by
feeding the output from HouseCanary AVMs (i.e., their estimates of home value) and data
HouseCanary’s models. This was equivalent, in HouseCanary’s words, to pouring Coke into a
“magic machine” to train the machine to produce more Coke. (HouseCanary closing 3/14/18
PM 62). But attorney argument is not evidence—and at trial, HouseCanary’s own expert,
Rhyne, repeatedly admitted that “Title Source did not use any HouseCanary data to train
4
While Title Source has analyzed each AVM for the sake of clarity and completeness, three
arguments apply to multiple trade secrets: (1) Rhyne’s testimony established that no trade
secret—AVM or otherwise—can satisfy the use element, as there is no evidence that any
trade secret was used in MyAVM (see Rhyne 3/6/18 PM 86; id. at 39-40); (2) HouseCanary
cannot claim that any model—AVM, similarity score, or complexity score—was acquired by
Title Source because, as Rhyne testified, Title Source “didn’t have any of the analytics inside
of [the models],” “all they had was the data [outputs],” (Rhyne 3/6/18 AM 119), and
HouseCanary cannot argue that any secret was acquired by “improper means,” as any
information Title Source did acquire about HouseCanary’s technology was provided
voluntarily by HouseCanary itself, see Educ. Mgmt., 102 F. Supp. 3d at 914; and (3) Rhyne’s
testimony establishing the absence of use in MyAVM’s development also defeats liability
and the damages element for every trade secret.
not use HouseCanary data to train its MyAVM product”); id. at 54 (“agree[ing]” that “MyAVM
did not use any HouseCanary information for its training data”)).
As Title Source’s lead developer testified, MyAVM’s training data was either owned by
Title Source or publicly available; the training data was not from HouseCanary. (C. Wang
2/28/18 PM 15 (explaining that MyAVM “learns on three parts of the data: Appraisal––Title
Source appraisal data and the demographic data to describe the zip code, and also the
macroeconomic data”)); (Rhyne 3/6/18 PM 60 (Rhyne agrees Title Source’s “data sources are
internal information, information from FNC, US Census, and publicly available data”)). In short,
HouseCanary’s own expert testified that Title Source never poured the “Coke”—i.e.,
HouseCanary’s data, reflecting its purported trade secrets—into the so-called “magic machine.”
After all, it would have made no sense for Title Source to train its model to replicate
mimic another model’s estimates of property values, but to mimic the real thing—actual
appraisal values, which Title Source already possessed. (Rhyne 3/6/18 PM 102-03; id. at 64-65).
Common-sense examples illustrate the point. HouseCanary’s claim would be like using polling
data, and not actual voting results, to develop a model that explains election outcomes. It would
be like using weather forecasts from the nightly news, and not actual weather data, to develop a
model that explains climate. That is precisely why Title Source used actual appraisal values to
train its model for MyAVM. See id. For this reason, there is neither evidence nor logic to
support the claim of reverse engineering using HouseCanary’s output data, and no reasonable
“discover [the] design, structure, construction, or source code” behind the HouseCanary models.
Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(5) (defining “[r]everse engineering” under
TUTSA). In fact, HouseCanary didn’t even allege it. Instead, HouseCanary alleged merely that
Title Source used a form of machine learning to attempt to replicate the outputs from
teach my model to work like your model whether lines are coded the same or not”)). But
machine learning is not synonymous with reverse engineering as defined in TUTSA—it does not
reveal “source code” or the “design [or] structure” of a computer program.5 Id. HouseCanary’s
to misdirect the jury by arguing that a former Title Source employee, Ryan Yang, attempted to
use HouseCanary’s models in creating a different model—his Home Value Estimation (“HVE”)
HVE is an “incredibly simple” and “one-page” formula that Title Source never used for
any purpose—much less a commercial one, which is what matters where trade secrets are
concerned. (Rosenberg 3/1/18 AM 93-94); see Sw. Energy Prod. Co., 491 S.W.3d at 722 (“use”
of a trade secret “means commercial use by which the offending party seeks to profit from the
use of the secret” (quoting Glob. Water Grp., 244 S.W.3d at 930)). As HouseCanary’s expert
testified, HVE was not even “code” in the programming sense of the word—rather, it was a
5
HouseCanary’s expert testified that the “fundamental nature” of machine learning renders it
inappropriate for examining “Source Code.” (Rhyne Robinson 2/23/18 AM 46-47 (“the
nature of machine learning” is unrelated to “Source Code”)).
code”); see also id. (“I don’t think of that [HVE formula] as software”)).
machine-learning model in MyAVM. (Rosenberg 3/1/18 AM 93-94; id. at 100). Because the
law requires commercial use, and because HouseCanary’s entire theory of harm and its damages
model are predicated on use in MyAVM, any argument with respect to HVE is nothing more
The Court need go no further to determine that any evidence with respect to HVE is
legally irrelevant and cannot support the jury’s verdict. But HouseCanary’s arguments with
respect to Ryan Yang’s HVE work also fail on numerous other independent grounds. To begin
with, HouseCanary ratified and approved Yang’s work on HVE—and waived any right to
complain about it. Precisely because there was nothing nefarious about Yang’s work, Title
Source openly disclosed Yang’s modeling in emails sent to HouseCanary before Amendment
One was signed—and HouseCanary did nothing. (Sicklick 3/8/18 AM 45-46).6 DX128; PX116;
(Sicklick 3/7/18 AM 93). Sicklick testified that he understood the emails to show that Yang was
“building a model and using [HouseCanary’s] data and understanding of attributes to improve
his model.” (Sicklick 3/7/18 AM 95; see id. at 98). But this was not improper—after all,
HouseCanary had agreed that Title Source had the “right to develop, use, or market products or
6
Sicklick attempted to claim there were provisions that were inserted into Amendment One in
response to PX116 (Sicklick 3/7/18 AM 95-98), but he later admitted that the language he
referred to was proposed in September 2015 prior to receipt of the email in PX116 (Sicklick
3/8/18 AM 39-44).
It was not until Title Source brought this lawsuit that HouseCanary conveniently
complained about Yang’s work. But by then it was too late. By choosing to enter into a
contractual relationship with Title Source with full knowledge of Ryan Yang’s work,
HouseCanary impliedly consented to and ratified that work—or, at a minimum, waived any right
to complain about it. See Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(3)(B) (defining
misappropriation as use of a trade secret without “implied consent”); see also Lesikar v.
Rappeport, 33 S.W.3d 282, 300 (Tex. App.—Texarkana 2000, pet. denied) (waiver and
ratification of contract claims); Fortune Prod. Co. v. Conoco, Inc., 52 S.W.3d 671, 677 (Tex.
2000) (ratification of fraud claims). HouseCanary’s inaction––in the face of knowledge that
Ryan was developing HVE––is an independent basis to reject HouseCanary’s TUTSA, fraud,
Moreover, as HouseCanary’s own expert admitted, the evidence does not show any
reverse engineering in the creation of HVE at all. Although HouseCanary’s attorneys insinuated
denied that it showed anything of the sort. (Rhyne 3/6/18 PM 101-02 (Rhyne rejects argument
that DX134 shows copying by Yang)). It merely showed that, in 2015, Yang was comparing the
accuracy of his own independently developed model (HVE) with that of HouseCanary’s Cascade
model (the model that is merely a composite of two Black Knight AVMs) to see which was more
accurate. (Rhyne 3/6/18 AM 128). Comparing is not copying. (Rhyne 3/6/18 PM 101-02).
Indeed, Rhyne testified that it was at least “equally likely” that Yang was merely trying
to improve his own model, HVE, to match the actual appraisal value, rather than engaging in any
independently fatal because, under the equal inference rule, “a jury may not reasonably infer an
ultimate fact from ‘meager circumstantial evidence which could give rise to any number of
inferences, none more probable than another.’” Hancock v. Variyam, 400 S.W.3d 59, 70-71
(Tex. 2013) (quoting Hammerly Oaks, Inc. v. Edwards, 958 S.W.2d 387, 392 (Tex. 1997)); see
also Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *9–10, 14 (rejecting trade secrets claim
where equal inference rule severed causation between misappropriation and damages).
Finally, HouseCanary broadly asserted that Ryan Yang’s mere presence on the MyAVM
development team—even in a minor, tangential role—was enough to show that he somehow used
proprietary HouseCanary information in MyAVM. But under Texas law, mere suspicion is not
evidence—and even an individual who had access to trade secrets (which Yang did not) cannot
be presumed to have used them later. See Greenville Automatic Gas Co. v. Automatic Propane
Gas & Supply, LLC, 465 S.W.3d 778, 787–88 (Tex. App.—Dallas 2015, no pet.) (evidence that
claimant’s former employee had extensive knowledge of its proprietary customer lists, and
served the same customers for his new employer, was “not evidence of use,” as claimant did not
show employee’s knowledge of trade secrets was actually used in creating the new employer’s
customer list); Trilogy Software, 143 S.W.3d at 465 (“Standing alone, [the programmer’s] receipt
of the e-mail [containing trade secrets] . . . does not raise a genuine issue of material fact that he
In the end, HouseCanary’s speculation about Ryan Yang and the development of his
simplistic, never-used HVE model—which has nothing to do with MyAVM—is simply a red
herring and does not support its misappropriation claims. The HVE allegations are legally
HouseCanary also claimed that it developed a “similarity score model,” which Chris
“comp[arable]” property was to the subject property. (Stroud 2/26/18 PM 155-56). This model
generated a similarity “score” between 0 and 100 that was listed on HouseCanary Value Reports
and was intended to appear in the appraisal app. Id. at 156-57. But yet again, HouseCanary
failed to show that Title Source either acquired or used HouseCanary’s similarity model, let
HouseCanary’s “similarity score model,” with its “mathematical distance” calculations and
“regression-based method,” was never provided to Title Source. (Rhyne 3/6/18 AM 119). As
HouseCanary admitted, it only provided Title Source with general “information about the
similarity score” and some “outputs from the similarity score model”—not the model itself.
What HouseCanary did provide are not trade secrets. As HouseCanary’s witnesses
admitted, the “outputs” of a model are not sensitive information, much less a trade secret.
(Sicklick 3/7/18 AM 104). And HouseCanary’s executives denied that the generic information it
provided about the similarity score was “proprietary” or a trade secret. This information
consisted of half a page in a sales presentation with practically the same information about
HouseCanary’s similarity score that HouseCanary provides publicly on its website. Compare
DX759, with PX518B. Chris Zaloumis, HouseCanary’s Vice President of Product, stated that
this presentation (DX759) contained only a “general description” of the inputs for the similarity
if it were to provide Title Source with “more technical specifics” than were in the
presentation. DX334.
Chris Stroud, HouseCanary’s Chief of Research, responded that HouseCanary could even
“provide a longer list [of information] while still keeping it generic to protect our IP”—in other
words, the description was so “generic” that even more detail could be included without
disclosing any proprietary information. Id. HouseCanary’s expert testified that this document
does not convey any information about the underlying modeling technology. (Rhyne Robinson
2/23/18 AM 88-89 (testifying that DX759 “doesn’t tell you how the model operates, no”)). And
he even admitted that the “geographic and property characteristics” that the document described
(Rhyne 3/6/18 PM 74) were “common sense” in the real estate industry and well known to Title
Source long before its relationship with HouseCanary. Id. at 55-57. A “general” description of a
model—which largely repeats public information, does not describe “how the model operates,”
and was “common sense” already known to Title Source—is not a trade secret. Tex. Civ. Prac.
& Rem. Code Ann. § 134A.002(6)(B) (trade secret must be non-public and provide “independent
economic value”).
Unable to present evidence that Title Source acquired the similarity model—and faced
with conclusive evidence to the contrary—HouseCanary instead pressed the theory that Title
based algorithm and start[ed] to train it to replicate the similarity score.” (Stroud 2/26/18 PM
171). This is all suspicion and no evidence. HouseCanary’s stack of inferences—that Title
Source had the motive to develop its own “similarity score,” that it could have used “machine-
learning” to acquire HouseCanary’s similarity score, and that it ultimately developed some,
even a scintilla of evidence that Title Source actually acquired HouseCanary’s similarity score
model, much less used it. See Marathon, 106 S.W.3d at 728 (“[A]n inference stacked only on
other inferences is not legally sufficient evidence.”); Lakeway Reg’l Med. Ctr., 2017 WL
672451, at *14 (no evidence of misappropriation when claimant was “simply speculating” that
defendants “must have used [it]s confidential information”). There is no evidence that anyone at
Title Source used machine learning to acquire HouseCanary’s similarity model. And, as
discussed below, any similarity metric developed by Yang neither reflected HouseCanary’s
testified that the similarity model was not used in MyAVM. As Rhyne testified, there is no
MyAVM—testified it did not contain any similarity model at all. (Manheim 3/6/18 AM 22
(“When I was conducting my code review [of MyAVM], I couldn’t––I didn’t see any evidence
of similarity score”)). Nor is there any evidence that any HouseCanary similarity score outputs
were used to “train” MyAVM’s machine learning model. (Rhyne 3/6/18 PM 39-40 (“Title
Source did not use any HouseCanary data to train MyAVM”)). These admissions are fatal to any
Unable to establish use in MyAVM, HouseCanary again resorted to pure speculation and
misdirection, pointing to a Title Source presentation that mentioned a generic “similarity score”
to assert that Title Source used HouseCanary’s similarity score. (Sicklick 3/9/18 AM 24-25).
But the presentation only states that some similarity metric may have been developed by Ryan
testified, while there is some evidence that Yang was “working on some similarity metrics,”
“that could be anything.” (Manheim 3/6/18 AM 40-41). Similarity scores are nothing new—and
HouseCanary can hardly claim a trade secret based on the mere idea of a similarity score. (E.g.,
Stroud 2/26/2018 AM 146).7 Indeed, Rhyne testified that Title Source already had a similarity
score before entering into an agreement with HouseCanary. (Rhyne 3/6/18 PM 68-71
(discussing PX130)). There is no evidence that Yang’s unknown metric used, or bears any
semblance to, HouseCanary’s similarity model, and HouseCanary’s experts testified that
HouseCanary’s similarity model was not used in MyAVM—which is all that counts as a matter
of law.
“evidence” or source code for any “similarity score or similarity model” of a similarity score in
MyAVM (Manheim 3/6/18 AM 43)—that a similarity score could have been used in MyAVM
because he had seen an empty folder on Yang’s computer labeled “similarity score.” Id. at 43-
44, 54. But Manheim admitted that Claude Wang, who developed MyAVM, did not have the
empty folder on his computer. Id. at 39, 44. He further admitted there could be “a lot of
different explanations as to why there is a similarity folder outside of [there] having to be Source
Code for it.” Id. at 43. And he admitted that he could not tell “anything about the history of that
folder, who created it, and when it was created,” or “if there was ever any code in it.” Id. at 51.
7
Moreover, full mathematical formulas for creating similarity scores are available on the
internet, including on Wikipedia. (Stroud 2/26/18 145-46; Rhyne 3/6/18 44). Indeed, this is
precisely the information that Yang used to create a simple similarity metric. (C. Wang
2/28/18 PM 154).
Manheim had regarding the folder was purely speculative, at odds with the actual facts (i.e., his
own admission that there was no similarity model used in MyAVM), and cannot support the
judgment. See Wal-Mart Stores, Inc. v. Merrell, 313 S.W.3d 837, 840 (Tex. 2010) (“An expert’s
failure to explain or adequately disprove alternative theories of causation makes his or her own
theory speculative and conclusory.”); Coastal Transport Co. v. Crown Cent. Petrol. Corp., 136
S.W.3d 227, 232 (Tex. 2004) (“Opinion testimony that is conclusory or speculative is not
relevant evidence.”).
In sum, HouseCanary’s own testimony establishes that Title Source never acquired its
similarity model or analytics and did not use that model––indeed, did not use any similarity
model––in MyAVM. Again, because HouseCanary’s entire theory of liability and damages is
predicated on the use of its purported trade secrets in MyAVM, this claim fails.
According to HouseCanary, its data dictionary was an index of the various types of data
available for sale from HouseCanary, along with brief descriptions of the type of data. See
Chris Stroud, admitted, most of the data dictionary was licensed from Black Knight—a third
party—and was not owned by HouseCanary. (E.g., Stroud 2/26/18 PM 27-28 (“Q: . . . [T]he
county level homeowner file and all of these property characteristics [in the data dictionary],
those all come from Black Knight, right? A: Yes, we license those.”); id. at 29 (“Q: Included in
this Black Knight data was, you know – whether [the home] has a garage, a pool, things like that,
right? A: Yes.”)). Similarly, any non-licensed data described in the dictionary was public
information. Id. at 142 (data dictionary is an index of “data that we licensed” from third parties
hinges on evidence that Yang reproduced some pages of a “data dictionary” from a HouseCanary
spreadsheet in August 2015. (Rhyne 3/6/18 91); DX136. But those pages were not
HouseCanary’s—they were indisputably created by, owned by, and licensed from a third party,
Black Knight (previously known as LPS). (Rhyne 3/6/18 PM 94 (“It is clearly data that
originated with LPS”); id. at 95 (“It originated, obviously, with LPS”); id. at 95 (“It originated
with LPS, no doubt about it.”); see also id. at 94 (Rhyne agrees that Yang “could get that stuff
from LPS directly”)); PX49 at 43-60 (data dictionary containing LPS logo); PX62 (Black Knight
licensing agreement). HouseCanary cannot have a trade secret in what it did not develop, and
Further, it is undisputed that Title Source did not acquire any portion of the data
dictionary by “improper means”—rather, the dictionary was freely shared by HouseCanary itself
as part of various promotional materials for its products. (Stroud 2/26/18 PM 26); PX49; PX64.
Indeed, as is true for all of the purported trade secrets, HouseCanary willingly gave Title Source
all the information Title Source received about the data dictionary—none of the purported trade
secrets were acquired by improper means. Given that all of the purported trade secret
8
House Canary’s License Agreement with Black Knight did not give HouseCanary any “title
to, or the right to enforce rights in” the excerpt of Black Knight’s data dictionary. Tex. Civ.
Prac. & Rem. Code Ann. § 134A.002(3-a). To the contrary, the License Agreement only
gave HouseCanary “right, title and interest in and to [so-called] Derivative Works,” PX62 at
15, which are works created by HouseCanary which “do not include the Licensed Materials
in the form provided by BK,” id. at 1 (emphasis added). Here, the undisputed evidence is
that the excerpt of the data dictionary was in “the form provided by BK,” complete with the
“LPS” (i.e., Black Knight) logo on the top. Compare DX136 at 2–8 with PX49 at 43–48.
Thus, it does not constitute a “Derivative Work[],” and Black Knight retains “all rights in and
to” these pages, including all “proprietary rights.” See id. at 2 (“BK reserves all rights in and
to the Licensed Materials, including, but not limited to . . . all . . . proprietary rights”).
provisions is no evidence of acquisition by improper means. See, e.g., Educ. Mgmt. Servs., 102
F. Supp. 3d at 914 (subsequent breach of confidentiality provisions “is irrelevant to the method
No use of HouseCanary trade secrets. Nor did HouseCanary provide any evidence that
Title Source used the data dictionary in MyAVM. See Tex. Civ. Prac. & Rem. Code Ann.
demonstrate use. See GE Betz, Inc. v. Moffitt-Johnston, 885 F.3d 318, 327 (5th Cir. 2018)
(employee’s “download of [trade secret] data and emails to herself on the eve of her departure”
were “no evidence of actual use”). And HouseCanary’s expert, Rhyne, admitted that he “didn’t
see any implementing of the HouseCanary data dictionary” in MyAVM. (Rhyne 3/6/18 PM
87).9 In fact, the undisputed evidence shows that Title Source used the data from its own
appraisals, along with data from a few widely known public sources, to train MyAVM. (C.
HouseCanary also cannot claim that Title Source used the data dictionary to identify and
copy the “ingredients” (variables) that HouseCanary used in its AVMs. The data dictionary does
not identify the specific variables used in HouseCanary’s AVMs; rather, it contains an
9
Rhyne claimed that there was some similarity between MyAVM and the pages Yang
reproduced because Yang’s compiled data dictionary contained headings regarding
“Neighborhood characteristics,” “demographics,” and “economic data,” and a PowerPoint
slide regarding MyAVM stated that Title Source uses “Demographic data at zip level,”
“Neighborhood information, “Macro-economic data,” and “Appraisal reports.” (Rhyne
3/6/18 AM 125). But Rhyne admitted that Yang developed a data dictionary that used the
“[s]ame set of characteristics” as MyAVM before HouseCanary shared the Black Knight
dictionary. Id. at 58 (emphasis added). And Rhyne admitted that Title Source—like every
other AVM developer—would use the same categories of data. (Rhyne 3/6/18 PM 56).
Further, as discussed just above, the variables used in MyAVM to predict property value
are the very same property and neighborhood factors used by Title Source’s human appraisers to
value a property. These attributes are “common sense” and were known to Title Source long
before it received HouseCanary’s data dictionary. (Rhyne 3/6/18 PM 55, 64). Moreover, Rhyne
admitted that Title Source—and, indeed, every other AVM developer—would use these same
types of data. (Rhyne 3/6/18 PM 56 (testifying that Title Source and other appraisal companies
“know what are the attributes, the property attributes that provide a home value”); Stroud
2/26/18 PM 98 (admitting that Title Source gave HouseCanary ideas about where to collect MLS
data, not the other way around)). And it is well established that a listing of publicly available
information does not, and cannot, make that information proprietary and preclude its use. See
Baxter & Assocs., 2017 WL 604043, at *9 (information is not a “trade secret” under TUTSA if it
In sum, the portions of the data dictionary at issue are not HouseCanary’s trade secret and
were not used in the creation of MyAVM. Further, as a matter of Texas law, Title Source’s
10
Texas courts routinely hold that data lists—for example, customer lists—are not confidential
information deserving trade secret protection if that information is readily accessible by
industry inquiry. See, e.g., Guy Carpenter & Co. v. Provenzale, 334 F.3d 459, 467-68 (5th
Cir. 2003) (“A customer list of readily ascertainable names and addresses will not be
protected as a trade secret.”); SCM Corp. v. Triplett Co., 399 S.W.2d 583, 586 (Tex. App.—
San Antonio 1966, no writ) (data compilation “was readily ascertainable by anyone, and was
therefore not a trade or business secret”); see also Numed, Inc. v. McNutt, 724 S.W.2d 432,
434–35 (Tex. App.—Fort Worth 1987, no writ); Allan J. Richardson & Assoc. v. Andrews,
718 S.W.2d 833, 836–37 (Tex. App.—Houston [14th Dist.] 1986, no writ); Research Equip.
Co. v. C.H. Galloway & Scientific Cages, Inc., 485 S.W.2d 953, 956 (Tex. Civ. App.—Waco
1972, no writ). HouseCanary has no right to prevent the use of this public information by
any other party.
establishing misappropriation, and Title Source has every right to use its own data, and widely-
Chris Stroud, HouseCanary’s Chief of Research, testified that HouseCanary created its
data compilation—i.e., database––from publicly available data and data licensed from third
parties. (Stroud 2/26/18 PM 152; see also id. at 144). As with HouseCanary’s other purported
trade secrets, undisputed evidence demonstrates that Title Source neither acquired nor used
provided Title Source with its data compilation. All that Title Source received from
HouseCanary were valuation reports containing individual pieces of data for the specific
property—for example, the number of bedrooms in a particular house or the year it was built.
(See, e.g., Sicklick 3/7/18 AM 116-17); DX303. These pieces of data may have come from the
data compilation, but they were not equivalent to the compilation itself. It was the compilation
as a whole—not the scattered pieces of data it contained—that was the trade secret put before the
Stroud, HouseCanary’s Chief of Research, testified that what made the data compilation
useful to HouseCanary was both its size—allegedly “the most comprehensive residential real
estate database available” (Stroud 2/26/18 PM 152)—and its structure—“a large part” of
developing a model is “getting the data infrastructure in place,” id. at 92. The individual points
of data were comparatively worthless. See id. at 108. And, indeed, they are either in the public
reports in the course of the parties’ contractual relationship. DX651. But in the context of
HouseCanary’s “residential real estate dataset” covering “100 million residential properties”
across the United States, (Stroud 2/26/18 PM 108, 110, 152), that is less than 0.2% of the value
reports that could have been generated by the data compilation. The amount of data Title Source
received was far too small to build a data compilation large enough to create a workable AVM.
Moreover, the individual data points Title Source received in the value reports were not
would have had to compile HouseCanary’s data itself—by extracting raw data from every one of
HouseCanary’s value reports and then collating, culling, and organizing that data into a working
dataset. There was no reason for Title Source to go through this rigmarole to get a mere fraction
of data already available to it from public sources and its own appraisals.
HouseCanary’s data compilation, it could not have used it, either. As previously discussed,
Rhyne expressly testified that no HouseCanary data was used to develop or train MyAVM.
(Rhyne 3/6/18 PM 39-40). Again, Title Source used the data from its own appraisals, along with
data from public sources, to train MyAVM. (C. Wang 2/28/18 PM 15).
A complexity score helps determine how difficult it would be for an appraiser to assess
the value of a particular property so that the task is priced appropriately and an appraiser with the
right skill level is assigned to it. Yet again, the undisputed evidence demonstrates that Title
Source neither acquired nor used HouseCanary’s complexity score. Indeed, because complexity
presentation, as part of its continued effort to market products to Title Source, to “build a model
of appraisal complexity.” (Stroud 2/26/18 PM 62); DX561. But the undisputed evidence, as
HouseCanary’s Chief of Research, Chris Stroud, admitted, is that a complexity model was
“never developed.” (Stroud 2/26/18 PM 187); see also PX314 (email noting decision made to
“shelve[]” the development of appraisal complexity score “indefinitely”). It defies reality that
contain no formulas, no source code, and no concrete information about how the complexity
score would actually work. DX561 at 3. The general descriptions in this sales pitch thus cannot
qualify as a trade secret as a matter of law. See Glob. Water Grp., Inc., 244 S.W.3d at 930
(where “[t]here is no discrete secret formula at issue,” “the imprecise nature of the information
weighs heavily against it being a trade secret”). And HouseCanary offered no evidence that it
11
At trial, HouseCanary confusingly referred to two different concepts as a “complexity score”:
(1) an “appraisal complexity model,” and (2) a “property score.” The only “complexity
score” that Stroud asserted as a trade secret at trial, however, was the “appraisal complexity
model.” (Stroud 2/26/18 AM 132). In deposition testimony (admitted at trial), Stroud
testified that the “complexity score” was also known as a “property score,” a metric
HouseCanary was obligated to provide to Title Source under the contract. Id. at 122. Unlike
the “appraisal complexity score,” the “property score” was intended to represent whether
enough information was available for a computer model (i.e., an AVM) to estimate the value
of a home. (Rhyne 3/6/18AM 113); PX3, Ex. A. At trial, Stroud disavowed his prior
statements equating the “complexity score” and “property score” and instead claimed that the
“appraisal complexity model” was the “complexity score.” (Stroud 2/26/18 AM 125).
Moreover, under Texas law, a trade secret must have “independent economic value.”
Tex. Civ. Prac. & Rem. Code Ann. § 134A.002(6)(b). And the sales pitch HouseCanary
manifestly useless, given that HouseCanary never developed the model to production and
“shelved [the project] indefinitely.” (Stroud 2/26/18 AM 125); PX314. Thus, the “project” itself
had no measurable commercial value to HouseCanary. See Tex. Civ. Prac. & Rem. Code Ann.
§ 134A.007(6)(B) (requiring a trade secret to have “independent economic value” from “not
being generally known”); see also Triple Tee Golf, Inc. v. Nike, Inc., 485 F.3d 253, 266 (5th Cir.
2007) (“For TTG to succeed on any of these claims, then, it must first prove that it in fact
possessed (1) proprietary information, that was (2) valuable to its business.”) (emphasis added);
BondPro Corp. v. Siemens Power Generation, Inc., 463 F.3d 702, 707 (7th Cir. 2006) (grant of
JNOV was proper because evidence that process was not commercially used provided a
“compelling” inference that the process has no measurable commercial value) (applying UTSA
as adopted in Wisconsin). For all these reasons, HouseCanary did not have or provide, and Title
No use of HouseCanary trade secrets. Rhyne admitted there is no evidence that Title
Source ever developed or used a complexity score at all as part of its business (it does not), let
alone in MyAVM. See, e.g., (Rhyne 3/6/18 PM 87, 99-101). That is not surprising since a
complexity score is not used to create an AVM, and it was not used as part of HouseCanary’s
AVM or Title Source’s MyAVM. Again, because HouseCanary’s entire theory of liability and
damages is predicated on the use of its purported trade secrets in MyAVM, HouseCanary cannot
do with AVMs, let alone MyAVM. See Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *14.12
Not only did HouseCanary fail to show that Title Source acquired or used any of its
purported trade secrets, it failed to show any harm or damages resulting from the alleged
requires “a causal link between the improper use of trade secrets and [the claimant]’s damages,”
but HouseCanary did not offer any link between any allegedly improper use of its five trade
secrets and the damages it demanded. See Lakeway Reg’l Med. Ctr,, 2017 WL 672451, at *6, 14
(requiring a “direct causal link between the misconduct, the plaintiff’s injury, and the damages
awarded”).
12
To the extent the jury could have understood “complexity score” on the jury verdict form to
refer to the “property score” (also known as the “suitability score,” to make matters more
confusing), HouseCanary provided no evidence of misappropriation of that either. There was
no evidence that HouseCanary owned the property score model—Black Knight did. See
(Stroud 2/26/18 PM 32-33 (admitting that the “property score was obtained from Black
Knight”)). Even if HouseCanary had some right to sue for misappropriation of somebody
else’s property score, it never showed that Title Source acquired the model. There was
likewise no evidence that the generalized descriptions of the property score provided to Title
Source were a trade secret. HouseCanary pointed only to a series of emails in which Yang
said he wished to “[f]ind out what’s behind” the property score, DX128, and “estimate” the
“property complexity” score, DX118; see also DX695 (Title Source email claiming that
Yang was “making progress on developing our complexity model using HouseCanary data”),
but emails stating Yang intended to “[f]ind out what’s behind” the property score is not
evidence that Yang succeeded in doing so. See Trilogy Software, Inc. v. Callidus Software,
Inc., 143 S.W.3d 452, 465 (Tex. App.—Austin 2004, pet. denied) (programmer’s “mere
receipt of [an] email” telling him to “take a look” at trade secret information is “no evidence”
that “he did, in fact, [] look at the information” and use it). As with the other purported trade
secrets, there is no evidence that Title Source ever used the property score, and HouseCanary
failed to demonstrate any specific harm from any misappropriation of the property score.
misappropriation of any specific trade secret, instead lumping them together into one
undifferentiated mass. Because Bratic and HouseCanary did not “attempt[] to measure” what
harm, if any, befell HouseCanary from the alleged misappropriation of any of the five trade
secrets specifically—each trade secret claim thus disintegrates. Hunter Bldgs. & Mfg., 436
S.W.3d at 21 (failing to “measure” harm from specific trade secret misappropriation mandates
* * *
impermissibly using a bona fide trade secret, thereby causing harm or damages. HouseCanary’s
smoke and mirrors misled the jury into believing that some trade secret was somehow
misappropriated. But a careful, step-by-step analysis of the governing law and relevant evidence
leads to only one conclusion—that HouseCanary failed to carry its burden of establishing each
element of misappropriation for any of the five trade secrets, and that no legally sufficient
evidence supports the verdict. See TM Prods., Inc. v. Nichols, 542 S.W.2d 704, 707 (Tex. Civ.
App.—Dallas 1976, no writ) (“[T]he rule providing that the court may render judgment non
obstante veredicto if a directed verdict would have been proper, does not mean literally ‘no
evidence at all’ but comprehends those situations in which the evidence is deemed to be legally
insufficient to establish an asserted proposition of fact.”). Title Source never acquired and never
used any of HouseCanary’s supposed trade secrets—and what was purported to have been
If even a single element of misappropriation is not satisfied, that trade secret claim
fails—and HouseCanary’s five trade secret misappropriation claims each independently fail as a
supposed contribution of each purported misappropriation claim to the damages, if even one
claim fails, the entire verdict collapses. See supra p. 18. Judgment must be rendered for Title
Because they are grounded in the same core theory of misappropriation, HouseCanary’s
fraud claims fail for the same basic insufficiency-of-the-evidence reasons that its TUTSA claims
do. But the fraud claims break down as a matter of law for two more independent reasons: (1)
they are preempted by TUTSA, and (2) there is zero evidence of fraudulent intent.
TUTSA broadly preempts common-law tort claims that are based on misappropriation of
trade secrets. Tex. Civ. Prac. & Rem. Code Ann. § 134A.007(a) (“[T]his chapter displaces
conflicting tort, restitutionary, and other law of this state providing civil remedies for
indicates that the law was intended to prevent inconsistent theories of relief for the same
underlying harm by eliminating alternative theories of common law recovery which are premised
on the misappropriation of a trade secret.” Super Starr Int’l, LLC v. Fresh Tex Produce, LLC,
To escape preemption, HouseCanary had to show that its fraud claims are “based on facts
unrelated to the misappropriation of the trade secret.” 360 Mortg. Grp., LLC v. Homebridge Fin.
Servs., Inc., No. A-14-CA-00847-SS, 2016 WL 900577, at *6–7 (W.D. Tex. Mar. 2, 2016); see
also Embarcadero Techs., Inc. v. Redgate Software, Inc., No. 1:17-CV-444-RP, 2018 WL
315753, at *3 (W.D. Tex. Jan. 5, 2018) (preemption upheld where “both claims stem from the
misappropriation claims. It explained at trial that the “same facts basically support . . . [its] fraud
claims” and its misappropriation claims. (HouseCanary closing 3/14/2018 PM 20; see also id. at
41 (claiming that Title Source’s purported use of HouseCanary’s information to build its own
model “turns out to be fraud . . . and it is also misappropriation of trade secrets”)); Ex. A
(HouseCanary’s 4th Am. Answer & Countercls.) ¶ 83 (alleging under “Count Three: Fraud and
matter of law. See Embarcadero Techs., 2018 WL 315753, at *3. Put another way,
Interactive, Inc. v. Philips Elec. N. Am., 92 F. Supp. 3d 982, 992 (D. Haw. 2015). As a result, its
common law fraud or fraudulent inducement: (1) a material representation that was (2) false
and (3) known by the speaker to be false (or made recklessly) (4) with the intent that it should be
acted upon by the party, (5) the party acted in reliance upon it, and (6) thereby suffered injury.
DeSantis v. Wackenhut Corp., 793 S.W.2d 670, 688 (Tex. 1990). Fraudulent inducement
requires additional proof that the misrepresentation induced a party to enter into a contract.
Haase v. Glazner, 62 S.W.3d 795, 798–99 (Tex. 2001). HouseCanary failed to prove any of
As explained above, HouseCanary’s fraud theory centers on its allegations that Title
Source misappropriated its trade secrets. But those allegations are completely unsupported. See
supra pp. 20-45. Furthermore, HouseCanary failed to demonstrate that Title Source intended to
counsel made vague and conclusory assertions that because Title Source supposedly “took [its]
trade secrets,” Title Source must have planned “to be doing this” all along. (HouseCanary
closing 3/14/2018 PM 20-21). But attorney argument is not evidence, and a verdict cannot be
based upon mere assertions. Nor did HouseCanary show any reliance on any particular
representation or inducement to enter the contract, as opposed to its own desire to make a sale.
Both HouseCanary’s fraud and TUTSA claims fail as a matter of law for yet another
reason: they are not tort claims at all, but sound only in contract, because they depend on the
existence of a contract and seek to recover for economic loss. Further, under the one satisfaction
rule, HouseCanary can recover only under one of its three overlapping theories of liability.
Consequently, the verdict on Jury Question Nos. 11, 14, 38, and 41 should be set aside.
Texas law squarely prohibits plaintiffs from recovering tort damages “when damage
allegedly resulting from the defendant’s tortious conduct is limited to the loss of a contractual
benefit.” ConocoPhillips Co. v. Koopmann, 542 S.W.3d 643, 664 (Tex. App.—Corpus Christi
2016), aff’d on other grounds, No. 16-0662, 2018 WL 1440639 (Tex. Mar. 23, 2018), reh’g
denied (June 22, 2018) (citing Chapman Custom Homes, Inc. v. Dallas Plumbing Co., 445
S.W.3d 716, 718 (Tex. 2014) (per curiam); Sw. Bell Tel. Co. v. DeLanney, 809 S.W.2d 493, 494
(Tex. 1991)). A two-prong test applies. First, “if the defendant’s conduct . . . would give rise to
liability only because it breaches the parties’ agreement, the plaintiff’s claim ordinarily sounds
only in contract.” DeLanney, 809 S.W.2d at 494. Second, “[w]hen the only loss or damage is to
the subject matter of the contract, the plaintiff’s action is ordinarily on the contract.” Id. Each
prong is independently sufficient to bar an action in tort. See Chapman, 445 S.W.3d at 718 .
that a contract exists between the parties.” DeLanney, 809 S.W.2d at 494. That is plainly the
case here. This dispute grew out of a $5 million contract between the parties. In HouseCanary’s
closing argument, its counsel asserted that “the same facts” supported its fraud, breach of
contract, and misappropriation claims against Title Source: “Basically we say that that they took
our trade secrets, information we disclosed to them about our models, they took our data that we
overarching theory of the case is that Title Source misused its trade secrets in violation of the
“duty to maintain secrecy . . . to limit use” in the parties’ contracts. Id. at 42.
For instance, HouseCanary repeatedly argued that Title Source improperly reverse
engineered HouseCanary’s trade secrets. See, e.g., (HouseCanary closing 3/14/18 PM 22-25).
But even if true, which it is not, reverse engineering is expressly permitted by TUTSA. See Tex.
Civ. Prac. & Rem. Code Ann. § 134A.002(4) (“‘Proper means’ means discovery by . . . reverse
engineering unless prohibited”). HouseCanary’s real complaint is thus that Title Source
allegedly violated provisions of the parties’ contracts that prohibited reverse engineering. See
(HouseCanary closing 3/14/18 PM 42 (“Reverse engineering, when you’ve signed contract after
contract specifically agreeing not to reverse engineer, is improper.”)); see also DX1085
confidential information. But, in any event, because the alleged reverse engineering “give[s] rise
to liability only because it breaches the parties’ agreement,” HouseCanary’s claim “sounds only
in contract.” DeLanney, 809 S.W.2d at 494. Because HouseCanary’s TUTSA and fraud claims
purported trade secrets––they both fail as a matter of law and must be rejected.13
HouseCanary’s claims are also independently barred by the second prong of DeLanney
because they seek to recover for economic loss “to the subject matter of the contract.” 809
S.W.2d at 494-95. Based on the calculations of HouseCanary’s damages expert, the jury
awarded the exact same damages—$201.6 million for the value of the trade secrets or $64.1
million for use of the trade secret––for HouseCanary’s TUTSA and breach of contract claims.
See Ex. B (Jury Verdict Form), Qs. 39, 43–45. And the jury awarded the same damages—lost
profits of $33.8 million—for HouseCanary’s fraud and breach of contract claims. See id. Qs. 13,
28.
HouseCanary cannot, of course, recover twice for the same injury—let alone thrice (for
misappropriation, fraud, and breach of contract based on the same facts and alleged injury),
which is what it is seeking to do here. See Stewart Title Guar. Co. v. Sterling, 822 S.W.2d 1, 7
(Tex. 1991). Where, as here, “any damages suffered” from misappropriation derive from
contractual obligations, the “claims sound only in contract,” Ally Fin., 2014 WL 261038, at *10,
and the “one satisfaction rule . . . limits a plaintiff’s recovery to one of several overlapping
theories,” Myriad Dev., Inc. v. Alltech, Inc., 817 F. Supp. 2d 946, 983 (W.D. Tex. 2011).
13
See e.g., Ally Fin., Inc. v. Gutierrez, No. 02-13-00108-CV, 2014 WL 261038, at *10 (Tex.
App.—Fort Worth Jan. 23, 2014) (affirming summary judgment because “the trade secret
was acquired through a breach of a confidential relationship,” which could not occur “apart
from [the] alleged breach of a contractual provision”); Educ. Mgmt. Servs., 2014 WL
12586781, at *2 (dismissing TUTSA claim “[b]ecause the Plaintiff has not alleged facts
establishing the breach of any legal duty independent of the non-disclosure agreement” and
the claim thus “sounds only in contract”).
principles of contract law. As a result, the misappropriation and fraud recoveries must be struck
HouseCanary also does not have legally sufficient evidence on its breach of contract
claims. HouseCanary claimed, and the jury found, that Title Source breached all three of the
and Amendment Number One to the Master Software License Agreement. See Ex. B (Jury
Verdict Form), Qs. 17, 23, 30, 36. Just like the TUTSA and fraud claims, each of
HouseCanary’s contract claims boils down to the same unproven theory of misappropriation
(which is again why the one satisfaction rule applies to prevent duplicative recovery).
(HouseCanary closing 3/14/2018 PM 20). As a fallback, HouseCanary also advanced the theory
that Title Source failed to comply with the agreements’ notice-and-cure provisions. See Ex. B
As demonstrated above, see supra pp. 20-45, there is no evidence of reverse engineering
or any other misuse of HouseCanary’s trade secrets in building MyAVM—in fact, the unrebutted
evidence conclusively establishes the opposite—with no evidence of any harm resulting from
any alleged misuse. See McLaughlin, Inc. v. Northstar Drilling Techs., Inc., 138 S.W.3d 24, 27
(Tex. App.—San Antonio 2004, no pet.) (breach of contract claims require proof of damages).14
14
HouseCanary did not separately prove breach of the MSLA, either, because that agreement
related only to the appraiser app. See PX2 § A (defining “Licensed Software” as the
“proprietary software application known as HouseCanary Appraiser”). The MSLA
prohibited reverse engineering of the appraiser app—not the value reports, which were not
included in the contract until Amendment One. Id., Ex. A, § 2.4 (prohibiting reverse
(Cont’d on next page)
PLAINTIFF’S MOTION FOR JUDGMENT NOTWITHSTANDING THE VERDICT Page 50
Nor is there any evidence that Title Source failed to comply with the notice-and-cure
provisions. Again, the evidence affirmatively forecloses that conclusion. Title Source was
entitled by contract to “immediately terminate” the parties’ agreements if the appraisal app’s
“uptime”—or the ability to use it to access and complete appraisals—fell below 90% for two
consecutive months. PX3, Ex. D, § 2. The evidence affirmatively demonstrated that at no point
during the parties’ two year relationship could Title Source’s appraisers use the app to access and
complete appraisals in the field, as it lacked key features required to complete appraisals. See
infra pp. 80-83. As a result, HouseCanary failed to satisfy the uptime requirement, allowing
HouseCanary’s servers were responsive. DX182 (“Pingdom” report). Even assuming this report
was properly admitted (and Title Source respectfully submits that it was not), it at most
established that the server hosting the app’s Application Programming Interface was reachable.
As HouseCanary’s witnesses admitted, the report provided no information about whether the app
itself was functional. (Rhyne 3/6/18 AM 165-70). The evidence was thus another red-herring; it
is insufficient to establish that HouseCanary satisfied the “uptime” requirement, Title Source was
well within its rights to “immediately terminate” the agreements, and HouseCanary’s contract
engineering of the “Licensed Software”); see PX3 (Amendment One) § 2 (granting license in
“the HouseCanary Value Report”). All of HouseCanary’s purported evidence of reverse
engineering is based on the value reports. With no allegation that Title Source misused the
app—unsurprising given that Title Source could not “use” the app at all—there is necessarily
no evidence of any breach of the MSLA, either.
HouseCanary failed to carry its burden of proving liability on any claim as to any trade
secret—much less all five—and liability is otherwise precluded as a matter of law. But the
record-breaking damages award fails for the independent reason that it is based solely on wildly
speculative opinion testimony that is no evidence of damages. See, e.g., Hunter, 436 S.W.3d at
17 (reversing and rendering based on legally insufficient evidence of damages even while
“presum[ing], without deciding, that the Trade Secrets were trade secrets” and “misappropriated”
by the defendants).
Texas courts routinely grant JNOV where, as here, “the underlying factual basis” of an
expert’s damages opinion is “merely speculative,” Aquila Sw. Pipeline, 48 S.W.3d at 246, or
“based on assumed facts that vary materially from the actual, undisputed facts,” Houston
Mercantile Exch. Corp. v. Dailey Petroleum Corp., 930 S.W.2d 242, 248 (Tex. App.—Houston
Because HouseCanary had no evidence of any use of any alleged trade secret by Title
Source, see supra at pp. 20-22, its damages expert, Walter Bratic, was hard pressed to show any
harm to HouseCanary. Instead of relying on objective measures of value in the record, Bratic
damages. As demonstrated below, even minimal scrutiny shows that Bratic’s opinions are
the trade secrets. (Bratic 3/9/18 PM 127-28 (“I then assumed it would be a two-year time period
for which they would be running these transactions to measure the value of the trade secrets to
TSI as an investor.”)). To support his claim that Title Source could expect to receive over $200
million in value from merely two years of use, Bratic conjured numbers out of thin air,
disregarded the actual contract price of merely $5 million, and committed other serious errors, in
an effort to maximize damages over this short two-year period. (E.g., Bratic 3/12/18 AM 80).
“investor” opinion, and the $64.1 million “reasonable royalty” opinion (Bratic 3/9/18 PM 183)—
woefully overvalue the purported trade secrets and fly in the face of the objective evidence and
market reality.15 This is a case where “damages may be ascertained with precision . . . because
the parties previously agreed on the value [and] an industry standard provides a clear measure.”
Sw. Energy Prod. Co., 491 S.W.3d at 711. The bargained-for contract price for Title Source’s
use of all of HouseCanary’s technology and data (not just its “trade secrets,” and including its
15
In addition, Bratic’s “reasonably prudent investor” opinion is legally barred by TUTSA,
which only allows for damages that reflect “the actual loss . . . and the unjust enrichment
caused by misappropriation,” or in the alternative, the “imposition of liability for a
reasonable royalty.” Tex. Civ. Prac. & Rem. Code Ann. § 134A.004; see also Sw. Energy,
491 S.W.3d at 711 n.7 (“Remedies available under the Act include injunctive relief; damages
measured by actual loss plus unjust enrichment not included in computing actual loss; a
reasonable royalty; and exemplary damages capped at two times actual damages.”).
Although the common law allowed damages for misappropriation to be measured by “the
value a reasonably prudent investor would have paid for the trade secret,” see Sw. Energy,
491 S.W.3d at 711, TUTSA “displaces conflicting tort, restitutionary, and other law of this
state providing civil remedies for misappropriation of a trade secret.” Tex. Civ. Prac. &
Rem. Code Ann. § 134A.007. Because the legislature chose to expressly list some—but not
all—of the measures of damages available under common law, this provision acts to prohibit
the recovery of the omitted measures, including the “reasonably prudent investor” theory.
approaching Bratic’s figures. (See Bratic 3/12/18 AM 80). The record, moreover, shows that
competing AVMs are available for free (Bratic 3/12/18 AM 48-49), and that HouseCanary itself
licensed two different AVMs from Black Knight for unlimited use, along with all of Black
Knights’ data, for less than $1 million per year, id. at 27-29. The record further shows that Title
Source paid $1.2 million per year to FNC, another technology vendor, for use of Compinator, the
AVM that HouseCanary was supposed to replace, along with a suite of other products. Id. at 37,
120.
Bratic was not free to disregard this objective evidence of value, see Sw. Energy Prod.,
491 S.W.3d at 720-21, and yet that is precisely what he did. Bratic’s opinion that the alleged use
of HouseCanary’s trade secrets would be worth tens of millions, or even more than $100 million,
per year strains credulity—and his baseless opinion is no evidence. Without competent evidence
any “causal link between the improper use of trade secrets and [his] damages” opinions.
Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *14. HouseCanary’s witnesses and technical
experts only identified one purported “improper use” of trade secrets that they alleged caused
16
It is preposterous to think that the value of HouseCanary’s purported trade secrets to just one
customer (Title Source) is more than 100 times the total revenue that HouseCanary earned
for the sale of all of its products (not just those containing these alleged trade secrets) to all
of its customers over the same period. Indeed, the “the actual revenues that HouseCanary
made on [its] products containing the trade secrets was $627,095 . . . in 2016.” (Bratic
3/12/18 AM 86). And, for the first seven months of 2017, HouseCanary made a “little over”
$1 million in revenues for all of its products. Id.
MyAVM or otherwise claim to connect his damages opinion with this purported use of trade
secrets. Rather, Bratic claimed to value what Title Source would have paid (either as a
“reasonably prudent investor” or in a “reasonable royalty”) for the trade secrets in January of
2015. (E.g., Bratic 3/9/18 PM 132). But even if Bratic had record support for his speculative
claims about what Title Source was willing to pay in early 2015—and he did not—these claims
had no connection whatsoever to any purported harmful use of HouseCanary trade secrets.
Misappropriation damages must flow from the allegedly harmful use—an expert cannot
invent damages when none exist. Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *14 (claimants
must show “a causal link between the improper use of trade secrets and the plaintiff's damages,”
and “[w]ithout that causal link, [claimants] cannot succeed on a claim for misappropriation of
trade secrets”). Thus, to support his damages opinion, Bratic was obligated to show “that the
alleged misappropriation of trade secrets, as opposed to other factors,” caused actual economic
loss to HouseCanary. Hunter Bldgs. & Mfg., 436 S.W.3d at 20 (citing Rusty’s Weigh Scales &
Service, Inc. v. North Texas Scales, Inc., 314 S.W.3d 105, 111 (Tex. App.—El Paso 2010, no
4006748, at *6 (Tex. App.—Houston [1st Dist.] Aug. 28, 2008, no pet.) (claim that harm from
misappropriation “is not necessarily subject to exact quantification” was legally insufficient to
Bratic did not offer any “causal link” between the alleged improper use in MyAVM and
his damages opinions—indeed, he did not connect his opinions to any improper trade secret use
17
Of course, as already explained, see supra pp. 20-45, HouseCanary’s own experts ultimately
admitted that no HouseCanary trade secrets were used in MyAVM after all.
link between the improper use of trade secrets and the [claimant]'s damages”). Bratic’s opinion
represented the hypothetical value that he believed Title Source would have expected to receive
if it had misappropriated all five trade secrets in early 2015. (Bratic 3/9/18 PM 132 (“I had to
measure the value to TSI at the time of the alleged misappropriation.”)). To support his claim
that Title Source would have paid hundreds of millions dollars for these secrets in January of
2015 (his erroneous “time of misappropriation”), he claimed that Title Source expected at that
time to use the HouseCanary trade secrets millions of times in the “generation of [a] values
report [sic],” either for itself or a third party, Quicken Loans. Id. at 126; (Bratic 3/12/18 AM 52-
53 (Bratic admits that “[a]ll of these numbers that [he] provided to the Jury were the
But there was no connection between these hypothetical “value reports,” or the purported
“expectations” of Title Source in early 2015, and any improper use of HouseCanary’s purported
trade secrets, let alone use in MyAVM. Indeed, Bratic admitted that he did not understand which
trade secrets even existed, much less what the harm from their improper use would be. For
example, when asked “[a]re any of your damage analysis based on the regression AVM,” he
answered “I don’t know what you’re referring to when you refer to ‘regression AVM’” (Bratic
3/12/18 AM 46)—despite HouseCanary’s claim that the “Regression AVM” is one of its trade
secrets.
Simply put, Bratic—and HouseCanary—did not even “attempt[] to measure” what harm,
if any, befell HouseCanary from the alleged misappropriation of its purported trade secrets.
Hunter Bldgs., 436 S.W.3d at 21 (failing to “measure” harm flowing from specific trade secret
misappropriation mandates judgment against trade secret claim). Thus, even if some trade secret
resulted. See Lakeway Reg’l Med. Ctr., 2017 WL 672451, at *14 (“Even if we assume that
[claimants] provided [defendants] with trade secrets, that a confidential relationship was
breached, and that a trade secret was used or disclosed improperly, . . . [claimants] did not show
a sufficient causal link between” misappropriation and damages, and “[w]ithout that causal link,
Bratic’s opinion was not only wholly divorced from HouseCanary’s theory of improper
use, but his patently unrealistic claims about the value Title Source purportedly expected to
derive from the trade secrets were based on a host of implausible and wildly speculative
Bratic first assumed, contrary to record evidence, that a reasonably prudent investor
would be willing to pay for high-volume use of an AVM on a per-use basis, even though a
HouseCanary executive testified that “[i]t doesn’t make sense to pay transaction by transaction”
for bulk uses of information. (Stroud 2/26/18 PM 200). Bratic then assumed—in a conjecture
this Court called “very speculative” and “a stretch” (Bratic Robinson 1/24/18 PM 89)—that Title
Source would expect to use the trade secrets to generate nearly 14 million “value reports” a year,
with 95 percent of those unsubstantiated uses attributed to Quicken Loans Inc., a separate
corporate entity that is not a party to this case. Ex. C (Bratic Presentation), at 15; (Bratic 3/9/18
PM 127; id. at 160). He arrived at that staggering number by adding: (1) his estimate that Title
Source “facilitates” 660,000 appraisals annually and (2) 13.2 million unsubstantiated expected
uses on behalf of Quicken Loans bankers. Ex. C (Bratic Presentation), at 15. But the evidence
trade secret) as part of every appraisal it completes, (Bratic 3/12/18 AM 59), the undisputed
evidence established that AVMs are never used when completing an appraisal. Indeed, the
record is clear that such use is prohibited by the professional rules and regulations governing
appraisers to protect the integrity of the appraiser’s thought process. (Brocker-Querio 2/6/18
AM 16). Bratic acknowledged this, but speculated that even if appraisers cannot “use the AVM
to prepare the final appraisal opinion,” “it doesn’t mean you wouldn’t use an AVM in the
appraisal process.” (Bratic 3/9/18 PM 98). But Bratic cannot rely solely on his own say-so in
asserting that each appraisal that Title Source facilitated would use automated valuation
technology. An expert cannot “improperly presume[] that every [product] sold”—i.e., every
appraisal estimate or appraisal report ever created—is the result of misappropriation based on his
mere ipse dixit. Hunter, 436 S.W.3d at 20. Indeed, Bratic admitted that approximately 90% of
the 660,000 appraisals that Title Source facilitates annually are performed by independent panel
appraisers over whom TSI has no control (Bratic 3/12/18 AM 66-68), and it was pure speculation
for Bratic to assume that any of these independent appraisers would use any of HouseCanary’s
trade secrets in performing their work. Bratic’s reliance on the speculative and unsupported
Bratic’s 13.2 million figure is even more speculative. To reach this figure, Bratic started
with an internal HouseCanary document suggesting that HouseCanary intended its transmission
pipeline to be able to handle up to 50,000 Value Reports per day. DX499; (Bratic 3/9/18 PM
104-05; Sicklick 3/7/18 AM 122-23 (testifying this was the maximum capacity or bandwidth of
HouseCanary’s pipeline for value reports)). But even if this did represent the maximum number
of value reports that HouseCanary could produce, there is no evidence that anyone—Title
every day. By Bratic’s estimate, Title Source accounted for only 660,000 appraisals per year
(approximately 2,750 uses per business day);18 thus, Bratic had to turn to wild assumptions about
Quicken Loans to support his claim that Title Source somehow expected to use valuation trade
Because there was no competent evidence of how and to what extent nonparty Quicken
Loans actually uses any AVM, let alone a Title Source AVM, Bratic could only speculate.
Based solely on an internal HouseCanary email suggesting that Quicken Loans’ bankers
sometimes process 2,000 bulk purchase leads per hour, DX988, Bratic assumed that Quicken
Loans’ bankers use an AVM for each bulk purchase lead and do so 24 hours per day. (Bratic
3/9/18 PM 106-07).19 But Bratic admitted that he had no basis for this claim other than his bare
“assum[ption] that Quicken Loans orders a value report every time [it] receives a customer
18
Bratic calculated his usage numbers based on an assumption of “[t]wenty business days a
month,” i.e., 240 business days per year. (Bratic 3/9/18 PM 110).
19
A “bulk purchase lead” is a list of prospective customers purchased from “some other
company.” (Bratic 3/9/18 PM 105).
20
Bratic repeatedly agreed that his calculation of 13.2 million uses was nothing more than an
assumption. See, e.g., (Bratic 3/12/18 AM 54) (“Q. Well you assumed 55,000 per day? A.
Yes) (emphasis added); id. at 54-55 (“Q. Combined you included 55,000 purchased bulk
leads and customer inquiries that you assumed TSI would provide an AVM value or Value
Report for, right? A. Yes.”) (emphasis added); id. at 52 (“[F]or the other two transactions, the
bulk purchase leads and customer inquiries, I assumed that TSI would be generating the
Value Reports.”) (emphasis added).
There is no evidence that Quicken Loans uses a Title Source AVM for any bulk purchase lead.
But Bratic would have this Court believe, based solely on his say-so, that every cold call to a
purchased lead results in a successful contact with someone who turns out to be actually
interested in a mortgage from Quicken Loans; that each discussion involves the use of an
AVM—and every other trade secret—regardless of whether there is any reason to do so; and that
this uninterrupted streak of success occurs with 2,000 new customers every hour of the day,
including in the middle of the night. See (Bratic 3/12/18 AM 116-17). This assumption is in fact
directly contrary to testimony from Title Source that it expected Quicken Loans to use an AVM
for only a small subset of prospective clients—only for refinance loans, and only for those that
reach the loan application stage. See, e.g., (Germany 2/26/18 AM 23-26) (explaining that there
is “a huge dip in people who are actually interested in doing a mortgage” between the lead
But Bratic was not done assuming. He then piled on an assumption that Quicken Loans
would use the AVMs and other trade secrets an additional 5,000 times a day in response to
21
Bratic has previously been excluded as an expert for precisely this sort of sloppiness. See
Interplan Architects, Inc. v. C.L. Thomas, Inc., 2010 WL 4065465, at *7 (S.D. Tex. Oct. 9,
2010) (excluding Bratic’s damages testimony in copyright case because “Plaintiff is required
to do more than simply identify the Infringing Stores and designate all revenue from those
stores as a result of the alleged infringement”).
22
On top of that, Bratic’s 660,000 figure already covers actual appraisals involving all of Title
Source’s clients, including Quicken Loans. (Bratic Robinson 1/24/18 PM 19-24 (admitting
he could be “double-counting”)).
an “assumed 55,000 [uses] a day,” actually exceeding the 50,000 maximum system capacity that
was his lodestar. (Bratic 3/12/18 AM 54). But again, but there is no actual evidence that
Quicken asks Title Source to use an AVM when responding to any, let alone every, customer
inquiry.
Bratic’s speculation about Quicken Loans transactions is not only unfounded and
benefiting Quicken Loans provides monetary gain to Title Source. Bratic never explained how
or why Title Source would expect to receive hundreds of millions of dollars in additional profit
for uses of trade secrets by “Quicken loan bankers” (Bratic 3/9/18 PM 99-100)—and there is no
evidence that Title Source receives even a penny of value for any transaction performed for
Quicken Loans. That is why the Court was precisely right when it observed that the inclusion of
transactions involving Quicken Loans was “very speculative” and “a stretch.” (Bratic Robinson
1/24/18 PM 89-90).
Implicitly recognizing that it cannot, and should not, recover damages here based on
Quicken Loans’ supposed use, HouseCanary has filed a separate lawsuit asserting the same
alleged misappropriation against Quicken Loans in federal court. HouseCanary, Inc. v. Quicken
Loans, Inc., No. 3:18-cv-01672-WHA (N.D. Cal. filed Mar. 16, 2018, voluntarily dismissed,
May 25, 2018, ECF No. 37); HouseCanary, Inc. v. Quicken Loans, Inc., No. 5:18-cv-00519-FB,
(W.D. Tex. filed May 25, 2018). Any liability on the part of Quicken Loans may be (and as a
23
“Customer inquiries” includes anybody who “logs on to the Quicken Loans web site on the
Internet or calls the 800-number and inquires, ‘What’s your interest rate?’ or ‘What’s your
loan terms?’” (Bratic 3/12/18 AM 53).
speculative damages must be eliminated from the verdict here as a matter of law.
Bratic then completed his inflated and unsupported damages calculation by multiplying
the nearly 14 million unsubstantiated uses of all five HouseCanary trade secrets by an $11 per
use figure that has no basis in evidence or reality and is contradicted by the undisputed evidence.
To invent his $11 figure, Bratic relied on a bullet point in an email indicating that in 2015 Title
Source was paying some unspecified “outside vendor $11 to get an AVM.” (Bratic 3/9/18 PM
116) (citing DX36)). But he ignored the uncontroverted record evidence that this was a limited,
highly specialized use of a composite of multiple AVMs “outside of the production workflow”
(i.e., the business of appraisals and mortgage lending) and was used only for an idiosyncratic,
limited purpose (defending against repurchase demands by investors challenging loans and the
removal of private mortgage insurance). (Petkovski 2/20/18 PM 16-17). Indeed, the very same
email on which Bratic relies goes on to say that “it would obviously be cost prohibitive to use
this service for each of the hundreds of thousands of leads bankers work,” but that is precisely
the use Bratic assumed for his hypothetical $11-per-transaction AVM. DX36.
Bratic’s claim is illogical. The fact that Title Source would pay a high $11 price for an
idiosyncratic, special-purpose AVM is not evidence that it would pay the same price for an
everyday AVM, just as a couple’s splurge on a special dinner for their anniversary is not
evidence that they spend as much for every meal, every day. Indeed, the fact that Title Source
was unwilling to pay $11 for an everyday AVM (a price it described as “cost prohibitive” for
such a use) affirmatively demonstrates that it did not value such an AVM at $11 per use. After
all, the standard Bratic was purporting to apply required him to assess what a reasonably prudent
value a reasonably prudent investor would have paid for the trade secret”). And his “evidence”
is an email stating that Title Source would not pay such a nonsensical amount for the bulk use of
an AVM.
The absurdity is further illustrated by comparison to market reality. See supra pp. 53-54.
Most importantly, as described above, Title Source paid merely $5 million per year to use
HouseCanary’s AVMs—along with the Appraiser App and other related products—an unlimited
number of times. See PX3 § 2.2 & Ex. E (providing license to “specified data and analytics in
the categories summarized in Exhibit E . . . for [Title Source’s] internal purposes” including
HouseCanary’s “Leading edge AVM’s”). The actual contract between the parties negates any
need for Bratic’s guesswork, because it proves the parties valued everything HouseCanary could
provide—including the (never-finished) app, not just the supposed trade secrets—at a bulk rate
of $5 million a year. (Bratic 3/9/18 PM 161). Where, as here, a contract offers “objective
evidence from which more certainty can be gleaned, it is incumbent on the [claimant] to produce
“critical misstep.” Sw. Energy, 491 S.W.3d at 720 (reversing trade secret damages award that
Even if the parties did expect that Title Source would use HouseCanary’s trade secrets
13.86 million times (and, as explained, they did not), that would mean each transaction was
impliedly valued at 36 cents—not $11. Either way, Bratic’s damages calculations are legally
not justified when objective evidence is available,” and here the parties’ contract, as well as
record evidence on the market price for unlimited uses of AVM products, was “objective
evidence that . . . bears directly on ‘the amount that a person desiring to use the trade secret
would be willing to pay for its use.’” Sw. Energy, 491 S.W.3d at 721. Bratic instead based his
opinions on speculative assumptions that are not only untethered to the facts, but starkly at odds
with them, and with demonstrable market reality. His opinion is thus no evidence, and as there is
no legally sufficient basis for the damages award based on his opinion, it cannot stand.
As an alternative, Bratic opined that Title Source would have been willing to pay a
“reasonable royalty” of $64.1 million for two years of use of HouseCanary’s trade secrets. This
calculation suffers from the same fatal flaws as his primary measure. It was based on the same
speculative number of uses, as discussed above, and runs counter to the same objective evidence
demonstrating how the parties, in fact, valued a license for the alleged trade secrets. It also
Bratic began his alternative analysis with the acknowledgement that (despite his earlier
$11 figure) HouseCanary charges customers $3 to $4 per use for small volume purchases of
value reports (50 reports). (Bratic 3/9/18 PM 116-17). Using $3.50 per use as the midpoint
24
Bratic made no attempt to explain why each of the five trade secrets on the jury charge,
lumped together, were worth $11 every time they were used. He also admitted that he has no
evidence connecting the $11 to any current product. (Bratic 3/12/18 AM 76). As explained
above, see supra pp. 54-57, the calculations also combine all of the trade secrets into one
“value,” without any evaluation of the worth of any individual trade secret.
If anything, Bratic’s alternative value-per-use figure of $3.50 further discredits the $11
per use figure, because it shows that not even HouseCanary values its trade secrets at that absurd
price. But even this $3.50 figure is itself speculative and disconnected from reality.
First, while HouseCanary’s website does offer members of the general public “50 value
reports” per month at a price that is “about 3.50 or so” per report (Bratic 3/12/18 AM 77), the
website also makes clear that “large volume” corporate users would be eligible for a customized
“enterprise plan” price. Id. at 79. Bratic admitted that Title Source would be eligible for one of
these bulk purchase plans, id., and thus would not be charged the $3.50. It belies record
evidence and common sense to equate, as Bratic does, the pricing offered for 50 uses with the
pricing for 14 million uses. Second, Bratic ignores the fact that this pricing is for entire Value
no explanation as to why the various trade secrets (whose outputs make up only small pieces of a
Value Report) would be worth the same as a full Value Report itself. Id. at 17-18. And there is
no evidence that Title Source ever created and used value reports or ever planned to. Even
Further, since both the $11 and the $3.50 per use values conjured by Bratic are based on
other products, they provide no evidence of damages under Texas law. In Texas, an expert
cannot claim that the cost to license one product should be the “reasonable royalty rate” to
license another, unless those products are actually the same. Sw. Energy, 491 S.W.3d at 720.
Particularly important here, a royalty rate must be derived from evidence reflecting the actual
value of the trade secret at issue, and not the cost to license something else. Id. That is only
average retail cost for consumers to download a few Value Reports per month—when
negotiating a huge contract for hundreds of thousands of transactions a month involving different
products? Bratic’s speculation defies both controlling precedent and common sense.25
royalty”—both depend on calculating the value that Title Source would have paid for the trade
secrets at the time of the alleged misappropriation (either to own the trade secrets wholesale
under the “investor” theory, or to license the trade secrets under the “reasonable royalty” theory).
In Bratic’s own words, his job was “to measure the value [of the secrets] to TSI at the time of the
alleged misappropriation.” (Bratic 3/9/18 PM 132 (emphasis added)). “[B]ecause the precise
value of a trade secret may be difficult to determine, the proper measure is to calculate what the
parties would have agreed to as a fair price for licensing the defendant to put the trade secret to
the use the defendant intended at the time the misappropriation took place.” Sw. Energy, 491
S.W.3d at 711 (emphasis added) (quotations omitted); see also Univ. Computing Co. v. Lykes-
Youngstown Corp., 504 F.2d 518, 536 (5th Cir. 1974) (“the law looks to the time at which the
misappropriation occurred to determine what the value of the misappropriated secret would be
25
Moreover, Bratic offered no evidence to support his claim that $3.50 represents the value of
using all (or some, or one) of HouseCanary’s alleged trade secrets—he never attempted to
value the trade secrets on an individual basis at all and thus had no support for his lump-sum-
together valuation. There is no evidence that whatever is included in the “$3 to $4” “value
reports” relied on by Bratic (Bratic 3/9/18 PM 116-17) is equal to the value of the five trade
secrets submitted to the jury, as Sw. Energy requires, 491 S.W.3d at 720.
could not properly value any of them. Instead, he simply assumed—with no explanation—that
the misappropriation of every alleged trade secret began in “January 2015.” (Bratic 3/9/18 PM
132, 144 (“I understood the alleged misappropriation occurred shortly after the Master Services
License Agreement was executed.”)). This flawed understanding formed the basis for his entire
valuation. But there is no basis in the record to suggest that any of HouseCanary’s trade
secrets—much less all five on the jury charge—were misappropriated in January 2015. The
record conclusively established just the opposite, because some alleged trade secrets indisputably
did not exist in January 2015––and Title Source did not receive information related to trade
Even if each of HouseCanary’s trade secrets were misappropriated at some later time (and
they were not), and even if each trade secret had some value to Title Source at that later time,
Bratic did not calculate this value—and thus his opinion is no evidence of damages. Id.; Sw.
Energy, 491 S.W.3d at 711 (value must be based on what “the defendant intended at the time the
Bratic, moreover, claimed that Title Source would have expected to begin using
HouseCanary’s trade secrets in January of 2017, and expected to continue using them through
the end of 2018. (Bratic 3/9/18 PM 127-28). But he had no basis for either of these dates. There
26
(See, e.g., Rhyne 3/6/18 AM 112 (AVM outputs were not disclosed to Title Source until July
28, 2015); Rhyne 3/6/18 AM 134-35 (identifying DX759, sent in late December 2015, as the
critical disclosure of the similarity score); (Stroud 2/26/18 AM 125-26) (complexity score
proposal not shared until March 18, 2016); Rhyne 3/6/18 AM 112 (first disclosure of tiny
piece of data compilation was not until July 2015); HouseCanary Opening 1/31/18 AM 86
(describing PX64, sent on June 19, 2015, as the email disclosing the data dictionary that
would be used)).
time the misappropriation took place” (i.e., in January of 2015, by Bratic’s speculation), to use
trade secrets in 2017 and 2018. See Sw. Energy, 491 S.W.3d at 711.
Bratic chose the date range seemingly at random. He vaguely asserted that he chose
January 2017 because “in late 2016 TSI announced that it was bringing a––its AVM making it
available to customers in early 2017.” (Bratic 3/9/18 PM 127-28). But this “early 2017”
MyAVM date is wrong—implementation and actual use of MyAVM did not begin until August
2017. (C. Wang 2/28/18 PM 76-77). And whatever Title Source “announced” about MyAVM
“in late 2016” has no bearing on what Title Source expected almost two years earlier in January
2015 at the time of the alleged misappropriation. See Sw. Energy, 491 S.W.3d at 711 (damages
must reflect “a fair price for licensing the defendant to put the trade secret to the use the
And by the time Bratic offered his opinion, HouseCanary had already publicly disclosed
all of its supposed trade secrets in January 2018, during the trial in this case—permanently
destroying their value. Accordingly, even assuming misappropriation, the damages clock started
in August 2017 and stopped in January 2018, for there cannot be ongoing harm from using what
is public.
Bratic’s opinion fails to base the expected period of use on any facts in the record, and
thus is no evidence of damages. See Houston Mercantile Exch. Corp., 930 S.W.2d at 248
(damages opinion was no evidence where expert provided no evidentiary basis to determine
whether the allegedly misappropriated products “were, in fact, [profitably used] for all, part, or
* * *
witness.” Coastal, 136 S.W. 3d at 232. As the above review of the key components of Bratic’s
opinion shows, that is all he proffered. Neither the facts nor the law permit HouseCanary to
millions of unsubstantiated transactions, with no evidence of actual use, at a per-use value that
flies in the face of objective evidence and reality. There may be difficult cases about where to
draw the line between competent expert opinion and unsupported assertions, but this is not one
of them. Bratic’s testimony is legally inadequate to support judgment for HouseCanary on its
trade secrets claims. The compensatory damages in Jury Question Nos. 39, 43, 44, and 45
cannot stand.
Bratic also offered an opinion—which the jury inexplicably adopted for both fraud and
breach of contract damages—calculating HouseCanary’s lost profits at $33.8 million for each
claim (after discounting to net present value). See Ex. B (Jury Verdict Form), Qs. 13 (fraud), 28
(breach of Amendment One). This opinion is as baseless as the others. “A party seeking to
recover lost profits must prove the loss through competent evidence with reasonable certainty.”
Atlas Copco Tools, Inc. v. Air Power Tool & Hoist, Inc., 131 S.W.3d 203, 206 (Tex. App.—Fort
Worth 2004, pet. denied). Lost-profits opinions “must be based on objective facts, figures, or
data from which the amount of lost profits can be ascertained.” Id. Bratic’s opinion, in contrast,
is sheer speculation and conjecture. It is no evidence, and the lost profits damages in Jury
Bratic calculated lost profits by adding together (1) what Title Source would have paid to
HouseCanary if the contract had remained in place for three years, minus his estimate of
appraisers who, he believed, would adopt the (never-delivered) appraisal app merely because
Bratic’s calculation of alleged lost profits under the contract fails as a matter of law.
Bratic used a three-year contract term to calculate lost profits, reasoning that absent breach, Title
Source would have paid HouseCanary the $5 million per year contract price for the full term of
the contract. (Bratic 3/9/18 PM 164). But as he acknowledged, there was no guarantee the
contract would remain in effect that long. Id. To the contrary, Amendment One specifically
included a provision allowing either party to terminate the contract on purely discretionary
grounds after one year. PX3 § 7.3. It was thus impermissibly speculative for Bratic to calculate
lost profits beyond the contract’s first year, particularly given the fact that the parties’
relationship ended during the first year. See, e.g., Atlas Copco Tools, Inc., 131 S.W.3d at 206–09
renewed every year would be renewed for six consecutive years despite letter terminating
contract); United Way of San Antonio, Inc. v. Helping Hands Lifeline Found., Inc., 949 S.W.2d
707, 711–12 (Tex. App.—San Antonio 1997, writ denied) (“[T]he mere hope for funding
renewable at the discretion of another party will not support recovery of any future funds the
There was no reason to assume that the contract would have continued for an additional
two years when, in fact, it ended during the first year. As a result, Title Source would have paid
assumed.27
Bratic added an additional $29,357,641 to his already inflated damages calculation based
on profits he asserted HouseCanary would have obtained from independent contractors. Ex. C
(Bratic Presentation), at 67. The parties’ contract required Title Source to “make the Licensed
Software available” for use by its appraisers, including those appraisers that it hires as
independent contractors. PX2 (MSLA) § 5.1. Bratic assumed, without record support, that by
2018 50% of the independent appraisers who work with Title Source—10,000 appraisers—
would decide to use HouseCanary’s app to conduct appraisals for other clients (i.e., other
companies not including Title Source). (Bratic 3/9/18 PM 169-71; Bratic 3/12/18 AM 72-73).
He then estimated that these independent appraisers would conduct 4,079,167 additional
appraisals with HouseCanary’s app over three years, at an estimated profit of $7.20 per appraisal
completed on the app.28 Ex. C (Bratic Presentation), at 67; (Bratic 3/9/18 PM 171).
27
Bratic, moreover, assumed that HouseCanary would have a profit margin of 95% on the
contract. This assumption is contradicted by the very evidence from which he purportedly
derived it—an internal summary of HouseCanary’s projected revenue (DX245) that referred
to “95%+ marginal economics.” (Bratic 3/9/18 PM 173). The document itself (plus simple
math) conclusively establish that, at best, HouseCanary’s profit margin was projected to be
around 76-80%—not 95%. At worst—i.e., if HouseCanary was unable to obtain additional
contracts––its profit margin would be negative. See DX245.
28
To estimate the profit from the independent appraisers, Bratic took the $5 million contract
with Title Source (ignoring that the contract price covers a number of other products) and
divided it by the 660,000 appraisals a year conducted by Title Source to conclude that each
appraisal was worth $7.58. (Bratic 3/9/18 PM 176-77). Plugging in his speculative 95%
profit margin, Bratic calculated that HouseCanary would earn $7.20 of profit from each
appraisal conducted using the HouseCanary app. Multiplying this number by the 4.1 million
additional hypothetical appraisals, Bratic concluded that independent appraisers would
generate $29,357,641 in additional revenue for HouseCanary. Id. at 177; Ex. C (Bratic
Presentation), at 67.
bandwagon is wholly unsupported. See, e.g., Ramco Oil & Gas Ltd. v. Anglo-Dutch (Tenge)
L.L.C., 207 S.W.3d 801, 824–25 (Tex. App.—Houston [14th Dist.] 2006 (pet. denied)
(concluding evidence of lost profits legally insufficient because “Plaintiffs’ proof of lost profits
is largely speculative, dependent on uncertain and changing market conditions, and based on
risky business opportunities and the success of an unproven enterprise”).29 The likelihood that
these 4 million appraisals would ever occur depends on a host of unknown factors—not the least
of which is how much the independent appraisers would actually like the app, a functional
version of which was never delivered to Title Source—not to mention the price HouseCanary
Bratic’s calculation further assumed that the adoption by independent appraisers (and
thus the lost profits) would grow over three years, when as a matter of law it was improper for
him to assume continuation of the contract after the first year. And his calculation of profit per
appraisal is not only tenuous, but also rife with internal inconsistency.30
Title Source cannot be held liable for lost profits based on this wildly speculative and
legally infirm set of assumptions which turn on the behavior of third parties outside of its
control. Because Bratic’s lost-profits opinion lacks any “objective, evidence-based support for
29
His conjecture that “50%” of appraisers would start using the app is the ultimate example of
pulling assumptions and numbers out of thin air. (Bratic 3/9/18 PM 169-71). This is nothing
more than mere guesswork.
30
Bratic’s estimate of profit per appraisal (calculated by dividing the $5 million contract by
660,000 appraisals, see supra note 28) is entirely inconsistent with his prior argument that the
contract contemplated that Title Source could pull up to 50,000 value reports a day, which of
course would result in markedly less profit per use.
The parties’ contract contained a bargained-for damage cap that should be enforced in the
event that a verdict on liability stands (and it should not). The parties limited contract damages
HouseCanary offered no argument or explanation why the damage cap should not have been
enforced. While the cap does not apply in cases of misappropriation, there is no evidence such
conduct occurred here. And, in any event, HouseCanary cannot have it both ways. If
HouseCanary’s fraud and breach of contract claims are mere repackaging of its misappropriation
claim, which is indeed the case, all three theories cannot stand under the one satisfaction rule.
But if the breach and fraud claims are independent, they are limited to the contractual cap of $2
million.
As discussed above, HouseCanary claimed, and the jury found, that Title Source
breached all three of the parties’ agreements—the Non-Disclosure Agreement, the Master
Software License Agreement, and Amendment One to the Master Software License Agreement.
At HouseCanary’s counsel’s suggestion, the jury awarded, as contract damages, the trade secret
damages of $201.6 million and $64.1 million for these alleged contractual breaches (in addition
to the $33.8 million separately awarded for breach of Amendment Number One in Question No.
28).31 See Ex. B (Jury Verdict Form), Qs. 28, 43–45. The $201.6 million and $64.1 million
figures represent, respectively, Bratic’s opinion on the investor “value” of the HouseCanary trade
31
Because each of these awards is also duplicative of the award for misappropriation, it is
undisputed that HouseCanary cannot recover this same amount on all four claims
(misappropriation and the three breach of contract claims)—it can at most recover for one.
secrets, which are highly speculative and provide no evidence of damages as discussed above.
These sums are not recoverable as a matter of law as a measure of damages for breach of
contract. Indeed, such measures are inappropriate for valuing breach of contract damages
because they “are governed by different theories.” Smith v. Nelson, 53 S.W.3d 792, 795 (Tex.
App.—Austin 2001, pet. denied). Contract damages typically “restore the non-breaching party
to the same economic position in which it would have been had the contract not been breached––
thus giving the party the benefit of its bargain.” Statewide Bank & SN Servicing Corp. v. Keith,
301 S.W.3d 776, 785–86 (Tex. App.—Beaumont 2009, pet. abated). In other words, contract
In contrast, the trade secret damages Bratic employed here––value of the trade secret and
reasonable royalty measures—focus on the value to the misappropriating party. See Sw. Energy,
491 S.W.3d at 711 (describing these damages theories as measuring “[v]alue to the defendant”
and “a proxy for the value of what the defendant appropriated”). As a result, these awards were
legally impermissible, because these calculations focus on different measures—one on the loss
of a contract expectancy, the other on the unrealized gain to the defendant. Compare Statewide
Bank, 301 S.W.3d at 785–86 (discussing expectation damages), with Sw. Energy, 491 S.W.3d at
711, 726 (discussing trade secret damages). The jury’s damages award in Question Nos. 28, 43,
Even if liability were to stand (and it should not), HouseCanary cannot recover on its
numerous overlapping theories and its wildly speculative damages figures. Instead, what started
out as a contract case should end as a contract case, and at most HouseCanary should recover
alleged breach did not destroy the value of the purported trade secrets, and it did not reduce their
value in the market (as Title Source neither sells its AVM outputs to third parties nor prevents
HouseCanary from marketing theirs to other users). As a result, the entirety of the damages in
this action can be nothing more than the value HouseCanary would have received under the
contract—the contract price of $5 million per year for the two year useful life of its purported
trade secrets.
B. The jury’s grossly excessive punitive damages award violates Texas law and
the federal Constitution.
After awarding compensatory damages of $235.4 million, the jury added a whopping
$470.8 million in punitive damages. This record-breaking award is based on emotion and
caprice, violates both Texas law and the U.S. Constitution, and must be set aside.
Punitive damages “are proper only in the most exceptional cases.” Transp. Ins. Co. v.
Moriel, 879 S.W.2d 10, 18 (Tex. 1994). TUTSA authorizes punitive damages only where a
plaintiff proves “willful and malicious misappropriation . . . by clear and convincing evidence.”
Tex. Civ. Prac. & Rem. Code Ann. § 134A.004. This demanding standard requires “intentional
misappropriation resulting from the conscious disregard of the rights of the owner of the trade
secret.” Id. § 134A.002(7). Punitive damages for fraud similarly require proof by clear and
convincing evidence that the harm results from fraud or malice. Tex. Civ. Prac. & Rem. Code
Ann. § 41.003(a). And malice requires evidence of “a specific intent by the defendant to cause
Texas law is clear that “the intent to commit the tort alone cannot justify an award of
exemplary damages.” Eagle Oil & Gas Co. v. Shale Expl., LLC, No. 01-15-00888-CV, --
misappropriation and misuse of . . . trade secrets is not legally sufficient evidence of malice.”)
(emphasis added). Instead, a plaintiff must prove that the defendant “specifically intended” to
inflict “substantial injury that was ‘independent and qualitatively different’ from the
compensable harms associated with the underlying causes of action.” Horizon Health Corp. v.
Acadia Healthcare Co., 520 S.W.3d 848, 867 (Tex. 2017) (citation omitted). Without this
misappropriation case, rather than the exceptional case involving egregious misconduct and
HouseCanary failed to present any evidence—much less clear and convincing evidence—
that Title Source willfully and maliciously misappropriated HouseCanary’s trade secrets or
otherwise acted with malice. See Horizon Health, 520 S.W.3d at 867. HouseCanary’s counsel
told the jury that “this case is about lying and stealing and cheating,” (HouseCanary closing
3/14/18 PM 62), but lawyer’s argument is not evidence of anything, much less of malice. See
Smith v. O’Donnell, 288 S.W.3d 417, 424 (Tex. 2009) (“the strict standard for proving malice
was not met” by “conclusory statements from [claimant’s] expert and . . . counsel”).
Beyond the alleged “evidence of the tort itself,” Horizon Health, 520 S.W.3d at 867,
HouseCanary “failed to pinpoint any evidence . . . that [Title Source] specifically intended to
harm [HouseCanary].” Myriad Dev., 817 F. Supp. 2d at 976 (emphasis added). There was no
evidence, for example, that Title Source was plotting to drive HouseCanary out of the market.
HouseCanary also failed to offer clear and convincing evidence that Title Source engaged
Amendment One was signed, Title Source disclosed to HouseCanary that it was developing its
own model and was interested in finding out “what’s behind” the complexity valuation and
complexity score. See supra pp. 28-29. This evidence precludes any argument that Title Source
an intent to cause substantial injury,” Rusty’s Weigh Scales, 314 S.W.3d at 112, the jury’s
The Due Process Clause “prohibits the imposition of grossly excessive or arbitrary
punishments on a tortfeasor” and requires “the measure of punishment [to be] both reasonable
and proportionate to the amount of harm to the plaintiff and to the general damages recovered.”
State Farm, 538 U.S. at 416, 426. See U.S. Const. amend. XIV, § 1. “Elementary notions
enshrined in our constitutional jurisprudence dictate that a person receive fair notice not only of
the conduct that will subject him to punishment, but also of the severity of the penalty that a
State may impose.” BMW N. Am., Inc. v. Gore, 517 U.S. 559, 574 (1996). Because “[p]unitive
damages pose an acute danger of arbitrary deprivation of property,” “[e]xacting review ensures
that an award of punitive damages is based upon an application of law, rather than a
Courts must be particularly vigilant to ensure that punitive damages do not violate the
Constitution when, as here, counsel inflames the jury and plays to their emotions.
HouseCanary’s counsel explicitly invited the jury to send a message to “Corporate America.”
(HouseCanary closing 3/14/18 PM 72). In urging the jury to “write in double the amount” of
isn’t about settling the disputes between one party and another. This is about you telling the
United States of America . . . that this . . . is where we draw the line on corporate unethical
behavior.” Id. at 71-72. HouseCanary further asked the jury to “stand[ ] up and . . . tell[ ]
Corporate America, I’ve had enough of this” by putting this case “in the Wall Street Journal.”
Id. at 72. It is in circumstances precisely like these that punitive damages are most likely to cross
constitutional lines. Indeed, the Supreme Court has held that “[a] defendant should be punished
for the conduct that harmed the plaintiff, not for being an unsavory individual or business,” let
alone for perceived deficiencies of all of corporate America. State Farm, 538 U.S. at 423.
must carefully consider three factors: “(1) the degree of reprehensibility of the defendant’s
misconduct; (2) the disparity between the actual or potential harm suffered by the plaintiff and
the punitive damages award; and (3) the difference between the punitive damages awarded by
the jury and the civil penalties authorized or imposed in comparable cases.” State Farm, 538
U.S. at 418. The first two factors cut squarely against the jury’s massive punitive damages
award, while the third does not apply (as there are no civil penalties for TUTSA or fraud).
courts weigh five factors. State Farm, 538 U.S. at 419. “The existence of any one of these
factors weighing in favor of a plaintiff may not be sufficient to sustain a punitive damages
award; and the absence of all of them renders any award suspect.” Id. While due process
requires courts to conduct an “[e]xacting review” of punitive damages awards, id. at 417–18, in
this case little analysis is needed to see that each State Farm factor cuts decisively against the
(2) Title Source did not act with “indifference to or a reckless disregard of the health
and safety of others.” Id.
(4) There is no evidence that Title Source “repeatedly engaged in prohibited conduct
while knowing or suspecting that it was unlawful,” BMW, 517 U.S. at 576, as this
is the only misappropriation case ever brought against Title Source and there is no
prior trade secrets violation.
Because there is no reprehensibility here, the punitive damages award cannot stand.
The punitive damages are otherwise excessive. State Farm holds that where
“compensatory damages are substantial, then a lesser ratio, perhaps only equal to compensatory
damages, can reach the outermost limit of the due process guarantee.” Id. at 425 (emphasis
added). That is the case here, where compensatory damages totaling $235.4 million are
“substantial.” See, e.g., Lompe v. Sunridge Partners, LLC, 818 F.3d 1041, 1073–75 (10th Cir.
2016) ($1.95 million substantial); Boerner v. Brown & Williamson Tobacco Co., 394 F.3d 594,
602–03 (8th Cir. 2005) ($4 million substantial despite “highly reprehensible conduct”). At a
bare minimum, the punitive damages must be no more than half of compensatory damages.
The factors that could justify a higher ratio—an “injury that is hard to detect” or a
“particularly egregious act [that] has resulted in only a small amount of economic damages”—
are conspicuously absent here. See BMW, 517 U.S. at 582. As those factors indicate, one
purpose of the ratio analysis is to ensure that punitive damages achieve their important purposes
million contract with a $2 million damages cap—with zero evidence of any use of the alleged
trade secrets by Title Source, according to HouseCanary’s own expert, and zero evidence of
damages are awarded. Such damages would already be sufficient to satisfy the twin purposes of
punitive damages—to punish and deter—and heaping punitive damages on top would be
arbitrary. See Pac. Mut. Life Ins. Co. v. Haslip, 499 U.S. 1, 21 (1991) (punitive damages awards
may “not exceed an amount that will accomplish society’s goals of punishment and deterrence”).
The punitive damages award violates principles of fair notice. Imposition of the
punitive damages aware would also violate principles of fair notice. Title Source could not
reasonably have known that the dissolution of its business relationship with HouseCanary would
subject it to nearly half a billion dollars in punitive damages imposed by the State of Texas over
alleged trade secrets that it never used, BMW, 517 U.S. at 574, let alone that it would be
III. Title Source is entitled to JNOV on its claim that HouseCanary breached
Amendment One.
The only true breach of contract in this case is HouseCanary’s failure to deliver the
revolutionary product it promised. In its final act, HouseCanary again used smoke and mirrors to
distract the jury from reality—this time, by showing the jury a present-day version of the app
that is nothing like the non-functional product that existed when the contract required delivery.
But the record evidence shows that HouseCanary never delivered a functioning product to Title
Source.
Under Section 4 of Amendment One, HouseCanary committed to having its appraisal app
“deployed in [the] field” by November 1, 2015. PX3 § 4. As explained in Title Source’s partial
functional product that could be used by appraisers in the field––not just a test version. See Ex.
D, Pl’s. Directed Verdict Mot. 4-8, Mar. 12, 2018; Ex. E, Pl’s Partial Summ. J. Mot. 14-16, Dec.
29, 2017. Even if the deadline were some other “reasonable time,” see DeClaris Assocs. v.
McCoy Workplace Sols., L.P., 331 S.W.3d 556, 563 (Tex. App.—Houston [14th dist.] 2011, no
pet.), the undisputed evidence established that the app was never “deployed in [the] field” at any
point during the parties’ relationship: No appraiser ever used the app to complete and submit an
That is because HouseCanary never gave Title Source an app that could function to
complete appraisals in the field. “[W]hen one party to a contract commits a material breach of
that contract, the other party is discharged or excused from further performance.” Bartush-
Schnitzius Foods Co. v. Cimco Refrigeration, Inc., 518 S.W.3d 432, 436 (Tex. 2017). The
materiality analysis should focus on “the nature of [the] agreement,” not the “various component
parts.” Computize, Inc. v. Longhorn Packaging Inc., No. 04-03-00138-cv, 2004 WL 86143 at *2
Here, the “nature” of the agreement was for HouseCanary to deliver an app that Title
Source’s appraisers could actually use to complete appraisals in the field, not to experiment with
in hopes of one day using. Because the app was intended to replace existing software employed
to complete the four most common forms used by appraisers and required by Fannie Mae, it is
hardly surprising that HouseCanary’s CEO admitted that he knew “having those [forms]
complete [would] be critical to rolling [the app] out across Title Source.” (Sicklick 3/8/18 AM
146 (emphasis added)); see PX133. Yet HouseCanary admittedly failed to deliver an app that
So too with the “mobile sketch” feature, which was needed for the appraiser to draw or
“sketch” floor plan drawings, without which “appraisals couldn’t be completed.” (Poindexter
3/2/18 AM 82; Rhyne 3/6/18 PM 140). Because the absence of that feature made it impossible
for the app to complete appraisals, it “prevent[ed] the parties from accomplishing the purpose of
the contract.” Noble Energy, Inc. v. ConocoPhillips Co., 532 S.W.3d 771, 777 (Tex. 2017).
HouseCanary did not deny that its appraisal app was incomplete and non-functional—it
even acknowledged (internally) that Title Source’s “criticisms [were] fair,” and that Title Source
should not be expected to pay for the app as it existed in March 2016. PX309. Implicitly
acknowledging that the deficiencies were material, HouseCanary tried to pin the blame for them
HouseCanary made much of the problems it had connecting the app to the system Title
Source used to manage appraisal assignments and submissions. See, e.g., (Rhyne 3/6/18 PM
143-44). But there was no evidence those problems prevented HouseCanary from delivering an
app that could complete appraisals. Indeed, HouseCanary’s own witnesses admitted that the
absence of a connection would not have prevented appraisers from completing an appraisal.
HouseCanary also tried to excuse its breach by claiming it had deprioritized mobile
sketch and the three forms to perform additional requests from Title Source. E.g., (Poindexter
3/2/18 AM 32-34). But for that excuse to hold water, Title Source would had to have known
about HouseCanary’s claimed trade-off between the app’s foundational features and the
(Tex. App.—Beaumont 2012, no pet.) (“knowledge of all relevant facts” required). But
was getting those “extra” features instead of mobile sketchwriting. PX2 at 12. Moreover, the
only “extra” features that HouseCanary identified at trial were things required by the contract all
along.32
The bottom line is that HouseCanary did not deliver a product that could be used to
complete and submit appraisals—thereby thwarting the basic purpose of the parties’ agreement
without any excuse for the failure to deliver. Title Source is entitled to judgment in its favor on
Jury Question No. 25, and to $8 million in damages for its breach-of-contract claim, (Ugone
2/28/18 AM 49); see also Ex. F (Ugone Presentation), at 68, as well as attorney’s fees.
PRAYER
For the foregoing reasons, Title Source respectfully requests that the Court disregard the
jury’s verdict on all issues submitted, render a take-nothing judgment against HouseCanary on its
counterclaims, render judgment in favor of Title Source on its breach of contract claim as to
Amendment One, and award Title Source damages in the amount of $8 million, plus interest and
costs as permitted by law. Title Source further requests the Court set for hearing the matter of
attorney’s fees to be awarded to Title Source. Title Source prays for such other relief to which it
may be entitled.
32
(See, e.g., Poindexter 3/1/18 PM 42; Poindexter 3/2/18 AM 33-34; Sicklick 3/9/18 AM 78).
the following counsel of record by email in accordance with the Texas Rules of Civil Procedure: