Liveperson, Inc. v. 24/7 Customer, Inc.

Filing 565

ORDER DENYING MOTION TO EXCLUDE CERTAIN TESTIMONY OF LIVEPERSON'S EXPERT WITNESS DR. WILLIAM. Signed by Judge Jon S. Tigar on July 19, 2019. (mllS, COURT STAFF) (Filed on 7/19/2019)

1
2
3
4

UNITED STATES DISTRICT COURT

5

NORTHERN DISTRICT OF CALIFORNIA

6
LIVEPERSON, INC.,

7

Plaintiff,

8
v.

9

[24]7.AI, INC.,

10

Defendant.

United States District Court
Northern District of California

11

Case No. 17-cv-01268-JST
ORDER DENYING MOTION TO
EXCLUDE CERTAIN TESTIMONY OF
LIVEPERSON'S EXPERT WITNESS
DR. WILLIAM CHOI
Re: ECF No. 534

12
Now before the Court is Defendant [24]7.ai, Inc.’s (“[24]7”) Motion to Exclude Certain

13
14

Testimony of LivePerson’s Expert Witness Dr. William Choi. ECF No. 533-3.1 The Court will

15

deny the motion.

16

I.

BACKGROUND

17

The background of this trade secret case2 has been extensively described in prior orders.

18

See, e.g., ECF No. 492. Briefly, Plaintiff LivePerson, Inc. (“LivePerson”) provides online chat

19

engagement services through a digital platform that it sells to website operators. Chat engagement

20

allows website operators to initiate real-time text-based communications with website users

21

directly on the website. LivePerson’s platform attempts to maximize the effectiveness of those

22

chat engagements by identifying when initiating a chat with a particular user is most likely to

23

produce a positive outcome, such as a sale. The platform uses “rules” that recommend initiating a

24
25
26
27
28

1

Because this order contains or refers to material subject to sealing orders, the Court initially filed
the order under seal and provided the parties seven days to propose redactions. ECF No. 559 at 1
n.1. The parties then filed a joint notice stating that they do not request any redactions. ECF No.
563. Accordingly, the Court now files the order on the public docket.
2

The parties also have a companion patent infringement case which is not at issue here. [24]7.ai,
Inc., et al v. LivePerson, Inc., Case No. 3:15-cv-02897-JST (N.D. Cal.).

1

chat based on variables such as the user’s navigation history on the website. LivePerson develops

2

rules and variables specific to each customer and combines that information into a Design

3

Requirements Document (“DRD”), which it shares with the customer. LivePerson’s chat platform

4

also collects and analyzes XML data based on users’ visits to the customer’s websites.

5

LivePerson makes those data reports accessible to its customers for use in their operations, and

6

LivePerson also uses that data to further develop its predictive models. See ECF No. 533-5 (Third

7

Supplemental Expert Report of William S. Choi, Ph.D.) ¶ 17 (“[E]ach of the ‘data’ trade secrets

8

consists of data tracking LivePerson metrics such as web visitor behavior, rule performance, chat

9

volumes and content that is output from LivePerson’s predictive chat platform and can be in the
form of either (i) XML data, or (ii) reports.”); id. ¶ 18 (“The XML data, in contrast, provides

11

United States District Court
Northern District of California

10

granular data about each interactive chat, including the transcripts of the chats themselves, which

12

can be further utilized to draw additional conclusions.”).

13

Thus, for purposes of this order, and as the foregoing would suggest, there are two types of

14

LivePerson trade secrets at issue in this motion: “rules” trade secrets, which involve the

15

combinations of rules and variables just identified, and “data” trade secrets, which consist of the

16

XML data gathered during users’ visits to LivePerson’s customers’ websites.

17

[24]7 also operates in the online chat industry. For a period of time, both companies

18

operated under a series of agreements to market and provide services to mutual customers, through

19

which [24]7 gained access to the rules and data developed for certain LivePerson customers.

20

LivePerson alleges that [24]7 improperly used LivePerson’s trade secret information to develop its

21

own competing chat platform and take LivePerson’s customers. LivePerson further alleges that

22

[24]7 continued to make use of LivePerson’s protected information while providing chat platform

23

services to those customers.

24

LivePerson has retained economist William S. Choi, Ph.D., to render an opinion on

25

damages. The Court previously granted a motion by [24]7 to exclude portions of Dr. Choi’s

26

opinion on Daubert grounds. ECF No. 498; Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579

27

(1993). Specifically, the Court found that “Dr. Choi’s opinion must be excluded because he does

28

not apportion trade secret misappropriation damages among particular alleged trade secrets, and
2

1

offers no methodology for the jury to calculate trade secret misappropriation damages on fewer

2

than all of the 28 alleged trade secrets in the case. And because the initial trial is a bellwether trial

3

in which only 15 of the 28 alleged trade secrets will be at issue, the jury’s verdict will necessarily

4

encompass fewer than all of the alleged trade secrets.” ECF No. 498 at 3 (citation omitted).
Following this ruling, the Court granted leave to LivePerson to submit a “supplemental

5
6

damages report . . . that addresses the Court’s Daubert ruling,” ECF No. 519, which LivePerson

7

did. [24]7 then filed this motion.

8

II.

9
10

United States District Court
Northern District of California

11

LEGAL STANDARD
Federal Rule of Evidence 702 provides:
A witness who is qualified as an expert by knowledge, skill,
experience, training, or education may testify in the form of an
opinion or otherwise if:

13

(a) the expert’s scientific, technical, or other specialized knowledge
will help the trier of fact to understand the evidence or to
determine a fact in issue;

14

(b) the testimony is based on sufficient facts or data;

15

(c) the testimony is the product of reliable principles and methods;
and

12

16
17
18

(d) the expert has reliably applied the principles and methods to the
facts of the case.
Trial courts serve a “gatekeeping” role “to ensure the reliability and relevancy of expert

19

testimony.” Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 152 (1999) (citing Daubert, 509

20

U.S. 579). They should screen “unreliable nonsense opinions, but not exclude opinions merely

21

because they are impeachable.” City of Pomona v. SQM N. Am. Corp., 750 F.3d 1036, 1044 (9th

22

Cir. 2014) (citation omitted). The reliability test under Rule 702 and Daubert “is not the

23

correctness of the expert’s conclusions but the soundness of his methodology.” Id. (citation

24

omitted). “Shaky but admissible evidence is to be attacked by cross examination, contrary

25

evidence, and attention to the burden of proof, not exclusion.” Primiano v. Cook, 598 F.3d 558,

26

564 (9th Cir. 2010). The proponent of the expert testimony has the burden of proving

27

admissibility. Lust ex rel. Lust v. Merrell Dow Pharm., Inc., 89 F.3d 594, 598 (9th Cir. 1996).

28
3

1

III.

DISCUSSION
A.

2

Dr. Choi’s Third Supplemental Report

In his Third Supplemental Report, Dr. Choi addresses the Court’s prior concerns by

3
4

valuing the alleged trade secrets using a three-step process. First, he apportions damages across

5

the particular LivePerson customers with which the specific trade secrets are associated. Second,

6

he allocates damages among the three trade secrets categories: rules, data, and tags. Third, Dr.

7

Choi allocates damages between individual trade secrets within the same category (i.e., among all

8

the rules trade secrets on the one hand and all the data trade secrets on the other).3
As Dr. Choi explains, the “rules” trade secrets consist of more than just the rules

9
10

themselves, but also consist of variables:
In some instances, the “rules” trade secrets comprise different trade
secret information contained within a single document. For example,
one “rule” trade secret may comprise the rules contained in a
LivePerson trade secret document, while a different “rule” trade
secret may comprise the variables contained in the same
document . . . . [T]he rules use the variables to function and therefore
the rules and variables work together to form LivePerson’s predictive
chat technology.

United States District Court
Northern District of California

11
12
13
14
15
16

ECF No. 533-5 ¶ 13.

17

Using that backdrop, Dr. Choi then opines that damages can be further allocated among the

18

individual rules trade secrets based on the total amount of predictive information contained in each

19

trade secret DRD. Id. ¶ 15. The amount of predictive information contained within each trade

20

secret, according to Dr. Choi, is reflected by the number of individual rules utilized by that trade

21

secret, because the time and cost of generating the trade secret increases with the number of rules.

22

In addition, the rules trade secrets for any given customer may
reasonably be assigned a relative value (i.e., an allocated weight)
based on the amount of information (namely, the number of rules)
described in the trade secret document. Specifically, after reviewing
the evidence, the number of rules in a trade secret document is a
reasonable indication of the volume of predictive information
contained therein, and hence the relative importance of the trade
secret within the context in which it is used. That is a reasonable

23
24
25
26
27
28

[24]7’s motion does not address the “tags” category in Dr. Choi’s opinion, but instead confines
its arguments to his opinions about rules trade secrets and data (or XML) trade secrets. The Court
therefore does the same.
4
3

inference given my understanding that generally the time and cost of
developing each rules trade secret increases with the number of rules,
as each separate rule must be individually constructed.

1
2
3

Id. Dr. Choi further opines that the more rules a trade secret includes, the greater the volume of

4

predictive information is included in the trade secret, which provided [24]7 with a

5

commensurately greater ability to take LivePerson’s customers:

6

It is my understanding that the greater volume of predictive
information, or the greater number of rules, provided [24]7 with more
knowledge of LivePerson’s predictive functionality and also provided
[24]7 with the ability to transition customers more seamlessly.
Accordingly, for all these reasons, individual rules trade secrets for a
given customer can properly be valued (weighted) relative to one
another in accordance with the number of rules contained in the
particular trade secret.

7
8
9
10

United States District Court
Northern District of California

11
12

Id. ¶ 16.
As to the XML trade secrets, Dr. Choi determined the value of the XML data as to each of

13

the customers [24]7 allegedly wrongfully appropriated from LivePerson, assuming that for each

14

customer [24]7 misappropriated a sufficient quantity of that customer’s data to construct its

15

competing product for that customer. Stated simply, the value of an XML data trade secret is the

16

value conferred on [24]7 by the transfer of the customer associated with that trade secret.

17

LivePerson proposes that that this value be awarded if it proves that [24]7 has misappropriated “a

18

sufficient amount of LivePerson data to take LivePerson’s customers and cause lost profits.” ECF

19

No. 542 at 5.

20

B.

Rules Trade Secrets

21

[24]7 acknowledges that Dr. Choi has now assigned different values to each trade secret,

22

but argues that the Court should exclude his testimony nonetheless because there is no basis for

23

Dr. Choi’s assumption that each rule should be assigned the same weight. It argues that the

24

Court’s prior admonition that “[s]ome trade secrets are more valuable than others,” ECF No. 498

25

at 4, must also apply to LivePerson’s rules and that Dr. Choi violates this admonition.

26

The Court is not persuaded by this argument. Other courts have approved calculations of

27

damages employing similar per-unit methodologies. For example, the District of Massachusetts

28

recently considered an almost identical argument to the one [24]7 makes here. Iconics, Inc. v.
5

1

Massaro, 266 F. Supp. 3d 461 (D. Mass. 2017). In that case, the plaintiff moved to exclude the

2

defendant’s damages expert’s opinion that apportioned profits by “assuming each customer

3

contributed equally to [defendant’s] operating profit and that each line of code contribute[d]

4

equally to the software’s overall value.” Id. at 476. As [24]7 argues here, the plaintiff in Iconics

5

argued that “some clients, some lines of code, and some trade secrets are more important and

6

valuable than others.” Id. The court rejected the plaintiff’s argument, reasoning that the expert

7

“justified his apportionment methodology and applied it in a defensible and replicable way. The

8

merits of that apportionment may be explored during cross-examination at the appropriate time.”

9

Id.
Similarly, Judge Freeman of this district recently denied a motion to strike expert damages

11

United States District Court
Northern District of California

10

testimony in a patent case in which plaintiffs’ expert “calculate[d] Defendants’ alleged savings in

12

R&D fees by totaling Plaintiffs’ mask design fees for the 139 projects they allege[d] contributed to

13

the development of the accused trade secrets, multiplying this by the ratio of total R&D expenses

14

to total mask fees (to estimate total expenses for the 139 projects), and using surface area ratios to

15

apportion for the features that relate to trade secrets.” VIA Techs., Inc. v. ASUS Comput. Int'l, No.

16

14-CV-03586-BLF, 2017 WL 3051048, at *6 (N.D. Cal. July 19, 2017). In other words, the

17

expert “us[ed] surface area as a proxy for design effort,” id., much as Dr. Choi uses the number of

18

rules as a proxy for the effort required to construct a trade secret. As [24]7 does here, Defendants

19

moved to exclude the expert’s testimony as unreliable, arguing that the expert “[did] not attempt to

20

measure the actual benefit received by Defendants from their alleged trade secret

21

misappropriation.” Id. Judge Freeman denied defendants’ motion to exclude the expert’s

22

testimony, finding that their concerns could be adequately addressed on cross-examination. Id. at

23

*7.

24

In addition to the opinions of these courts, the Court has also considered the testimony of

25

[24]7’s damages expert, Suzanne Stuckwisch. She admitted that there is no defined methodology

26

for allocating damages among trade secrets:

27
28

Q: There is no defined methodology in the law or in this industry
that you know of, is there for allocating damages among trade
secrets?
6

1

A: That is correct.

2

ECF No. 541-8 (Stuckwisch Depo.) at 7-8. In fact, Ms. Stuckwisch conceded later in her

3

deposition that she would apply Dr. Choi’s methodology to allocate damages for the lost profits

4

associated with Allianz, one of the additional customers that [24]7 allegedly obtained from

5

LivePerson. Id. at 426:18-21 (“Q. And then you would have the jury apply Dr. Choi’s allocation

6

methodology for Allianz; correct? A. Yes.”); id. at 421:4-7 (“Q. And am I correct that you are not

7

providing an independent allocation methodology? A. Correct.”). [24]7 does not explain or

8

qualify this testimony in its reply brief.
For all these reasons, and like the courts in Iconics and VIA, this Court finds that Dr.

9
10

Choi’s opinion is sufficiently sound and reliable that it can be admitted at trial.

United States District Court
Northern District of California

11

C.

XML Trade Secrets

12

As to Dr. Choi’s XML secrets valuation, [24]7 argues that each XML trade secret

13

“represents all of the data collected over several years in connection with one of three customer

14

accounts at issue in the first trade secret trial,” but “Dr. Choi offers no methodology for the jury to

15

allocate damages if it finds only a subset of the XML data was misappropriated.” ECF No. 534 at

16

6.
The jury will either find that [24]7 misappropriated enough of a particular customer’s

17
18

XML data to enable it to enroll that customer when it otherwise would not have, or it won’t. If it

19

makes that finding, [24]7 fails to explain why Dr. Choi’s testimony is not an appropriate

20

benchmark for LivePerson’s damages. Instead, it makes arguments about the completeness of

21

LivePerson’s identification of its customer data that are not relevant to the Court’s consideration

22

of Dr. Choi’s testimony.

23

///

24

///

25

///

26

///

27

///

28

CONCLUSION
7

1

For the foregoing reasons, [24]’s motion to exclude Dr. Choi’s testimony is denied.

2

IT IS SO ORDERED.

3
4
5

Dated: July 12, 2019
______________________________________
JON S. TIGAR
United States District Judge

6
7
8
9
10

United States District Court
Northern District of California

11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
8