Apple Inc. v. Samsung Electronics Co. Ltd. et al

Filing 2130

Declaration Reply Declaration Of John R. Hauser In Support Of Apples Motion For A Permanent Injunction filed byApple Inc.. (Attachments: # 1 Exhibit A, # 2 Exhibit B, # 3 Exhibit C, # 4 Exhibit D, # 5 Exhibit E, # 6 Exhibit F, # 7 Exhibit G)(Jacobs, Michael) (Filed on 11/9/2012)

1
2
3
4
5
6
7

HAROLD J. MCELHINNY (CA SBN 66781)
[e-mail]
MICHAEL A. JACOBS (CA SBN 111664)
[e-mail]
RACHEL KREVANS (CA SBN 116421)
[e-mail]
JENNIFER LEE TAYLOR (CA SBN 161368)
[e-mail]
MORRISON & FOERSTER LLP
425 Market Street
San Francisco, California 94105-2482
Telephone: [phone]
Facsimile: [phone]

8
9

Attorneys for Plaintiff and
Counterclaim-Defendant APPLE INC.

10

WILLIAM F. LEE
[e-mail]
WILMER CUTLER PICKERING
HALE AND DORR LLP
60 State Street
Boston, MA 02109
Telephone: [phone]
Facsimile: [phone]
MARK D. SELWYN (SBN 244180)
[e-mail]
WILMER CUTLER PICKERING
HALE AND DORR LLP
950 Page Mill Road
Palo Alto, California 94304
Telephone: [phone]
Facsimile: [phone]

11
12
UNITED STATES DISTRICT COURT
13
NORTHERN DISTRICT OF CALIFORNIA
14
SAN JOSE DIVISION
15
16

APPLE INC., a California corporation,

17
18
19
20
21
22
23

Plaintiff,
v.

Case No. 11-cv-01846-LHK (PSG)
REPLY DECLARATION OF
JOHN R. HAUSER IN SUPPORT
OF APPLE’S MOTION FOR A
PERMANENT INJUNCTION

SAMSUNG ELECTRONICS CO., LTD., a
Korean corporation; SAMSUNG
ELECTRONICS AMERICA, INC., a New
York corporation; and SAMSUNG
TELECOMMUNICATIONS AMERICA,
LLC, a Delaware limited liability company,
Defendants.

24
25
26
27
28
REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

1
2

I, John R. Hauser, hereby declare as follows:

3

1.

I am the Kirin Professor of Marketing at the MIT Sloan School of Management at

4

the Massachusetts Institute of Technology (“MIT”). I have been retained as an independent

5

expert witness in this case. I previously submitted the Expert Report of John R. Hauser, March

6

22, 2012 (the “Hauser Report”). I have attached a copy of my Report as Exhibit A to this

7

declaration. My CV is attached as Exhibit A to the Hauser Report.

8
9

2.

I was asked by counsel for Apple to design and conduct two surveys—one for

smartphones and one for tablets – to determine the price premium, if any, that Samsung

10

consumers are willing to pay for the features associated with the patents at issue.1 My analysis

11

shows that, for both smartphones and tablets, Samsung consumers are willing to pay a significant

12

price premium for the tested features that are covered by the patents at issue. These results

13

demonstrate that there is substantial demand for these patented features. The results were

14

reported and admitted at trial as Plaintiff’s Exhibit 30 as a part of my testimony.

15

3.

On October 19, 2012, Professor Tülin Erdem, Professor Yoram (Jerry) Wind, and

16

Dr. R. Sukumar submitted expert declarations on behalf of Samsung.2 Counsel for Apple has

17

asked me to respond to these declarations. I provide no new opinions and do not expand the

18

Hauser Report or my deposition or trial testimony in this declaration. I provide this declaration to

19

explain why the criticisms in these declarations are mistaken and inconsistent with work

20

previously done by these individuals. Exhibit B contains a list of materials upon which I have

21

relied in coming up with my opinions in this declaration.

22

4.

For the purposes of my declaration, I have summarized the main criticisms of

23

Professor Erdem, Professor Wind, and Dr. Sukumar into four categories. These experts state

24

erroneously: (1) that my conjoint surveys did not measure demand for the features associated with

25
26
27
28

1

The relevant patents are: US 7,844,915 (“’915 Patent”), US 7,469,381 (“’381 Patent”),
and US 7,864,163 (“’163 Patent”).
2
Declaration of Tülin Erdem in Support of Samsung’s Opposition to Apple’s Motion for a
Permanent Injunction, October 19, 2012 (“Erdem Report”); Declaration of Yoram (Jerry) Wind,
October 19, 2012 (“Wind Report”); and Declaration of R. Sukumar in Support of Samsung’s
Opposition to Apple’s Motion for a Permanent Injunction and for Damages Enhancements,
October 19, 2012 (“Sukumar Report”).

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

1

1
2

the infringed patents; (2) that conjoint surveys are hypothetical and do not accurately measure

3

what people are willing to pay in the real world; (3) that my surveys lead to unrealistic results

4

because my survey methodology is flawed; and, (4) that my survey design might have led to bias.

5

Below, I address each of these categories.

6

(1) My surveys measure the extent to which demand is driven by features associated with

7

the infringed patents.

8
9

5.

My assignment was to measure the price premium that Samsung owners would be

willing to pay for the patent-related features. To do this, I used the widely known technique

10

called conjoint analysis. As Professor Wind has written, “[p]ut simply, conjoint is a method that

11

lets us separate out the various features of a product to see how the consumer values each feature

12

against the others.”3

13

6.

Professor Wind and Dr. Sukumar assert that my conjoint study was not designed to

14

measure consumer demand, but only to measure willingness to pay.4 This distinction has no basis

15

because the very definition of consumer demand is based on the concept of willingness to pay.

16
17

7.

At his deposition, Professor Wind acknowledged that willingness to pay is directly

related to consumer demand:

18

“Q: How does the, how does measuring the willingness to pay for a
product with these features relate to your question of determining
whether it is a feature that makes a significant contribution to the
purchaser’s decision to purchase?

19
20

A: Well, if the people would be willing to pay more for a product
with the feature and actually buy a product at a higher price, then
obviously it is indicated that this feature is important for them, and
therefore, it affects their willingness to buy the product.”5

21
22
23

“Q: It is an accepted definition – in economic literature, one of the
accepted definitions for demand is based on willingness to pay,
correct?

24
25
26
27
28

3

Testimony of Yoram (Jerry) Wind before the Copyright Royalty Board, Library of
Congress In the Matter of Adjustment of Rates and Terms for Preexisting Subscription Services
and Satellite Digital Audio Radio Services (“Wind Copyright Board Testimony”); available at:
http://www.loc.gov/crb/proceedings/2006-1/soundex-wind.pdf.
4
Sukumar Report ¶ 5, Wind Report ¶¶ 40–45.
5
Deposition of Yoram (Jerry) Wind, November 7, 2012 (“Exhibit C”), at 20:9-19.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

2

1
A: Yes.”6

2
3

8.

It is a fundamental economic principle that demand is defined in terms of a

4

consumer’s willingness to pay. The following quotes from bestselling economics textbooks

5

explain the relationship between demand and willingness to pay:

6

“[T]he height of the demand curve measures the value buyers place
on the good, as measured by their willingness to pay for it.”7

7
“[Y]ou will sometimes find it useful to think of a demand curve as a
curve that represents a consumer’s ‘willingness to pay’ for a
good.”8

8
9

“The demand curve reflects a consumer’s marginal willingness to
pay: the maximum amount a consumer will spend for an extra
unit.” 9

10
11
12
13
14
15
16
17
18
19
20
21
22
23

9.

Measuring consumer preference lies at the heart of conjoint analysis. As I

explained in my Report, “[t]he general idea behind conjoint analysis is that consumers’
preferences for a particular product are driven by features embodied in that product.”10 I
surveyed actual consumers of the Samsung smartphones and tablets that were accused in this case
to ascertain how their preferences for product features of Samsung smartphones and tablets would
inform their choices. The price premiums that I calculated were derived from their choices and
showed their willingness to pay—and hence demand—for the features I tested.
10.

Professor Wind’s assertion that I acknowledged my analysis did not show

consumers bought Samsung smartphones or tablets because they were equipped with the patented
features misconstrues my testimony.11 I stated only that I did not ask those taking the survey to
compare Samsung smartphones and tablets with other devices. Such a comparison is not
necessary to measure the demand for the patented features in Samsung’s smartphones and tablets.
In fact, restricting respondents’ choices to Samsung smartphones and tablets, and limiting the

24
25
26
27
28

6

Exhibit C, at 117:5-8.
N. Gregory Mankiw, Principles of Economics, South-Western Cengage Learning, 6th
edition, 2012, p. 138.
8
David Besanko and Ronald R. Braeutigam, Microeconomics, John Wiley & Sons, 2nd
edition, 2005, p. 138.
9
Jeffrey M. Perloff, Microeconomics, Pearson Education, 5th edition, 2009, p. 271.
10
Exhibit A ¶ 17.
11
Wind Report ¶ 14.
7

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

3

1
2

universe of respondents surveyed to those who recently purchased one of the smartphones or

3

tablets accused in this case, results in the most directly relevant measure of demand for the

4

patented features in the accused products.

5
6
7
8
9

11.

Notably, Dr. Sukumar’s conjoint survey on behalf of Samsung also did not ask

those surveyed to compare the surveyed product with other devices.12
(2) Conjoint surveys accurately measure real world demand.
12.

In his declaration, Professor Wind asserts that the price premiums measured by my

surveys are not reliable due to the fact that “there are no actual consequences to purchasing more

10

expensive products in a hypothetical survey.”13 In short, he suggests that the surveys are

11

unreliable because actual dollars were not used. This is mistaken, inconsistent with Professor

12

Wind’s own practice, and inconsistent with a large body of academic research in this area. As

13

explained in my Report, the conjoint surveys that I undertook are based on methodology

14

commonly used in the field of marketing research. Conjoint analysis was introduced to the field

15

of marketing research in 1971 and is generally recognized by marketing science academics and

16

industry practitioners to be the most widely studied and applied form of quantitative consumer

17

preference measurement.

18

13.

Importantly, it has been shown to provide valid and reliable measures of consumer

19

preferences, and these preferences have been shown to provide valid and reliable forecasts of

20

what consumers will do (or would have done) under scenarios related to those measured.14 For

21

example, under the auspices of MIT’s Virtual Consumer Initiative, my colleagues and I undertook

22

large-scale tests of the validity of Web-based conjoint analysis. Predictions were highly accurate,

23

predicting future choices consumers would make in a subsequent test with real money at stake

24

and predicting what would happen in the marketplace. One of the scientific papers discussing the

25
26
27
28

12

R. Sukumar, Report: Results from A Survey Measuring Use and Valuation of Four
Patented Features (Patented Features: 460, 893, 711, 871) and A Survey of World Clock Patented
Feature Usage (Patented Feature 055), March 22, 2012 (“Sukumar Survey Report”), Exhibit G:
Screen Capture from On-line Survey.
13
Wind Report p. 36.
14
John R. Hauser and Vithala Rao, “Conjoint Analysis, Related Modeling, and
Applications,” Advances in Marketing Research: Progress and Prospects, Jerry Wind and Paul
Green, Eds., (Kluwer Academic Publishers) 2004, pp. 141–168.
REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

4

1
2

validity test received two awards as the best paper in the marketing sciences literature for 2003

3

(awarded in 2004) and for the best paper based on a dissertation (awarded in June 2005).15

4

14.

Professor Wind’s own testimony and academic writings contradict his criticism.

5

At his deposition, Professor Wind admitted that in over 40 years of experience with conjoint

6

analyses, he can recall only one study in which consumers used actual dollars:

7

“Q: You can’t remember even one that you’ve ever done – out of
all the conjoint analyses that you’ve done, you can’t remember even
one, sitting here today, where the consumers used actual dollars?

8
9

A: Actually I can remember one.

10

Q: One. In how many years you’ve been doing this, 35, 40?

11

A: Since 1970.”16

12

“Q: And, in fact, you’re not alone in this. It’s not as though you’re
the only person out there who is not using actual dollars. In fact,
it’s extremely rare for anybody to use actual dollars in a conjoint
analysis, right?

13
14
15
16

A: Correct.”17
15.

Further, Professor Wind stated that the use of hypothetical dollars does not impact

the accuracy or reliability of a conjoint analysis.

17
18

“Q. Right. And the fact that real dollars are not used is not
something that impacts the accuracy or reliability of the study?

19

A. Correct. Real dollars is not the factor. The real factor is the
realism of the framing of the questions.” 18

20
21
22
23

16.

Notwithstanding the fact that they do not use actual dollars, conjoint analyses are

acknowledged as accurately reflecting consumer preferences, and have been relied upon for
making real-life marketing decisions for decades. Professor Wind acknowledged this in
testimony he gave to the Copyright Royalty Board of the Library of Congress:

24
[C]onjoint analysis is a fixture in the commercial world. Thousands
of conjoint studies have been administered in the 35 years since the

25
26
27
28

15

Olivier Toubia, Duncan I. Simester, John R. Hauser, and Ely Dahan, “Fast Polyhedral
Adaptive Conjoint Estimation,” Marketing Science, 22, 3, (Summer) 2003, pp. 273–303.
16
Exhibit C, at 55:6-13.
17
Exhibit C, at 56:8-13.
18
Exhibit C, at 56:23-57:3.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

5

1
technique was first introduced to marketing. Conjoint analysis
underlies innumerable corporate decisions regarding product
design, pricing, positioning, and segmentation decisions where
millions and even billions of dollars are at stake.”19

2
3
4

…

5

At least as critical as the widespread use of the conjoint
methodology is its validation. The continued and repeated use of
conjoint analysis by industry is the best indication of its proven
validity. In addition, conjoint analysis has been validated by studies
that compare different research methods.20

6
7
8

…

9

Since its introduction, businesses and governments and numerous
academic researchers, consulting firms, and marketing research
firms have embraced conjoint analysis as a reliable and valid
method of analyzing consumer preferences when introducing new
products or studying markets. Indeed, businesses and governments
regularly make billions of dollars of decisions based on the results
of conjoint analyses.21

10
11
12
13
14
15
16
17
18
19

17.

Not surprisingly, the conjoint survey conducted by Dr. Sukumar that Samsung

submitted in this case was also a hypothetical survey that did not use real dollars.22
(3) My survey methods were appropriate and the results are valid.
18.

In my conjoint study, as set forth in my Report and in my deposition testimony, I

performed numerous sensitivity analyses and statistical tests. For example, I calculated market
premiums using two related methodologies, showing that they yielded similar results; I checked
the sensitivity of my results, finding them robust to a number of potential variations to the

20

model;23 I undertook multiple tests of internal validity;24 and I tested for demand artifacts and

21

consumer heuristics.25 These tests confirmed that it is appropriate to use the results to estimate

22

the price premiums of the patented features.

23
24
25
26
27
28

19

Wind Copyright Board Testimony, p. 13.
Wind Copyright Board Testimony, Appendix H at 2.
21
Wind Copyright Board Testimony, Appendix H at 3.
22
Sukumar Survey Report.
23
Exhibit A ¶¶ 96–109. See also Hauser Production April 25, 2012 for an analysis
showing that my results are robust to reweighting the survey data as suggested by Dr. Sukumar.
24
Exhibit A ¶¶ 82–90.
25
Exhibit A ¶¶ 19, 36, 42, 47; Hauser Production April 25, 2012; Deposition of John
Hauser, April 27, 2012 (“Exhibit D”), at 244:16–245:5.
20

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

6

1
(a) Professor Wind’s criticisms

2
3

19.

Professor Wind claims my conjoint surveys yield irrational results. More

4

specifically, he uses a statistical model to predict that when faced with two identical but

5

differently priced items, a number of consumers will choose the more expensive one; or that

6

when faced with two identically priced items in which one is objectively better than the other

7

(e.g., all their features are the same, but one has 16 GB of memory while the other has 32 GB of

8

memory), a number of consumers will choose the objectively less attractive item.26 These claims

9

are based on predictions that Professor Wind makes and not on any irrational choices actually

10

found in the surveys. Professor Wind admitted that his results were based on predictions rather

11

than actual choices and that he would have corrected his report to make that clear.27 Further, the

12

approach is flawed.

13

(1) Professor Wind’s criticism is based on his predictive model and not any

14

actual irrational answers in the survey.

15

20.

Importantly, Professor Wind’s claims are not based on any actual irrational

16

choices made by respondents. In fact, Professor Wind failed to identify a single nonsensical

17

choice by a respondent who took the survey:

18
19

“Q: Okay. So you’re reporting predictions that you think are
nonsensical, not actual results of the survey, not actual responses
from the people who took the survey?

20

A: Correct.”28

21

“Q: My question is, you’re not aware of any instance where even
one person made the choice that you’re predicting would occur in
actual fact?

22
23

A: I cannot answer it in terms of we were or not. We didn’t do the
analysis. There may be a lot of them. We did not do this
analysis.29

24
25
26
27
28

26

Wind Report ¶¶ 46-51.
Exhibit C, at 91:7-92:1.
28
Exhibit C, at 88:23-89:2.
29
Exhibit C, at 95:22-96:2.
27

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

7

1
2

All of the raw survey data from my Report was made available to Samsung shortly after I signed

3

my Report in March 2012. Professor Wind admits that these data were sufficient to have allowed

4

him to test for inconsistencies in the actual responses themselves but he did not do so.
(2) Professor Wind’s predictions are flawed.

5
6

21.

Instead of testing the actual data for inconsistencies, Professor Wind made

7

selective predictions based on a statistical model. Professor Wind’s makes hypothetical

8

predictions by exploiting statistical noise in the model I use; his predictions do not correspond to

9

the actual choices respondents made. By focusing on responses at the individual level (which

10

would emphasize statistical “noise”) rather than responses at the aggregate level (which would

11

emphasize accuracy), his predictions pick up too much statistical noise, which increases the

12

probability of irrational responses. 30 As I discussed in my report and deposition, the data as a

13

whole, which is made up of thousands of choices made by hundreds of respondents, allow for

14

very accurate estimations of market-level information, such as the average market premiums

15

consumers are willing to pay for certain features.31 Professor Wind’s focus on exploiting

16

individual-level noise has no bearing on whether the market-level information is accurately

17

measured.

18

22.

Even at the individual level, Professor Wind’s results are a statistical artifact. In

19

my report I calculated my market premium estimates using two different methodologies, which

20

are known as “first choice” and “randomized first choice,” that yielded largely similar results for

21

the question I was asking.32 In my survey, the two approaches agreed and thus reflect the robust

22

nature of the answers and the estimates.

23
24
25
26
27
28

30

For example, if I have a fair coin, and observe 1,000 flips of a coin from 500
consumers, the probability of heads is very likely close to 50%. If I try to make a prediction of
the fairness of the coin based only on observations from an individual consumer who flips the
coin twice, that prediction is not likely to be accurate since, for example, that individual consumer
has a 25% probability of getting heads twice. See Exhibit A ¶¶ 27–29.
31
Exhibit A ¶¶ 27–29; Exhibit D ¶¶ 234:12–15, 254:17–255:2.
32
Exhibit A ¶ 99 and fn. 67.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

8

1
23.

2

Professor Wind uses only the second approach, known as “randomized first

3

choice,” for his predictions and ignores the “first choice” approach. In fact, for the question he

4

asks, it is “first choice” approach that is the justifiable approach. “Randomized First Choice”

5

begins with the individual-level information that the model has estimated (which as I explained

6

earlier, is inherently less imprecise than aggregate-level information) but then adds a random

7

error term. When used appropriately, the random error term accounts for imprecision and

8

enhances predictive accuracy. But when misused, as in Professor Wind’s analysis, the error term

9

forces a probability that any option, no matter how absurd, can be chosen as a respondent’s first

10
11

choice.
24.

Therefore, if a person wishes to look solely at whether the data predicts irrational

12

choices at the individual level, it is not appropriate to add this random element. Introducing the

13

random element artificially increases the likelihood of irrational choices at the individual level,

14

the level at which Professor Wind performs his analysis. In contrast, the alternate methodology,

15

known as “first choice,” does not introduce the random term described above and is more useful

16

to evaluate whether individuals are making irrational choices. Professor Wind testified at his

17

deposition that he was unable to say whether the irrational results he predicts would disappear if

18

he used the “first choice” analysis because he had not done the work.33 I have done that work and

19

it predicts that all respondents would choose the expected product, as reflected in Exhibit E to this

20

declaration. That is, using the “first choice” approach, the model does not predict irrational

21

choices.

22

25.

Professor Wind also criticizes the survey for coming up with willingness to pay

23

estimates in the smartphone survey that are “comparable to (or greater than) the average price

24

respondents actually paid for their Samsung smartphones.”34 This criticism is invalid. First, the

25

sum of the willingness to pay of the patented features associated with the three patents that were

26

tested is less than what respondents reported that they had paid for their phones. Second, the

27
28

33
34

Exhibit C, at 102:2–11.
Wind Report at ¶ 69.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

9

1
2

retail price of the phone fails to account for the true cost of ownership, which often includes a

3

two-year service contract and can exceed $2,000. This was also discussed in my trial testimony.35
(b) Dr. Sukumar’s criticisms

4
5

26.

Dr. Sukumar’s assertion that my willingness to pay estimates for certain features

6

such as memory and number of apps are low and not realistic36 is conceptually flawed. It is

7

conceptually flawed because Dr. Sukumar did not take into account that the market premiums

8

calculated are averaged over all respondents. Thus, he asserts that Samsung charges $100 to

9

double the storage capacity of one of its tablets,37 but he misses the point that Samsung offers the

10

tablet both with and without the double capacity. Thus, some consumers who are willing to pay

11

the $100 and others are not. As a result, the average willingness to pay is likely to be below the

12

$100 price Samsung charges to double memory capacity.

13

27.

The survey data reflect these differences in individuals’ preferences. In the survey,

14

7% of smartphone respondents value going from 16 to 32 GB of memory by more than the $100

15

market price and 11% value going from 16 to 64 GB of memory by more than the $200 market

16

price.38 This supports the validity of my surveys’ findings.

17

28.

In addition to the conceptual flaw in Dr. Sukumar’s argument, he relies on a

18

selective reading of my analysis. He relies on a robustness check that I undertook and reported in

19

a footnote of the Hauser Report39 and ignores the central results that I discussed at length in my

20

report.40 The central results, referred to as the “market simulation approach,” provide different

21
22
23
24
25
26
27
28

35

Tr. at 1942:23-1943:9.
Sukumar Report ¶¶ 16–18.
37
Sukumar Report at ¶ 17.
38
Comparing two otherwise identical smartphones, one of which has 16 GB of memory
and costs $199 and another which has 32 GB and costs $299, 7% of consumers choose the latter
under a first choice simulation. Comparing two otherwise identical smartphones, one of which
has 16 GB of memory and costs $99 and another which has 64 GB and costs $299, 11% of
consumers choose the latter under a first choice simulation.
39
Exhibit A, fn. 72. This analysis tests whether my estimates of the market premiums
associated with the patented features are sensitive to how I aggregate the individual-level
information to calculate the market-level information within the data. I find that my touchscreen
results are not sensitive.
40
Exhibit A ¶¶ 96–105.
36

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

10

1
2

average willingness to pay estimates for the distraction features that are reasonable, even under

3

Dr. Sukumar’s construction.

4

29.

For example, Dr. Sukumar estimates price premiums for 16, 32, and 64 GB of

5

memory to range from $0.04 to $24 for smartphones and from $10 to $57 for tablets. Using the

6

primary methodology in my report instead, the average price premiums range from $16 to $28 for

7

smartphones and from $26 to $51 for tablets. These estimates are reasonable for the average

8

consumer, given that some consumers may value these features highly while others may not.

9

Similarly, Dr. Sukumar finds premiums below $0.05 for different levels of the number of apps

10

feature, but using the primary methodology in my report yields premiums of $6 to $13. These

11

estimates are reasonable in light of other information that reflects diminishing returns to the

12

number of apps.41

13

(4) There is no evidence that my survey design created survey bias.

14

30.

Professor Wind and Dr. Sukumar make a number of assertions regarding the

15

“design flaws” in my surveys and conclude that this may have resulted in survey bias. As a

16

threshold matter, these assertions are speculative. Neither Professor Wind nor Dr. Sukumar has

17

presented any results that show actual bias. If Samsung believed that my survey design caused

18

bias, it could have obtained evidence to support that claim by conducting a modified version of

19

my surveys to prove this. More than sufficient time existed to do so since March of 2012 when I

20

provided my data and report. In his deposition, Professor Wind acknowledged that it could have

21

been done in the two months after his engagement but he was personally too busy to do so:

22

“Q: You point out in your report that, in your opinion, there are a
number of design flaws with Dr. Hauser’s survey. Did you
consider doing another survey in revising some of those flaws to
see what would happen?

23
24
25
26
27
28

41

Randall Stross, “Microsoft + Nokia = a Challenge for Apple,” New York Times, 4/2/11
(“What is often missed is the diminishing returns after 1,000 applications,” says Thomas R.
Eisenmann, a professor at the Harvard Business School. “If a platform attracts the thousand-mostpopular apps, then it provides almost anything a reasonable person would want to do with a
smartphone.”) available at http://www.nytimes.com/2011/04/03/business/03digi.html.
REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

11

1
A: No. We didn’t have time. This was very, you know, timecompressed. And there was no way I could have designed a survey
in this time period.”42

2
3

“Q: Okay. So if you didn’t have anything else on your plate, sounds
like you could have done it. It would have been tight, but you could
have done it?...In the time you were approached by Samsung.

4
5

A: Yeah. I was not asked to do the design, nor did I propose it
because I basically had many other commitments and could not
have done it.”43

6
7
8

(a) My survey included an appropriate number of features and distraction features.

9

31.

In addition to the patented features, my survey included several distraction features

10

to make the choice task more realistic and to avoid demand artifacts. My conjoint survey had five

11

distraction features, including features deemed important by Professor Wind, such as camera,

12

connectivity, and size and weight.44 Professor Wind nevertheless asserts that my conjoint survey

13

omits relevant features and thus may be biased.45 I do not agree with that assertion.

14

32.

At the outset, I note that Dr. Sukumar designed a conjoint survey in this litigation

15

where he included only four smartphone features, excluding price.46 He included no distraction

16

features and omitted essentially all of the smartphone and tablet features that Professor Wind

17

asserts are important.47

18

33.

An effort to include all features of a product, especially a complex product, would

19

render the results of a conjoint survey unreliable. Therefore, conjoint surveys typically include a

20

limited number of features, usually eight or fewer.48 Professor Wind has himself conducted

21

conjoint surveys where he has used a limited number of features. For example, in a conjoint

22

study of referral decisions of obstetricians, Professor Wind only included seven factors and

23
24
25
26
27
28

42

Exhibit C, at 27:16–23.
Exhibit C, at 28:8–18.
44
Wind Report ¶¶ 70–72.
45
Wind Report ¶¶ 70–72, Erdem Report ¶¶ 37–49.
46
Sukumar Survey Report, p. 3.
47
Wind Report ¶¶ 70–72.
48
Bryan K. Orme, Getting Started with Conjoint Analysis: Strategies for Product Design
and Pricing Research, Research Publishers, 2nd edition, 2010, p. 51. (“Typical full-profile
conjoint studies in practice involve about eight or fewer attributes….”).
43

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

12

1
2

excluded several others, although they were considered important.49 Similarly, in a litigation

3

study of dental lab technicians conducted by Professor Wind, he only included two factors – price

4

and distribution mode.50 Consistent with Professor Wind’s papers, I included six features in

5

addition to price in my conjoint survey of smartphone users.

6

34.

Although I excluded certain features from my conjoint survey, I instructed the

7

respondents to assume that all those excluded features are the same across all smartphones. This

8

instruction ensured that features that were not included would not affect respondents’ choices.

9

This is a standard assumption made by survey designers. Indeed, Dr. Sukumar made the same

10

assumption in his conjoint survey of smartphone users; he instructed respondents to “[a]ssume

11

that all other features of the iPhone 4 not shown to you are the same.”51

12

(b) I used animations where appropriate.

13

35.

Dr. Sukumar and Professor Wind each assert that the partworths of some of the

14

features (touchscreen, camera, and connectivity) may have been inflated because of the survey

15

design. I disagree.

16

36.

First, Professor Wind’s exhibits are not sufficient to demonstrate bias toward the

17

animated features as he claims. They are equally consistent with the conclusion that the features

18

that were animated – camera , connectivity and touchscreen – have higher willingness to pay

19

estimates because consumers, in fact, are willing to pay higher price premiums for them.

20

37.

Second, the estimated willingness to pay for the animated features is not uniformly

21

higher in my surveys than the willingness to pay for features that were not animated. In contrast

22

with Professor Wind’s Exhibit 21 for smartphones, Exhibit 26 provides a different picture for

23

tablets. Levels for certain features that were not animated such as storage/memory and

24

size/weight have greater partworths than some of the touchscreen feature levels that were

25
26
27
28

49

Douglas K. Richardson, Steven G. Gabbe, and Yoram Wind, “Decision Analysis of
High-Risk Patient Referral,” Obstetrics and Gynecology, Vol. 63, No. 4, April 1984, pp. 496-501.
50
Expert Report of Jerry Wind, United States of America v. Dentsply International, Inc.,
June 14, 2000.
51
Sukumar Survey Report, Exhibit G: Screen Capture from Online Survey.
REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

13

1
2

animated. This is consistent with the conclusion that consumers taking the survey were valuing

3

the features themselves, not merely biased by the fact that some of the features were animated.

4

38.

Dr. Sukumar makes a separate assertion that the design of my survey differentiated

5

the touchscreen feature from other features in other ways. In particular, he asserts that the

6

touchscreen feature was differently depicted in the survey, among other ways, through the choice

7

of background colors and how touchscreen levels were crossed out when moving from one level

8

to another.52 Again, Dr. Sukumar does not provide any proof that this differentiation resulted in

9

bias. He also could have tested my survey for design bias in any of these areas but failed to do so.

10

39.

Further, such differentiation also applied to features other than the patented

11

features. My survey has six features in addition to price. The six features are not identical; the

12

descriptions were designed appropriately for the features. For example, the memory feature

13

shows images associated with memory and the apps features shows images associated with apps.

14

Rather than cherry-pick some differences, Dr. Sukumar could also have argued, albeit

15

erroneously, that there were biases because the memory feature is the only one with the color blue,

16

while the number of apps feature is the only one that shows a phone. Dr. Sukumar fails to explain

17

why small differences in the presentation of the patented features led to bias but similar

18

differences in the presentation of the other features did not. Finally, I pretested my survey for

19

such potential biases. None were found.53

20

(c) An “outside option” was not necessary and its inclusion may have created bias.

21

40.

Professor Wind asserts that not allowing respondents to choose an “outside option”

22

(that is, to choose not to purchase the phone or tablet at all) may lead to an inflated willingness to

23

pay for the features that are tested in the conjoint survey. As a threshold matter, the inclusion of

24

an outside option may be necessary when calculating “primary demand” – whether users would

25

purchase a smartphone as opposed to a basic phone. But inclusion of an outside option is

26
27
28

52
53

Sukumar Report ¶¶ 14–15.
Exhibit A ¶ 47.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

14

1
2

unnecessary to determine the extent to which the patented features drive Samsung consumers’

3

demand for the accused Samsung products.

4
5

41.

At his deposition, Professor Wind admitted that the literature on the topic of the

“outside option” is split on whether it introduces its own bias.

6

“Q: Now, Professor Wind, you know that in the literature there are
many articles that debate whether including an outside option can
bias – including outside option can bias the survey and distort the
results in some circumstances, right?

7
8

A: I think there is mixed messages, if you want to, in the
literature.” 54

9
10

42.

Even the publication Professor Wind relies for the assertion that an “outside

11

option” enhances the realism of the survey acknowledges that inclusion of an “outside option”

12

raises concerns, including the creation of bias.55

13

43.

Professor Wind has himself designed conjoint studies without an “outside” option.

14

For example, in one study respondents were asked to allocate 100 points over various alternatives

15

and they could not choose 0 for all alternatives.56 In other words, an outside option of not

16

choosing any of the alternatives was not provided to respondents. In another litigation-based

17

study of travel agents conducted by Professor Wind, respondents were asked to choose from

18

American Airlines flights or Continental flights.57 Respondents could not select an outside option

19

of not choosing either of the airlines. Similarly, the conjoint survey of smartphone users designed

20

by Dr. Sukumar in this litigation does not offer an outside option to respondents. Dr. Sukumar’s

21

survey instructed respondents to “[a]ssume that the three products shown below are the only

22

iPhone 4 products you can choose from.”58

23
24
25
26
27
28

54

Exhibit C, at 130:5-12.
Jeff D. Brazell, Christopher G. Diener Ekaterina Karniouchina, William L.
Moore,Válerie Séverin and Pierre-Francois Uldry, “The no-choice option and dual response
choice designs,” Marketing Letters, Vol. 17, No. 4, December 2006, p. 256.
56
Expert Report of Jerry Wind, United States of America v. Dentsply International, Inc.,
June 14, 2000, Appendix E, Instructions for Responding to Materials in Envelope B.
57
Yoram (Jerry) Wind, Abba M. Krieger, and Paul E. Green, “Applying Conjoint
Analysis to Legal Disputes: A Case Study,” Wharton School, University of Pennsylvania, January
2002, p. 5.
58
Sukumar Survey Report, Exhibit G: Screen Capture from Online Survey.
55

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

15

1
2

(d) My survey did include noninfringing alternatives.

3

44.

Professor Wind and Professor Erdem assert that my conjoint survey did not take

4

into account the noninfringing alternatives that Samsung had for each of the patented features.59

5

They claim that I compared the availability of a patented feature with the non-availability of the

6

patented feature, instead of comparing the patented feature with its noninfringing alternative. As

7

a result, they claim, my willingness to pay estimates are overstated. Their assertions are factually

8

incorrect.

9

45.

My survey included noninfringing alternatives for each of the three alternatives.

10

These non-infringing alternatives were animated and shown at the outset of the survey when the

11

different touchscreen choices were explained to the survey respondents. In his deposition,

12

Professor Wind admitted it was an “oversight” that his report did not acknowledge that the

13

noninfringing alternatives were presented in the animations.60 A copy of the touchscreen

14

animations that were included in the smartphone survey is attached to this declaration as Exhibit

15

F. Very similar animations were presented for the tablet survey, although using a tablet instead of

16

a smartphone. Thus, survey respondents were asked only to compare the relative value of the

17

patented features with noninfringing alternatives, not the value of the patented features with no

18

comparable feature at all.

19

46.

Although his declaration states to the contrary, Professor Wind acknowledged in

20

his deposition that my surveys did include these noninfringing alternatives and that they were

21

presented in animated form.61 While he now seeks to claim that a respondent may not have

22

understood the noninfringing alternative, this criticism rests on a fundamentally false premise. I

23

intentionally designed my surveys to present reasonable noninfringing alternatives, which were

24

chosen based on consultation with the technical experts in this case.

25
26
27
28

59

Wind Report ¶¶ 61–65, Erdem Report ¶¶ 59–62.
Exhibit C, at 80:7-81:8.
61
Exhibit C, at 76:14-77:2.
60

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

16

1
(5) Other criticisms by Samsung’s experts.

2
3

47.

Professor Erdem and Dr. Sukumar also assert that the patented features are

4

unlikely to be driving demand.62 Neither Professor Erdem nor Dr. Sukumar has provided any

5

sort of scientific analysis that would justify their conclusions. Professor Erdem admitted in her

6

deposition that she did not conduct any empirical research at all.63 She also testified that she

7

could not think of any patented feature that would drive smartphone demand.64 As Professor

8

Wind explained in his deposition, the decision to purchase any type of product is typically driven

9

by multiple factors, and it is rare that any one factor would be the sole determinant of demand.

10

“Q. But if the test for whether something were driving consumer
demand is whether that was the sole determinant of the reason to
purchase the product, that test could almost never be met in your
experience?

11
12

A. I would find it kind of strange to find a factor that is a sole
determinant. Perhaps with the exception I mentioned before of a
brand name, that a brand name represents a totality of images,
associations, perceptions of a consumer of the brand; and a
consumer may decide, you know, BMW is such a terrific
brand…;65

13
14
15
16

48.

Professor Erdem also asserts that consumers do not choose as they do in the

17

surveys I designed, but use heuristics.66 Accordingly, in the purchase of a smartphone or tablet,

18

consumers “do not separately evaluate each sub-feature.” Her assertions are problematic for a

19

number of reasons. First, Professor Erdem shows no actual evidence that smartphone or tablet

20

consumers use heuristics (and what sorts of heuristics) when making choices. Second, the use of

21

heuristics can be tested. Accordingly, it would have been entirely possible for Professor Erdem to

22

survey Samsung consumers to test whether they use heuristics. This, she has not done. Third,

23

unlike Professor Erdem, I actually tested for the presence of heuristic-based choices – in

24
25
26
27
28

62

Erdem Report ¶¶ 51–58, Sukumar Report ¶ 10.
Deposition of Tülin Erdem, November 7, 2012 (“Exhibit G”), at 22:4-10; 34:17-35:11;
76:3-8; 165:19-23.
64
Exhibit G, at 26:7-19.
63

65
66

Exhibit C, at 26:9-27:2; see also id. at 12:14-13:15.
Erdem Report ¶¶ 47–50.

REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

17

1
2

particular, the lexicographic heuristic – and found that surveyed respondents did not exhibit such

3

behavior.67

4
5

I declare under penalty of perjury that the foregoing is true and correct and that this
Declaration was executed this 9th day of November 2012, at Cambridge, Massachusetts.

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

67

The lexicographic heuristic is one where, faced with varying levels of features A and B,
a consumer has an infinite preference for feature A. In other words, she would always choose
(even the smallest additional amount of) feature A regardless of the amount of feature B. See also
Exhibit D, at 244:16–245:12.
REPLY HAUSER DECL. ISO APPLE’S MOTION FOR PERMANENT INJUNCTION
CASE NO. 11-CV-01846-LHK (PSG)
pa-1560004

18