District Court for the Eastern District of Virginia Alexandria Division
Employees Insurance Company v. Google Inc.
Civil Action No. 1:04cv507
December 15, 2004
Transcript of bench trial before the honorable Leonie M.
Brinkema United States District Judge
transcription of stenographic notes:
THE COURT: Good morning, counsel.
MR. OSSOLA: Good morning, Your Honor.
MR. PAGE: Good morning, Your Honor.
THE COURT: Let's see, I think when we recessed, Mr.
Page, you were in the midst of your cross examination of Dr. Ford. Do we
need to bring him back in?
MR. PAGE: Briefly, Your Honor.
THE COURT: All right.
MR. PAGE: I have a few additional questions.
THE COURT: All right, Dr. Ford, if you'd come back
up to the witness stand? You're still under affirmation from your testimony
on Monday, sir.
GARY T. FORD, PH.D., PLAINTIFF'S WITNESS,
PREVIOUSLY AFFIRMED, RESUMED CROSS EXAMINATION (Cont'd.) BY MR. PAGE:
Q. Good morning, Dr. Ford.
A. Good morning, Mr. Page.
Q. I'm sorry to have to bring you back, but I just
have a few more questions for you. In your direct, you discussed briefly
the coding tables that are at tab I of your report. Could I ask you in a
little more detail how you go about selecting the categories that you use
A. In general, the coders from in this case Target
Research Group review perhaps 10 percent of the responses as they come in
and read them independently and assign, assign their responses to various
categories. At that point, I review what they've done. I look at the
categories that they've put together and, if necessary, suggest some changes.
Q. Okay. So the categories are a function of what
the actual responses are. You don't select them in advance; is that correct?
A. That's correct.
Q. Okay. And if you'd look at page 5 of tab
I of your report? This is Plaintiff's Exhibit 231.
A. Yes, I'm there.
Q. If you see near the bottom, there's a category
34, and that's labeled "It was a GEICO search/under search/results of search"?
Q. Is that the code that you would use if someone
indicated that their reason for giving a particular answer was that "GEICO"
was the search term?
A. Yes. That seems to be -- there may be
other codes, too, for it, but that would be one code.
Q. But that's what that code would indicate?
Q. And when you put your tables together, if,
if there are no responses in a given code, do you simply leave that, that
code out of the table?
A. Or, or put it in as a zero.
Q. Okay. For example, if you could turn to your
Revised Table 10, which is at tab 10?
Q. These are the responses to Question 1b, which
is the reasons why respondents would click on sponsored links --
Q. -- if they wanted to purchase automobile
Q. And are there any responses on this table that
correspond with the code 35 that we just -- sorry -- 34 that we just saw?
A. I would -- let me be more specific. I would
assume that code 34 would be collapsed into perhaps code -- the first
response on Table 1b.
Q. That says GEICO --
A. In other words, sometimes there's -- if we look
on page 5 here, there are -- and we look on -- and we look under the "GEICO
(Net)," there are several responses that go in the, in the GEICO -- that are
combined under the "GEICO (Net)." Now, some of those responses may be
folded into the responses on Table 10 Revised in the first category.
Q. So if I'm correct, you --
A. And it could also be in the second one, too.
Q. So you coded the answer that I searched for
GEICO and the answer "it says GEICO" into a single category?
A. It could be. I don't know off the top of my
head because of the way the nets were brought.
Q. Okay. And in fact, in Table 10, the
single-most prevalent answer is -- or the two most prevalent answers are "Says
get quotes" and "Says save money," correct?
Q. Okay. Yesterday, you said that your Questions
2 and 3 were more traditional likelihood of confusion questions; is that
A. I did.
Q. Okay. In traditional confusion surveys, do you
typically present the junior and senior marks sitting next to each
other when asking questions?
Q. Okay. But in this case, you really didn't have
much choice, did you?
A. That is correct.
Q. So you end up doing what you call a traditional
confusion survey with the allegedly infringing material and the senior mark
sitting right next to each other, right?
A. They -- there's no other way to do it in this
case, and that's also why I thought Question 1 was more than, too. In
Question 1, obviously, the respondent has the ability to respond by
selecting the place they would click first from the organic results or from
the sponsored links, and so I think you have to have -- that's one of the
reasons I used both of those questions.
Q. I see. So you feel it's less of a problem for
Question 1 than for Questions 2 or 3?
A. I don't necessarily feel it's a problem for
Questions 2 and 4 3. I think it was important to do that, and it's also
important to have a control to try to get at those kinds of issues.
Q. But you would, of course, expect that you'd get
higher affiliation answers when people are staring at the name than an
unaided affiliation question; would you agree?
A. I would think that in general, that's so, but
given the characteristics of this situation, which is that they entered
"GEICO" as a search term and they were asked for one of the sponsored links
where they would expect to go if they clicked on that sponsored link, I
don't see any practical way how to do that other than what was done.
Q. I see. Could I get a look at your survey
questionnaire, which is at tab G, I believe? Tab G. I want to look at your
A. Excuse me, 2a?
Q. No, 3a is actually the one I have.
A. That's it.
Q. That's perfect. Let me see if I understand this
correctly. Question 3 was only asked of people who had not already answered
"GEICO" to Question 1 or 2, correct?
A. To Question 2 only.
Q. Question 2. So any respondent who answered "GEICO" to
Question 2, they were done. They never got to Question 3?
A. They were, they were asked 2b and 2c, but other
than that, they were done.
Q. Right. So that the only people who were
answering Question 3 were people who got what I'll call the right answer
being asked about the sponsored links, in other words, gave an answer that
wasn't "GEICO" but for the most part was the actual link that those
ads went to, correct?
A. The only people who were asked Question 3 were
people who had not given a "GEICO" response to Question 2, which is what it
says in the directions to that question.
Q. Okay. So Question 3a was asked of people who,
for example, had said "netquote," and you then asked them, "Do you think the
company that sponsors this listing is associated or connected with any other
company or companies?" Correct?
A. That's correct.
Q. Aren't you just telling those respondents, "No,
wrong answer. Try again"?
A. Well, you can -- you have the opportunity to
look at the results individually for those, those responses, and there is --
it seems to me those are traditional questions that we would get in a
likelihood of confusion survey, generally followed by a third question that
asked whether there's a need, permission kind of aspects. So in this case, I
was asking two questions that are frequently asked in these types of studies,
and the results are reported separately. It can be aggregated separately so
we can determine whether there is any large jump in responses for the people
who were asked the second question and not the first.
Q. But in fact, you've only asked this -- you've
turned everybody who got the right answer and said, "Try again," and then
you've added them in as confused to the people who said "GEICO" to the
second, while at the same time, you haven't asked this question of people
who said "GEICO," so you haven't given them an opportunity to say, "Well,
yes, I think it may also be netquote." Correct?
A. That's correct.
MR. PAGE: Okay. I have no further questions.
THE WITNESS: But -- thank you.
THE COURT: All right. Mr. Ossola, any redirect?
MR. OSSOLA: No, Your Honor.
THE COURT: All right, thank you, Dr. Ford. You may
THE WITNESS: You're welcome. Thank you.
THE COURT: All right, Mr. Ossola, is there any
other evidence GEICO wants to present in its case-in-chief?
MR. OSSOLA: No, Your Honor, other than the evidence
that's already been submitted on the papers beyond that which is presented
in open court.
THE COURT: That's fine. And just for the record,
because this case went in in a somewhat -- the plaintiff's case has gone in
in a somewhat unusual fashion, for any appellate purposes if that were to be
necessary, we need to clear up what exhibits are actually in evidence. In
particular, Mr. Page, you referred to several of your exhibits during the
course of the plaintiff's case. I believe I have every one of those
recorded, but we would just need to make sure that that's done, and you
should check with Ms. Travers if there's any question about that. All
MR. OSSOLA: Yes, Your Honor.
THE COURT: All right. Mr. Page?
MR. PAGE: Your Honor, at this time, Google would
like to move for judgment as a matter of law in this case.
THE COURT: All right, sir.
MR. PAGE: At the opening of this case, I said that
the Court would be able to dispose of this case before hearing testimony on
damages. That got a bit of a laugh, but I was quite serious. Now that the
Court has heard plaintiff's case, we believe you can rule in Google's favor
as a matter of law. On the key issue in this case, whether the use of
trademarks as keywords violates the Lanham Act, GEICO's own evidence proves
Google's case. To the extent users are confused at all, that confusion
comes from either the use of GEICO's trademark in the text of the sponsored
link or from the natural and nonactionable assumption that any site that
offers to compare insurance rates will likely include GEICO. It's important
to note this assumption is only wrong because GEICO as a business decision
has chosen to make it wrong. That's their choice. As you heard from their
witnesses, it's an integral part of their business strategy to sell only
direct to customers and not to allow other companies to quote rates that may
not be accurate. It's a perfectly legitimate strategy and, combined with a
great marketing campaign, has resulted in a remarkable success story.
That's their choice, but Google cannot be liable for confusion that's a
natural result of that business strategy.What GEICO's evidence shows is that
the use of trademarks as keywords does not itself create confusion. As they
bear the burden of proof on this point, that's the end of the analysis. If
the Court were to deny this motion, this afternoon or tomorrow you would
hear from Google's survey expert, Dr. Jacoby, whose own survey clearly
establishes the same point. When respondents enter the word "GEICO" as a
search term and are presented with sponsored links that offer car insurance
quotes, some of them say yes when asked if they think they can get
information about GEICO there, but they also say yes at a frequency that is
actually slightly higher when asked the same question about Allstate, even
though "Allstate" appears nowhere on the search term, the organic results,
or on the sponsored links. So on the question of whether the use of
trademarks as keywords violates the Lanham Act, the answer is clearly no.
Turning to the question of GEICO's mark in the text of ads and whether that
gives rise to confusion, we believe that the flaws in GEICO's survey, which
is their only evidence on this point, mandate a finding as a matter of law
in Google's favor, but the Court need not throw out that evidence in order
to rule in our favor. As GEICO concedes, such use is already banned by
Google's trademark policy, and Google vigorously enforces that policy.
GEICO has told you that nonetheless, some ads
occasionally slip through, although they have presented no evidence of how
often that occurs, but as I noted in opening argument, the inability to
achieve perfect enforcement of that policy does not give rise to
contributory liability. In order to prevail on that claim, GEICO would have
to establish far more: that Google affirmatively encouraged or knowingly
assisted in a violation of trademark law by the alleged infringers.
There's no such evidence before the Court because
Google does not condone or encourage infringement. Quite to the contrary,
we were the first search engine to implement a trademark enforcement policy,
both because it is the right thing to do and because it makes business sense.
Trademark owners are our own customers on the
advertising side, just as the users are our customers on the search side.
Thus, we carefully balance the dual goals of protecting trademark rights on
the one hand with providing our users with the most complete and relevant
information possible on the other hand. We believe that our current
trademark policy strikes exactly the correct balance, neither over- nor
under-protecting trademark rights. Therefore, we ask that the Court enter
judgment as a matter of law in Google's favor. Thank you.
THE COURT: All right, thank you. Mr. Ossola?
MR. OSSOLA: Your Honor, good morning. Mr. Page
starts out by describing the issue that he believes there's a failure of
proof with respect to as a matter of law as whether the use of keywords --
whether the use of trademarks as keywords in and of itself gives rise to a
likelihood of confusion and claims that GEICO has failed to establish that
proposition, but I would submit to you that that is not the proposition that
GEICO is advancing in this case.
The claim of infringement, the claim of likelihood
of confusion is one that is based on the sponsored links that result from
under Google's system the user's entry of the search term "GEICO," and it is
the relationship between the search term and those sponsored listings that
gives rise to the likelihood of confusion.
One cannot in our view view -- look at the question of the use of "GEICO"
as a keyword by itself as presenting the real question to be decided based
on the evidence presented in this case. And as we have attempted to make
clear, it is the two categories of sponsored listings that at least thus far
are generated by GEICO as a keyword that gives rise to the likelihood of
confusion. One of those categories is the bulk of the sponsored listings
that the searches that we have put into evidence demonstrates, and that is
those that include "GEICO" in the text, the title or heading of the
As to that category of sponsored
listings, as we said in our opening statement, it seems to me that there is
a complete sufficiency of proof, particularly when there is no proof to the
contrary that will be offered by Google to conclude that there is a
likelihood of confusion, and I say that for two reasons. One is simply by
looking at the searches themselves that had -- that establish a connection
between a search term and a sponsored listing. Secondly, Dr. Ford's survey,
if it proves nothing else, certainly establishes based on the sponsored
listings that he tested, both that included the "GEICO" as a search term and
the one that did not, that there is an overwhelming likelihood of confusion
that is associated with sponsored listings that include the trademark in the
text. That is something that's established by his survey.
It's established by the other evidence submitted, the searches
themselves, and particularly given the evidence that came in with respect to
GEICO's focal point of its advertising of its business on GEICO as a, a
trademark that is used to prompt potential customers to search for a rate
quote from GEICO, the fact that a GEICO rate quote cannot be found, the
evidence shows, from any site other than GEICO's, and the fact that a
consumer would be misled into believing that a third-party source that is
offering a comparison of rate quotes that either refers to GEICO
specifically or does not refer to GEICO specifically suggests to the
consumer that they can find a rate quote from GEICO at those sites, and it's
clear, I think, from the evidence that they cannot and that those ads are
Now, Mr. Page refers to the
fact that Google already -- in its policy already provides that sponsored
listings that have the trademark in the text should not be displayed, and as
we pointed out, that has always been the case under Google's policy. Its
policy did not change with respect to that.
And I think the evidence is clear that that has
happened. It has continued to happen, and despite the policy, the reality
is that those sponsored listings have been displayed despite the objection
of GEICO, and in evidence are those objections, in evidence are the numerous
instances in which sponsored listings containing the search term "GEICO"
continued to appear.
So far from slipping through, it seems to me that
the bulk of the search terms that are in evidence establish that those are
the type of categories of listings, that is, with "GEICO" in the title, that
have, in fact, occurred in the marketplace and, we believe, clearly give
rise to a likelihood of confusion.
With respect to the question of whether Dr. Ford's
survey establishes a likelihood of confusion as to those sponsored listings
that do not have "GEICO" in the title, we submit that Dr. Ford's survey is
sufficient to prove that proposition. He did, in fact, test both sponsored
listings that include and do not include "GEICO" in the text. The result --
the respondents in the survey were asked independently to react to each of
those five sponsored listings.
The evidence shows that the responses with respect
to a source affiliation were not appreciably different among those
respondents. There was a slight dimunition in the level of confusion
associated with the fifth sponsored link, but it's still in excess of 50
So we believe that the evidence clearly establishes
that there is a likelihood of confusion based both upon the use of "GEICO"
in the text, and in this case, given the fact that "GEICO" is a mark
associated with rate quotes, that it is also sufficient to establish a
likelihood of confusion with respect to those sponsored listings that do not
contain "GEICO" in the text.
Thank you, Your Honor.
THE COURT: All right. Well, since this is a bench
trial, I could be wrong, but I'm going to assume that this motion is raised
under rule 52(c). I think that's the right rule.
MR. PAGE: That is correct, Your Honor.
THE COURT: And what the Court is being called upon
to do at this point in the trial is to evaluate the strength of the
plaintiff's evidence to see whether or not the case should go forward on any
or all of the issues in this case, and having carefully considered the
plaintiff's case, the Court is going to grant in part and deny in part the
As we all know, there are five essential elements to
a Lanham Act claim. First, the plaintiff must establish that it possesses a
mark that is protectable, and of course, that's not an issue in this case.
"GEICO" clearly is.
The remaining elements require basically a focus on
what the defendant does. Question No. -- the second element is that the
defendant uses the mark; three, that the defendant's use of the mark in
commerce; four, that the defendant's use of the mark is in connection with
the sale, offering for sale, distribution, or advertising of goods and
services; and fifth and what has in my mind been the key issue in why I
could not grant summary judgment initially in this case or grant the motion
to dismiss that was previously filed was that the use of the mark is done in
a manner that's likely to cause confusion.
Having heard the plaintiff's case, the Court is
satisfied that the plaintiff has not established that the mere use of its
trademark by Google as a search word or keyword or even using it in their
AdWord program standing alone violates the Lanham Act because that activity
in and of itself, there's no evidence that that activity standing alone
The Court also finds that there was insufficient
evidence presented in the plaintiff's case to let this case go forward on
the question of whether Google violated the Lanham Act after it barred -- or
after it began to bar the use of the GEICO mark from either the titles or
the text of the sponsored ads that appear as a result of use of the AdWord
And the reason I find insufficient evidence of that,
as we sort of -- I sort of indicated during the testimony of Dr. Ford's, is
that, frankly, Dr. Ford either wasn't asked or chose not to actually query
that particular issue because the survey focused on either ads that had "GEICO"
in the title -- I mean, the page that people were looking at had five
sponsored ads, the first four of which had "GEICO" either in the title and/or
in the text. The fifth one did not, but it was on a page that had the four
previous sponsored links.
The test -- the control in this case used Nike, but
what I don't feel was presented to the Court that needed to be presented to
the Court would have been ads of insurance that did not have "GEICO" in it,
and I was not satisfied therefore that the plaintiff's evidence was
However, as to the narrow issue of whether the
plaintiff has presented sufficient evidence at this point to let the case
continue on the issue of whether Google is liable for violating the Lanham
Act for those sponsored ads using GEICO's name in either the title or the
text that appear next to a -- and a GEICO organic listing as a result of the
AdWord program, at this point, there has been enough evidence of confusion
as a result of Dr. Ford's report to deny the motion and to allow the case to
continue so the Court can consider the defendant's evidence as to whether or
not this situation creates a likelihood of confusion in the marketplace and,
if so, what damages might be appropriate.
Now, counsel, is my ruling clear, and any questions
about what I have just done?
MR. OSSOLA: Your Honor, I have just one.
THE COURT: Yes.
MR. OSSOLA: You made reference to the -- what you
believed to be the insufficiency of evidence after Google began to bar the
use of the trademark in the sponsored link. I think that's what you said.
I just want to make clear that Google has always barred --
THE COURT: And I should clarify that. I realize
that, but in this particular case, they -- I meant to confine that tothe
facts of this case; that is, I didn't mean to suggest that Google did not
have that policy in place, but the reality was that we do have evidence that
there were -- "GEICO" was being used in titles and in text until Google was
advised to stop that, and then following its own policy, it implemented that
MR. OSSOLA: I understand.
THE COURT: All right.
MR. OSSOLA: But the insufficiency evidence on that
point is that with respect to those sponsored listings that do not have "GEICO"
in the text or heading of the ad, as to those -- that category of sponsored
listings, the Court is finding that there is not sufficient evidence of
likelihood of confusion.
THE COURT: In this case.
MR. OSSOLA: In this case.
THE COURT: Correct.
MR. OSSOLA: Thank you.
THE COURT: All right? Mr. Page, was there anything
you needed clarified?
MR. PAGE: No, I think that covers it, Your Honor.
THE COURT: All right. Now, I think as we all know,
there's serious legal issues that are involved in this decision that I've
just rendered, in particular, the status of how trademarks as used in
keywords, etc., and I think I would like time to write a more detailed
opinion on these legal issues, and so depending upon how you-all want to
proceed, as you know, I've also been encouraging you to see if you can
resolve the case, and I've given you now some very clear parameters as to
how the case would continue.
Unless there's any objection, what I propose is that
we terminate the trial at this point, I don't mean end it, but stop right
now what we're doing to give the Court a brief amount of time, which given
the holiday season might be two or three weeks, although we're going to try
to do it sooner than that, to get a written opinion out on this ruling,
consistent with this ruling, and to allow you-all the opportunity to see
whether or not there can be a resolution of what is left in the case. Does
that meet with all of your approval?
MR. PAGE: Yes, Your Honor, that would be fine.
MR. OSSOLA: Yes, Your Honor.
THE COURT: All right. Now, I'm not going to set
another date at this point for when we would start up Round 2 if that
becomes necessary. We'll get the opinion out as quickly as possible, and
then you can get back to me as to how you want to proceed.
In the meantime, since this courtroom is not going
to be used again until January 3, and I'm hoping we have it resolved by then,
you can if you want leave your exhibits up here. If the case ultimately has
been resolved, then there's no need for any of the exhibits to stay at the
courthouse, and we'll ask you to come and get them.
If the case goes on to the, what I would call the
second phase, then obviously, we'll need to get the appellate record cleaned
up in terms of just what exhibits are in or out. All right?
MR. OSSOLA: Yes.
THE COURT: Anything further we need to address at
MR. OSSOLA: Your Honor, just the second phase, if
we get there, would be -- would include damages as well; is that correct?
THE COURT: Yes. It would be the defendant's
opportunity to rebut liability, and obviously, if I were to find that as a
matter of law there was no issue left on liability and find in the
defendant's favor, then we wouldn't have to have a damages phase, but if I
find liability, then there would be a damages phase. All right?
MR. OSSOLA: What's been established to date then is
that the category of sponsored listings that have "GEICO" in the title and
text are -- have been found to be confusing and misleading under the Lanham
THE COURT: No. I have found that you've presented
enough evidence at this point to avoid a motion for judgment as a matter of
law. In other words, there certainly is a prima facie case that you've
established that they are confusing based on Dr. Ford's survey.
Now, I haven't heard the rebuttal evidence, so I'm
not making a specific finding on that. I'm at this point weighing the
quality of the evidence, the amount of evidence. You've presented enough to
let the case go forward.
MR. OSSOLA: Thank you, Your Honor.
THE COURT: All right?
MR. PAGE: Thank you, Your Honor.
THE COURT: All right? And, counsel, that one piece
of paper that was on each of your chairs was a joke. It was not serious.
But we happened to see it and thought it was appropriate.
All right, I wish you-all a happy holiday, and we'll
get back to you as quickly as possible. We'll recess court for the day.
MR. OSSOLA: Thank you, Your Honor.
MR. PAGE: Thank you, Your Honor.
(Recess from 10:31 a.m., until 11:09 a.m.)
THE COURT: All right, for the record, Mr. Ossola
has brought to the Court's attention the following matter: My understanding,
Mr. Ossola, is that your view of the defendant's evidence is that Google
will not be able to offer any evidence to counter GEICO's evidence that
those sponsored sites that contain "GEICO" either in their title or in their
text that have been at issue in this case are confusing. Is that correct?
MR. OSSOLA: That's correct, Your Honor.
THE COURT: And it's my understanding that, what,
the expert -- is it Dr. Jacoby, Mr. Page? Whose report would have --
MR. PAGE: That's correct. It would be Dr. Jacoby's report.
Mr. Ossola is correct. Our survey addressed Google's current policy, so the
sponsored links that we tested do not contain the word "GEICO." We were
testing the keyword issue.
THE COURT: All right.
MR. PAGE: So Dr. Jacoby's survey says nothing
either way as to whether a link with GEICO is confusing.
THE COURT: All right. Then the Court should have
clarified its earlier -- well, it couldn't clarify -- should have said in
its earlier ruling and says now in response to your motion that the evidence
before this Court does establish that those sponsored sites that contain "GEICO"
either in the title or in the text are likely to confuse for purposes of the
Lanham Act requirements, and therefore, to the extent that the
defendant's motion was based on that issue, that will also be denied.
And that reduces the issues that would have to go
forward were there -- if there is a second phase to this trial. The sole
issues that would remain -- and correct me if I'm wrong -- are, No. 1,
whether or not Google is liable for any Lanham Act violation based upon
those sponsored sites, and 2, if Google were liable, then what damages would
Now, is that a correct phrasing from your
standpoints as to what would be left in this case?
MR. PAGE: That is correct, Your Honor. The issue
that will remain is, is since the Court has found a likelihood of confusion
for ads that have "GEICO" in them, the question is whether Google is
contributorily liable or whether only the advertiser would be liable.
THE COURT: Well, actually, because the advertisers
are not in this case, the only question before us would be whether or not
Google is liable.
MR. PAGE: That's correct, Your Honor.
THE COURT: Right. Mr. Ossola?
MR. OSSOLA: I believe that's right. Under
contributory liability or any other theory that may be asserted by GEICO,
that remains to be seen.
THE COURT: Correct.
MR. OSSOLA: Phase 2.
THE COURT: So there has been -- just so we're clear
for the record, there has been no finding of liability -- there's been no
finding that Google is liable at this point. What we have found, however,
is that this particular group of sponsored sites does violate the Lanham Act.
MR. OSSOLA: Yes, Your Honor.
THE COURT: All right?
MR. PAGE: That's correct.
THE COURT: Do you think we need to clarify anything
further for the record?
MR. OSSOLA: No, Your Honor, thank you.
THE COURT: All right. Thank you very much for
bringing that to our attention. We'll recess court once again.
(Recess at 11:12 a.m.)
CERTIFICATE OF THE REPORTER
I certify that the foregoing is a correct transcript
of the record of proceedings in the above-entitled matter.
Anneliese J. Thomson