Google Inc. v. American Blind & Wallpaper Factory, Inc.

Filing 47

Google Inc. v. American Blind & Wallpaper Factory, Inc.

Doc. 47 Att. 3

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 1 of 27 1

UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF VIRGINIA ALEXANDRIA DIVISION GOVERNMENT EMPLOYEES INSURANCE COMPANY, Plaintiff, vs. GOOGLE, INC., and OVERTURE SERVICES, INC., Defendants. . . . . . . . . . . . . . . . . . . . . . . . Civil Action No. 1:04cv507

Alexandria, Virginia November 19, 2004 10:00 a.m.

TRANSCRIPT OF MOTIONS HEARING BEFORE THE HONORABLE LEONIE M. BRINKEMA UNITED STATES DISTRICT JUDGE APPEARANCES: FOR THE PLAINTIFF: CHARLES D. OSSOLA, ESQ. CHRISTOPHER WINTERS, ESQ. Arnold & Porter LLP 555 Twelfth Street, N.W. Washington, D.C. 20004 and JOHN F. ANDERSON, ESQ. Troutman Sanders LLP 1660 International Drive, Suite 600 McLean, VA 22102 MICHAEL H. PAGE, ESQ. Keker & Van Nest, LLP 710 Sansome Street San Francisco, CA 94111-1704 and MELANIE D. COATES, ESQ. Wilmer Cutler Pickering Hale and Dorr LLP 1600 Tysons Boulevard, Suite 1000 McLean, VA 22102 (Pages 1 - 27) COMPUTERIZED TRANSCRIPTION OF STENOGRAPHIC NOTES

FOR DEFENDANT GOOGLE:

Dockets.Justia.com

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 2 of 27 2

1 2 3 4 5 6 7 8 9

APPEARANCES:

(Cont'd.) DAVID S. FLEMING, ESQ. Brinks Hofer Gilson & Lione 455 North Cityfront Plaza Drive Suite 3600 Chicago, IL 60611 and RICHARD C. SULLIVAN, JR., ESQ. Reed Smith LLP 3110 Fairview Park Drive, Suite 1400 Falls Church, VA 22042 MICHAEL KWON, ESQ. ANNELIESE J. THOMSON, RDR, CRR U.S. District Court, Fifth Floor 401 Courthouse Square Alexandria, VA 22314 [phone]

FOR DEFENDANT OVERTURE:

ALSO PRESENT: OFFICIAL COURT REPORTER:

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 3 of 27 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Google. THE CLERK:

PROCEEDINGS Civil Action 2004-507, Government Employees Will counsel please

Insurance Company, et al. v. Google, Inc. note their appearance for the record. MS. COATES:

Melanie Coates, local counsel for Google,

along with Michael Page from Keker & Van Nest. MR. PAGE: THE COURT: MR. PAGE: Good morning, Your Honor. Good morning. With me is Michael Kwon, in-house counsel at

MR. ANDERSON:

Good morning, Your Honor.

John Anderson,

Chuck Ossola, and Chris Winters here on behalf of the plaintiff, GEICO. THE COURT: Good morning. Now, just so I'm clear,

Overture has not filed any dispositive motion, correct? MR. OSSOLA: THE COURT: That's correct. So I'm not looking at an Overture motion a

week or two down the road? MR. OSSOLA: MR. SULLIVAN: No, Your Honor. That's my understanding. This is

Good morning, Your Honor.

Richard Sullivan, with Reed Smith. before from the Brinks Hofer firm. THE COURT: yourselves. All right.

You've met David Fleming We're here for Overture. Well, good, you can answer

You've not filed anything? That's correct, Your Honor. We have not

MR. FLEMING:

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 4 of 27 4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

filed and do not plan to file a summary judgment -THE COURT: All right. But you did appear at the

pretrial conference, and you've filed your list of witnesses and exhibits, etc.? MR. FLEMING: THE COURT: sure. All right. Well, what we have here are the -- let me I have GEICO's motion for Yes, we have, Your Honor. All right, very good. I just want to make

address the administrative issue first.

leave to file under seal the complete memorandum in opposition to the motion for summary judgment and certain exhibits. I'm just sort of letting the word get out, this sealing business is becoming a real problem, and the developing case law in the Fourth Circuit convinces me that the standard for sealing documents, especially when you get to the summary judgment stage, is extremely high threshold. I think it's going to be an ongoing problem, but I must tell you-all that my practice is going to be I'm not spending a lot of my time worrying about sealing things. If lawyers present

me with a really narrow sealing request that is based on compelling justification, then obviously, I will grant that motion, but if it's a broad request and I'm finding that I'm spending a lot of time looking at documents for which there's no reasonable basis to be sealing, then I'm not going to spend my time looking for the one or two meritorious issues that might be

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 5 of 27 5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

amongst that. That's how I take attorney-client privilege documents the same way. If you give me a hundred and I do a random sampling

and the first five I pick out are not subject to the privilege, I'm not going to waste my time looking at the other 95. it's been waived. Now, in this case, a couple of the things about percentage of income based upon this kind of business, etc., some of those things I can understand of a proprietary nature, although now that Google is public or going public, I don't know where you are at this point, probably a lot of that stuff is actually not so much in the private domain anymore, but some of the specific attachments, I didn't see any way in which they would really truly be subject to a seal, and as I said, the Fourth Circuit's just, what, in the last two weeks in that Washington case sent very clear signals that it doesn't expect litigation to be done in privacy. So I'm torn. I think you did better than most of the lawyers I've had recently in terms of trying to be somewhat selective, but I still think it was broader than was necessary, and if this case were being heavily watched by the media, we would most likely have a motion from them to unseal. So I'm going to go ahead and be generous at this time rather than -- and I'm going to grant the motion, and so you can file under seal the original unredacted, and we'll leave the case I think

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 6 of 27 6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

as it is, but, I mean, this is sort of a message for the future that it's got to be much more narrowly tailored or I won't be granting any sealing. MR. OSSOLA: THE COURT: All right? Yes. However, I am not going to struggle if I

have to write an opinion in this case with -- because again, I'm working off of a national security background, where you have to run it by the CIA to get anything even published as a judge. not going to live like that in this case. So assuming we have a trial and I have to do findings of fact and conclusions of law, I'm not going to worry nor am I going to ask my law clerk to struggle with making sure that we don't include proprietary information in that opinion. If it's come out I'm

during the trial or if it's in the papers I've had to consider, it's coming out. So I just want you to know I cannot spend the time doing that. It's very labor intensive, as you know, because you went Okay?

through the process of redacting stuff yourselves. MR. OSSOLA: THE COURT: MR. OSSOLA: THE COURT: problem. All right. Understood, Your Honor. All right.

And the CIA is not involved in this case. Thank goodness. But in any case, that is a

So I've granted GEICO's motion for leave to

file, and that leaves then just the defendant Google's motion for summary judgment.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 7 of 27 7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that.

All right.

Again, as you know from our previous round

on the motion to dismiss, it was my view that I could not grant that motion in part because this issue about likelihood of confusion is such a fact-specific issue that it would need to be played out, you know, in a trial forum. I must tell you honestly I still think, although it has become a more solid basis right now, I still think that most likely that's where this case has to ultimately be resolved, but I'll hear some argument if you want to -MR. PAGE: Okay. Hopefully, I can change your mind on

In assessing the likelihood of confusion evidence and the

likelihood of confusion in this case, it's important to separate out the two very different theories of liability put forward by GEICO. The first theory is that the use of their trademark in the text of an advertisement gives rise to a likelihood of confusion. That theory at least on an injunctive level applies to

both Overture and to Google on a liability basis but not on an injunctive basis, because unlike Overture, Google does not permit advertisers to use the, use the GEICO trademark in the text of their ads now, so there's no point in an injunction that would require us to stop doing something we don't do. The second theory, it's a very different theory, is that the use of "GEICO" as the keyword that triggers an advertisement itself gives rise to a likelihood of confusion. That is what

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 8 of 27 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Google allows advertisers to do because we believe first that it is not a trademark use at all, but we've already addressed that issue with Your Honor, and second, that it does not give rise to a likelihood of confusion, and on that issue, there are no triable issues of fact anymore because GEICO's own survey, although it didn't set out to do it, tested that theory conclusively. As you'll recall, GEICO's expert, Dr. Ford, used in his survey a stimulus, which was a Google search page -- and it's Exhibit B to his report, which is Exhibit B to the Coates declaration -- he used a stimulus that consisted of a Google search page where the search term that was entered was "GEICO." The sponsored links consisted of five sponsored links. Four of those sponsored links contained the word "GEICO," and all five of those sponsored links offered comparison rates for insurance, get insurance rates here, free insurance quotes, that sort of thing. And he tested confusion on that stimulus.

And all of the experts agree that there are three possible sources of that confusion: One, what term did you search

for; two, does the trademark appear in the ad and does that give rise to the confusion, which was true of four of the five ads; or three, is it background confusion that's simply the result of the context that's nonactionable. As an example, if you put up a shoe store with a sign that says "Hundreds of Name Brands Here," you don't use anybody's trademark at all, and you survey 100 people in front of that store

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 9 of 27 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

and say, "Do you think this store sells Nike shoes?," many of them are going to say, "Yes." It's a perfectly logical supposition

that they have hundreds of name brands, "Sure, yeah, I think they sell Nike." Similarly, if you show someone an ad that says "Compare Insurance Rates Here" and ask them, "Do you think that you can get GEICO rates here?," their supposition is going to be, "Yeah, they compare insurance rates." That may be confusion, but it's not

confusion that stems from the trademark use. So in order to figure out what the source is of that confusion, a proper survey needs to do controls. It needs to do

tests that isolate one source from the other to show where the confusion comes from. And Dr. Ford tried to do a control, but he got it very wrong. What he did was he eliminated one source of actionable

confusion, the word "GEICO" in the ads, and he eliminated the context. He got rid of the ads that say "Compare Insurance Rates

Here," and he kept the second arguably actionable source of confusion, "GEICO" as the search term, and his results were that when people entered "GEICO" as a search term but the word is not in the ad and the ad does not offer insurance comparisons, the confusion rate is zero. So as to the use of "GEICO" as a keyword, their only evidence on confusion is that it is zero. conclusions you can draw from that: There are only two

Either A, that there's a

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 10 of 27 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

valid control and he tested it correctly, in which case it proves our case; or it's an invalid control, in which case they have no evidence to present at trial on confusion at all, because that is their only evidence of confusion. We, we did our own survey, and we set up a control that did it correctly. What we did was we showed people a search page

with "GEICO" as the search term and ads that offered comparison insurance rates, "Compare Insurance Here," at the actual ads that they complained of in their complaint, and we tested confusion on that. We then did a control sell, exactly the same stimulus. They searched for "GEICO," they saw the same insurance ads, and we asked them, "Do you think you can get Allstate insurance here?" And even though there was no search for Allstate, there was no mention of "Allstate" anywhere on the search page, the level of confusion was higher for Allstate than it was for GEICO, which establishes that the -- to the extent people think they can get GEICO quotes from a site that says "Compare Insurance Here," it's not because of what they searched for; it's because they think they can get insurance quotes there because they can, and that's what it says, and they assume -- in GEICO's case incorrectly -- that those quotes include GEICO. But that's a belief the user brings to the situation from their own experience and has nothing to do with any alleged trademark use even if you believe that triggering an ad that

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 11 of 27 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

doesn't use the trademark is, in fact, a trademark use. As -- because Google only allows people to trigger ads off trademarks and does not today allow them to put "GEICO" in the text of the ads, that's the only issue before this Court as to injunctive relief. THE COURT: All right. You-all want to respond to that?

Who's going to be -- let me have that issue responded to. MR. PAGE: THE COURT: MR. OSSOLA: wrong. Okay. Mr. Ossola? Your Honor, the -- Mr. Page has our theory

This is not a case where we're singling out any one factor This is a case -- and the survey

and looking at it in isolation.

reflects it -- where what is being looked at is the presentation of the search results of the sponsored listings on the page in conjunction with the organic listings; the context in which that occurs, which, of course, is the subject of the survey in terms of what was presented to users; the fact that -- which will be presented at trial, auto insurance quotes from GEICO are not available through any of these sites, GEICO insurance quotes are only available through GEICO; the fact that Google's own policy in the history of the evolution of their trademark policy shows that they started out being concerned about not about misleading consumers and did not allow what they're now allowing in terms of allowing advertisers to bid on competitors' trademarks and admits in the spring and during the course of this litigation, they've

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 12 of 27 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

modified their policy again, and now they don't allow the use of trademarks in the text, but the evidence will show that notwithstanding that change, it has continued. During the depositions of Google witnesses in this case, it still continued. We did live feeds at the time, and we still

saw "Google" appearing in the text of ads. So the effectiveness of the change in terms of injunctive relief, well, that's a matter for the Court to address after hearing the evidence. The appropriateness of injunctive

relief with respect to use of the mark in the text is part of this case. It's essentially been admitted by Google, and its own

internal user study, which is alluded to in our opposition which will be presented at trial, showed overwhelming levels of confusion associated with this, and there are hundreds of instances of past infringement that will be essentially admitted at trial as a result of Google's own admissions through its user survey and its change in policy. All of that will be considered by the Court in, first, a determination of liability, and second, whether or not injunctive relief is appropriate. Google says, "We've changed the policy. time to clean all of it up. It's taken some

No injunction is needed."

What you'll hear from us is there have continued to be problems throughout the litigation of this case. It's changed its

policies back and forth a number of times, and injunctive relief

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 13 of 27 13

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

is indeed important. But let me step back to, to the theories of liability. What the Court will consider at trial is likelihood of confusion based on the entire setting, not just the use of trademarks in the text, but also, of course, the use of trademarks of competitors as keywords, which Google originally prohibited and then allowed and we will submit and show for financial reasons only. And what was surveyed and studied by Dr. Ford showed that whether the trademark is in the -- whether the GEICO mark is in the text or whether the GEICO mark is just the trigger for the auto quote sponsored listings, he found very high levels of confusion, 60 to 65 percent levels of confusion, which he said and which the defendant's survey expert acknowledged are high levels of confusion if that is sustained by the Court. Google is simply asking you to interpret the meaning of our survey, which is bolstered not only by Dr. Ford but also by another experienced expert, Ivan Ross, they're asking you to interpret our survey the way they think it should be interpreted, and not only is that inappropriate at the summary judgment level; you have a classic disagreement among experts about the meaning and significance of the survey results. Those survey results, as counsel alluded to, involve different controls. The Court is going to have to decide which that

are appropriate and which were not, but I will say this:

defendant's findings of confusion before they started cutting back

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 14 of 27 14

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

with the controls were also strong, over 40 percent affiliation confusion, and then they use an elaborate series of controls to get that number down to below 5 percent. inappropriate. We think that's

We will attempt to demonstrate that at trial.

But the point is you're being asked to interpret the meaning of these controls at summary judgment stage, and I think the law in the Fourth Circuit on questions such as likelihood of confusion is pretty clear that that is generally not done and can only be done with great care. And I think here there's another reason why the Court should avoid getting into an interpretation of the meaning of our survey and, of course, crediting Google's survey on summary judgment without hearing from the experts, and that is, their attack is predicated on what our control proves, and that is -fundamentally misapprehends the purpose of a control. Our control and their control doesn't prove anything in and of itself. It doesn't test for confusion. It is simply there

to filter out the background noise that should be filtered out in assessing what are the reasons why we got these results 60-65 percent, close to 70 percent in some findings of confusion as to the source of the sponsored listings and whether they were affiliated with GEICO. And so the, the Nike control which was used by Dr. Ford was simply used to screen out background noise. You cannot take

that control and say it proves lack of likelihood of confusion.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 15 of 27 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

That is not an appropriate use of a control, and I think all the survey experts will, will confirm that at trial. So I would submit that this is a situation where you will hear from two experts, a primary survey expert and a rebuttal survey expert, from GEICO at trial; the findings of the GEICO survey will be presented; the findings of the Google survey will be presented; we will offer a rebuttal witness which will -- who will attempt to demonstrate to Your Honor that the Google survey was an elaborate attempt after finding high levels of confusion to cut them back based on controls that were not proper; but at the end of the day, the Court will have to assess who's right and what methodology was correct against the backdrop of the other evidence in the case, including Google's conduct, including the changes in its trademark policy and what they mean, and including its own internal studies which suggested that they knew that there was a problem with, with selling competitors' trademarks to other companies, which is why originally they didn't do it, and then there were reasons why they began doing it, and then because of the overwhelming levels of confusion that they found in their own internal study, they changed their policy to prohibit the use of trademarks in text, and that has not been effective. THE COURT: MR. PAGE: All right. Your Honor, counsel just said a couple of

things that are very telling, the first of which is that their theory isn't that confusion comes from one source or another but

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 16 of 27 16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

that it's the whole package, the whole gestalt, and that's what their expert said: "I said I wanted to test it all together and

then none of it together." You can't do that because you're counting the noise as well as the actionable confusion. have a control. They would love to be able to count every human being who thinks that when they look at a site that says "Compare Insurance Rates Here," that they can probably get GEICO insurance there, but that's not actionable, just like you can't sue the store that says "Millions of -- Hundreds of Name Brands Available" because consumers think they have a certain brand. take the noise out. The other telling thing he said was that the point of a control is to filter out the noise. The way you do that is you You need to It's why every survey has to

measure the overall confusion from both the actionable -allegedly actionable conduct and the noise, and you get a number, and then you do a control where you remove the actionable parts and you leave in what you're calling the noise and you measure that, and you subtract the noise from the overall confusion. They did that, but what they called the noise was "GEICO" as a search term. They left that in, and they measured

it, and they took out the context of the ads and the use, and when they measured what they were calling noise, the result was zero. THE COURT: All right.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 17 of 27 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 issues?

MR. PAGE:

If I could move on to a couple of other

THE COURT: this is a great case. before. MR. PAGE: THE COURT: your argument.

Actually, actually, as much as I enjoy -I've said that probably too many times

Don't -Wait. I think there's a lot of merit to

I think the issue about surveys is always

problematic and whether the proper technique has been used and whether the proper controls has been used, very interesting issues. I'm concerned, frankly, about the procedural posture of the case at this point, and I'm cognizant of the fact that we're set as a bench trial, not as a jury trial, so the use of resources is somewhat different. Resolving this case in the context of a trial, where, as you know, the burdens are different, the presumptions are different, etc., is a better forum in which to address these issues, so I am going to deny the motion for summary judgment not because I necessarily find it is without merit but because I think it is far better to resolve these issues in the context of a complete record so that that particular issue is out of any appellate proceedings that go down the road. If I'm reversed, so be it on the merits but not on a procedural error of resolving the case prematurely. All right?

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 18 of 27 18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

MR. PAGE:

Okay.

Your Honor, briefly, if I could

address a different issue, which I think -- on which you can very easily resolve this case on summary judgment immediately? realize I'm fighting uphill -THE COURT: right, let me hear. MR. PAGE: damages here at all. Which is very simply, there is no claim of The traffic that flows to GEICO and the I'm looking forward to hearing that. All I

money they make off the traffic that flows to GEICO from Google is nothing that they are entitled to in the first place. Google built a search engine that daily sends millions of -- or millions of dollars in the course of a year to GEICO for nothing. They are under no obligation to do so and -THE COURT: Wait. Isn't the reality of this case not a That's the sense that I have of

damage claim but the injunction? it. MR. PAGE:

That is correct, Your Honor, but if there is There's no harm to remedy.

no damage, no injunction will lie.

What they are complaining about is that a small subset, even if you grant all of their, their theories that an ad that says "Compare Insurance Rates" is somehow fraudulent because they didn't let anybody compare their rates and they claim no one's allowed to and that's misleading, even though it doesn't say "GEICO" to begin with, even if you grant all of that, they're not harmed. It is just a dimunition in a free flow of business to

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 19 of 27 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

them to begin with. These aren't two different issues. This isn't -- I

grant that if I run into your car, I can't claim that I'm off the hook because I painted your house for free. Clearly, that's not a

defense, but you can't sue me for failing to finish the trim if I decide to paint your house for free, and that's what they're doing here. They are claiming that the, the free flow of business that goes to them from Google is being diminished a little bit, and that's not a damage claim. good to begin with. judgment today. THE COURT: MR. OSSOLA: All right, let me hear a response to that. Your Honor, you're being asked to accept an This case is about That's just a dimunition in a free

And on that basis, you can grant summary

absolutely extraordinary proposition. sponsored listings.

Sponsored listings is the primary, almost the It's an advertising program

entire source of revenue for Google.

where they're paid on each time a consumer clicks on an advertisement. That's what this case is about, and there is a substantial damages component to it because every time, as we will show, every time a consumer clicks on one of these sponsored listings, believing as we will demonstrate that there's some connection that they'll either get a GEICO quote or this is GEICO or this is sponsored by GEICO, we have lost a click, and as we

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 20 of 27 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

will show you, a very substantial portion of GEICO's business now comes through the Internet, and our damages expert will present to you calculations that will show how many clicks have been lost. Clicks through to the GEICO site is what I'm talking about. We will then show you what the historic percentage of conversion rates is of the number -- and it's low; it's low -- but the number of clicks multiplied by that conversion rate and taking into account the value of an insurance policy over time is very high, and you will hear that this has, that this has cost GEICO a substantial amount of lost profits. What you are being asked to accept is that because Google includes through its algorithm "GEICO" in the organic listings, that free benefit outweighs any damages associated with trademark infringement. So if we violate the law, it's okay. You

should disregard those damages because we're including you in a benefit that we're offering to others through our organic listings. That's not the way the law works, and in fact, you know, this is such an extraordinary proposition, we will ask you before the trial to strike it because a damages theory that says that you can, that you can commit infringement and cause damage should be disregarded and, in fact, overwhelmed because there's some charitable contribution being made in another context that is not at issue in the case is simply not defensible. THE COURT: All right. It's an interesting argument.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 21 of 27 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

MR. PAGE: THE COURT:

Your Honor, I think -It got my attention. I heard it. I'll be

listening for it during the course of the trial.

I'll save I

Mr. Ossola the time, don't try to strike it from the trial.

mean, we're going to have a trial so that all the issues can be litigated and whatever decision I render is based on a complete record. All right? Now, we're set for this trial to begin on Monday, December 13. Am I correct in assuming that you're going to go -I mean, Overture has sort

the case will go against Google first? of been the quiet player in all of this.

I would think the lawyers are working well together. Talk among yourselves and make sure that you have a good, clear order of proof. I suspect -- because there are definitely from my

understanding of the records, definitely differences between Overture -- factual differences between Overture and Google, the way the pages look, the way the searches are done. Some of the

legal issues may be the same as to both defendants, but I want to make sure that, you know, the evidence is properly demarcated there. Obviously, I've been better educated about the Google

aspect of this case than I have the Overture aspect. The last question, I think I mentioned this to you last time or at the last pretrial, is there any work being done behind the scenes to try to resolve this as a business resolution? MR. OSSOLA: Your Honor, there are settlement

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 22 of 27 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

discussions ongoing with Overture, and those are substantial. Mr. Page and I continue to talk about, about settlement with Google. I don't know if he'll disagree, but I'm not sure our

conversations have risen to the level of settlement negotiations yet, but we have had conversations, and we will, I think, continue to have conversations. So we're cognizant of the complexity of the case. We're

cognizant of the importance of the case to both of these parties. So those conversations are continuing. MR. PAGE: discussions. Yeah. I would agree we, we have had I would say

We will continue to have discussions.

that there has been some movement. direction, however. THE COURT:

I won't characterize the

I would assume that technologically, it's

almost possible, if not definitely possible, what if Google decided -- well, GEICO indicates in their papers that Google has made a decision that it does not allow the Google name to be bid upon. Does the Google name appear, though, within your searches? In other words, if somebody just wants to learn about Google and they go on -- I haven't actually bothered to do this exercise at home, but I assume I can pick up Google as an informational purposes like any other keyword in a general search? MR. PAGE: Absolutely, Your Honor. And it is, it is not

true that we prevent people from bidding on the Google keyword. We did at one point; we do not do that and haven't for quite some

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 23 of 27 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

time.

It's a rare event. I think most advertisers are -- find it unlikely that

people will go to the Google site to search for the Google site or for information about Google when they're already there, so it's a rare example. We do have a policy where we discourage advertisers from describing things as Google specials or, you know -- and try to give an impression that they are somehow different from the rest of the advertisers. So we do not let them use "Google" in the

text, just like we don't let them use other trademarks in the text. I do want to clarify something Mr. Ossola said. not repeatedly changed our policy. We have

Our policy was until June of

this year that when a trademark holder requested it, we would block their trademark both as a keyword and in the text. In June, that policy changed, and we now continue to block trademarks in the text, but we do not block the use of trademarks as keywords. THE COURT: MR. PAGE: anything. That was the one change in policy.

But you had or have the ability to do that? We have the ability to, to block just about

It's challenging to execute because we have hundreds of

thousands of advertisers doing millions of ads, each of which they input themselves. We have a whole lot of bells and whistles where

you can have automatic insertion of whatever the user searched for. You can have multiple different ads that rotate. You can

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 24 of 27 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

have ads that get paused and reactivated. So when we made the change in June, it took us some time to go back and pick up -- when we were blocking ads as keywords as well, you didn't have an issue with keyword insertion in the text because it wasn't running to begin with. When we changed that, we started having ads that would pop "GEICO," for example, in on automatic insertion. time to go back manually and get rid of those. It took some

We then had to

deal with ads that when we did that search, had been paused so they weren't there, and reactivated. So it admittedly has taken us some time to implement the process of enforcing our procedures, but that doesn't give rise to derivative liability, and we continue to improve the process, and I'm sure Mr. Ossola will tell you that although they do occasionally find instances, they are very rare at this point, and they disappear very quickly. THE COURT: MR. OSSOLA: Okay. Your Honor, can I just say one other thing I think there are two aspects of this.

about what Mr. Page said?

As a matter of policy, almost philosophy, Google has allowed anybody to bid on the trademark -- on any keyword, including a trademark of a competitor. That's a business decision they made.

They didn't used to do that, that's what I meant by change, but that's the decision they've, they've made. THE COURT: All right.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 25 of 27 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 case?

MR. OSSOLA:

So they've chosen to allow that, but on the AdWords is

technological side, you know, this is their program. their program.

It has elaborate controls and policies and

relevance criteria. They can certainly, of course, as a matter of policy and as a matter of technology take whatever steps might be necessary to fix problems prospectively or retrospectively. It is a

complicated exercise to do it retrospectively, which is one of the reasons we're here, because we do not want to continue doing this, which is finding repeated problems, policing the mark. But there are two aspects of this, and they're both complicated. THE COURT: Who is the magistrate judge assigned to this

MR. OSSOLA: THE COURT:

Judge Buchanan, Your Honor. All right. Well, again, if you think that

getting with a judge who can keep you sort of focused would help, make sure you call her well in advance of the trial date. we'll see you here. Now, I'm going to have this trial upstairs in the 7th floor courtroom, which will give you access to a much better technology setup. I would assume during the course of this trial If not,

we will have live online presentation of some evidence, and you have the ability to get into the -- work with Ms. Travers on that, but I think we can probably fulfill any of the technology needs up

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 26 of 27 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

there.

Okay? MR. OSSOLA: MR. PAGE: We'll do that. That was one question I did want to address

to the Court. THE COURT: MR. PAGE: Yes. If the Court would, would like, we could

arrange so that we could have live Internet access so that witnesses can say, well, for example, if you do this. THE COURT: MR. PAGE: That -I think that would be helpful. We could

certainly do it with canned examples as well but -THE COURT: All right. It's going to be an analog You can't cut into the DCN

connection because of firewall issues. itself. MR. PAGE: THE COURT: people. Okay.

Ms. Travers and Mr. Bachman are our tech

You can work that out with them. But yes, I had anticipated this case will need that kind

of facility, and we have that.

So upstairs, right above this

courtroom, Courtroom 7, and I'm sure we'll be in contact with you informally by telephone to set up other logistics that need to be addressed. I don't expect any more pretrial motions in this case, right? MR. PAGE: Neither do I, Your Honor.

Case 5:03-cv-05340-JF

Document 47-4

Filed 12/22/2004

Page 27 of 27 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

THE COURT: MR. OSSOLA:

All right, very good. Your Honor, your comment about damages, the

only reason I made it is to see what your reaction is, and given your reaction, we'll present our arguments at trial. THE COURT: court until 11:30. MR. PAGE: MR. OSSOLA: Thank you, Your Honor. Thank you, Your Honor. (Which were all the proceedings had at this time.) Very good, all right. We're going to recess

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