- Jared Spool, User Interface Engineering
- Dan Thies, SEO Research Labs
- Cheryl Ho, LinkShare
- Jorgen Wouters, Consumer Reports WebWatch
- Dana Todd, SEMPO
- Danny Sullivan, Search Engine Watch
- Matt Cutts, Google
- Mara Hannula, Marriott International
- Mike Moran, IBM
- Rob Smahl, PriceGrabber
- Michael Yang, Become.com
- Brian Bowman, Dogpile
- Rob Smahl, PriceGrabber
- Beau Brendler, Consumer Reports WebWatch
- Jessie Stricchiola, SEMPO
- Laura Lippay, Consultant
- Ken Shuman, Bite Communications
- Barbara Coll, WebMaMa
- Bill Kelm, SEO-SEM
- Niall Kennedy, Technorati
Note: This is an edited transcript of the proceedings.
Beau Brendler: To tell you about our esteemed panelists today – and, as you can see, there are quite a large number of them up there. They don’t have protection from tables or podia, or whatever the plural is, or anything; they’re just kind of sitting there, waiting.
I’d like to introduce today’s moderator, Jared Spool. Many of you know Jared already. He is a charter member of the Consumer WebWatch Advisory Board – stuck with us all this time – and is a thought leader in the field of Web site usability and design. At User Interface Engineering, a research firm that studies the frustration introduced by new technology – something I think we’re all quite familiar with – Jared’s official title is Founding Principal, which means he spends all of his time in the principal’s office, which he says brings him back to his childhood.
I didn’t write this stuff.
Jared is also a professor at the Tufts University Gordon Institute and the chair of the annual User Interface Conference. He’s going to walk us through the important issues surrounding consumer trust and credibility with search engines to explore who really is paying for our searches. Jared.
Jared Spool: Thank you. We have the world’s largest collection of microphones, so I’m sure it’s going to be exciting all the way through.
Today, we’re going to talk about what we started this morning in talking about the issues surrounding search and particularly this idea of credibility for consumers and really try and understand all the different parts of the elephant here. There’s a lot of different things going on. There’s a lot of different issues that are involved. What are all the different pieces? From a variety of perspectives, we’ve put together a panel here that really sort of represents almost every possible angle you could look at this issue from. We’ve got people from the search engines themselves, we’ve got people who sell advertising, people who buy advertising, people who are analyzing the market space. So we’re going to look at all of this.
Real quick, just so you know who we have here. Starting at your left, going across, we have Dan Thies, who is the President of SEO Research Labs. Next to him, we have Cheryl Ho, who heads up Search Marketing Group for LinkShare. We have Mike Moran, who’s a distinguished engineer from IBM. Brian Bowman – do I have that right? No, I missed somebody. Matt – no, okay, I’m sorry. Brian is VP of Marketing and Product Management at Infospace. Dana Todd is President of the Search Engine Marketing Professionals Organization. Rob Smahl is VP of Marketing and Communications for PriceGrabber.com. Michael Yang is the founder and CEO of Become.com, and if he has a very large smile on his face, it’s because he just $7.2 million of venture. Something I’ve never managed to do. It more than doubles what I’ve been able to do. Matt Cutts, from Google, who also has a smile on his face, for obvious reasons. Danny Sullivan, founder and editor of SearchEngineWatch.com. Mara Hannula, from Marriott International, she is the Senior Director of E-Commerce Marketing. And Jørgen Wouters, who is with Consumer WebWatch and you heard from him this morning.
So that’s our panel and we’re going to start with paid placement, because it seems to come up first on the search engines. And, in particular, we’re going to start with Dan Thies, who you may not know, is the author of two books on search engine marketing. His company, SEO Research Labs, provides consulting and training services to search engine marketing firms.
So in the issue of paid placement, Dan, do we need to take extra care to protect consumers, or is basically the status quo acceptable?
Dan Thies: I think that, as we saw this morning, search engines like Yahoo! have actually backed away from disclosure that may have been in place a year ago. But overall, I don’t think that the disclosure of the paid advertising, paid listings, is nearly as much of a problem as other areas of what appears on the search results.
Jared Spool: So, all of this talk this morning was about people not being able to tell, people being confused, people being upset when they found it – are you thinking that’s at this point overblown? Is that–?
Dan Thies: To an extent. I mean, obviously, it’s been tested in front of live humans and some of them didn’t know what “Sponsored links” means or something like that, so maybe the wording can be changed. I just don’t think that it’s the primary issue that you’d look at in terms of what’s going on in search results and how it affects consumers.
Jared Spool: So if that’s the issue, if it’s not really such a big deal, should we just leave things as they are?
Dan Thies: I think that, compared to other issues going on in the space, you’ve got bigger fish to fry. I think that all of the search engines could do better in disclosing a lot of things, including making a better effort to disclose what are the paid advertisements. I personally like the wording “paid advertisement” as an adequate description, as opposed to “sponsored listings.”
Jared Spool: Okay. We’re going to come back to some of these other things in a minute, but before that, first I want to say, if you – we want to make this as interactive as possible, so if you have any questions, just put up your hand and Tracy – oh, it starts already – Tracy will bring a microphone to you.
Speaker: What are the [unintelligible]?
Jared Spool: Well, we’re going to talk about those. We have a nice, small piece. We’re going to talk about some of these other issues in just a little bit, but I want to stay on paid placement just for a minute.
Jorgen Wouters, Consumer Reports WebWatch. He’s a journalist that’s been working with them, and he told me before this that he can’t wait for this week to end, because he just put out a report, launched a Web site and prepared an excellent presentation this morning. And next week, what? You’re going to be at a cabin in Beaverkill River, is that what you said?
Jørgen Wouters: Drinking and reading. In that order.
Jared Spool: Excellent. Reading is always better after drinking, I’ve found. The books just take on that extra piece of life.
But you, this morning, went to great lengths about these issues. Your thinking, obviously, is that the status quo isn’t right. But is there a possibility that we’re just looking for problems that aren’t really there?
Jørgen Wouters: I think the problems are there. I do think it is a problem. I think one thing we all need to bear in mind is that we probably all spend more time on this than we probably should, and we’re in so deep that I think we often forget that the majority of people out there are not nearly as aware of this stuff as we are.
As I discussed this morning, there was some interesting research WebWatch came up with in 2003. Full disclosure: I did not write this report, so I’m going to be reading from some of the findings; I am somewhat familiar with it. This was this anthropological study that Beau referred to where – in this study, we took some Web-savvy consumers, and we put them through their paces doing Internet searches. And among the findings were that most participants had little understanding of how search engines retrieve Web pages or how they rank or prioritize links on a results page.
The majority of them never clicked beyond the first page. They trusted search engines to present only the best or most accurate, unbiased results on the first page. And as I said this morning, that means that two in five links, about 41% of the links selected by the participants in this study were paid results. They had no idea. Yes?
Speaker: Question. Just to get a better understanding of the data set, do you know what percentage of all online users, assuming this is U.S.-based, have gone online in the past six months as their initial entry?
Jared Spool: So the question was, do we know what percentage of users who go online have done it in the last six months in the U.S.? I’m not sure I have that statistic.
Jørgen Wouters: That’s a good question, I’m not sure I have the–
Speaker: The reason I ask–
Jared Spool: Oh, as the first time they go in.
Speaker: Yeah. The reason I ask is only because I think most of us here who have been in the Internet in the industry for well over at least 7 or 8 years, to define an Internet-savvy user as someone who entered in the past six months or only has six months’ experience, that’s not our definition. So that’s one issue.
The second issue is, what percentage of U.S. users, of the 90 million, does that six-month entry group represent? So I would like to see that data. Because if they only represent, you know, seven percent of all users, the concerns that they might have might not be representative of this overall. I’m not saying that the concerns that we’re raising here are not good concerns, because I’ve actually — just in the last session — realized a few things that, because I’m so micro-focused in the vacuum of my world, I hadn’t really thought about to the extent that you guys are bringing them up. But I think for this data set, I’d like to have more info.
Jørgen Wouters: Right, and, again, I – this is not a report that I generated, so I can’t rattle that off for you, but it is on the Web site. The reports are there, and the methodology and the whole nine yards. So I promise you it’s there. If it’s not up here.
Again, as I pointed out this morning, the bottom line is that when, during this study, a lot of the users had some fairly negative and emotional reactions. What I’m just trying to underscore is, what we take for granted, a lot of people have no idea is going on. And this is not about saying advertising is bad. It’s all just about being upfront and being clear and conspicuous and doing the things that the FTC pointed out. This was not invented after the Internet, these are long-standing principles. And I think it is an issue and I think it’s something we need to focus on.
Jared Spool: Okay. So, Dan, I want to come back to you and ask: Let’s say we go and we make sponsored ads more prominent — does that hurt advertisers?
Dan Thies: It in my opinion doesn’t hurt advertisers. I think one of the things that a lot of advertisers aren’t aware of is exactly how their ads are displayed and, in many cases, depending on where you buy them, how what you’re buying as a search advertisement isn’t even being displayed on search results, but it might be a pop-up window or some other format. So as an advertiser myself, I would much prefer that the people who are clicking on my ads are aware that it’s an advertisement, rather than that it be sort of sneaky, concealed, sort of hidden in-between other things. Or if it’s the first thing on the page, and people just click on the first thing on the page all the time.
We — one of the keywords that we used to bid on was “keywords.” And we had to put a negative match in for “AOL” and “type,” because you’d get people who would come to our site having typed in “type AOL keywords here.” That’s someone who logged onto AOL, hit the search button and clicked on the first link. They weren’t searching for that, I hope. And certainly they weren’t interested in what we had, and I would think that the amount of money we spent on that before we clued in to where all this was coming from would have been reduced quite a bit if it had said “Paid Advertisement” in big red letters above that, as opposed to it looking like a search result.
Jared Spool: Okay, but that’s sort of an unusual–
Dan Thies: It is, but still, there’s a whole element of strategy in the pay-per-click side of search engine marketing, where you try to avoid being the number-one listing, because it gets clicked just because people don’t know what it is.
Jared Spool: Okay. Now, Cheryl — this is Cheryl Ho; she comes from LinkShare. She’s been with LinkShare for nine years and she’s currently responsible for building successful page search campaigns that are integrated with LinkShare’s client affiliate marketing programs.
Do your clients get hurt if an ad is denoted as an ad?
Cheryl Ho: Do they get hurt if an ad–?
Jared Spool: I mean, if it’s prominent that that’s an ad, does that hurt your client’s ability to get those clicks or to get that business, do you think?
Cheryl Ho: Yeah, I think that’s interesting, because you say advertisement [unintelligible] and I think there’s a negative connotation to the word “advertisement.” And I think a lot of the angry frustration from the study that was done is because it was flat-out said these are ads, do you know these are advertisements? And of course I think the natural consumer behavior would be anger, distrust, whatnot.
And I would actually step back and think about what the definition of an advertisement is. Because it’s not like if I search on “Palm Pilot,” an ad for Ford trucks can appear. I mean, that’s a true ad. I mean, these are listings that are relevant to the users search. So when we talk about what’s relevant to a consumer, I think that’s what’s most important.
And I actually want to talk about some other statistics. We heard a lot of statistics this morning from the research that was done and there was an article written by Kevin Ryan, which I forwarded to you last night, that I thought was very interesting that shared some statistics that I just want to share with the group here. It was a study done — I think it was a survey on I think it was a thousand consumers. And they were asked what’s most important to you in search results. Eighty percent said it was the relevancy of the results. And 19% said it was the quality of the Web site and the clarify of the titles and description. And there was 1% that said it matters to me if it’s an ad, if it’s paid or if it’s free.
Same study asked users what is the definition of “sponsor.” And 90% responded that those are Web sites that paid to be there. So 90% understood that those listings were paid. That advertisers paid to be there.
And then there was another question about, what do you feel about paid results? And 60% said that I don’t care what the paid results show, I care about quality of the content. So when we talk about users and these ads, what’s most important? What’s most important to the consumer, it’s the quality of the content on the site. And that’s where I think that Dan was referring to, in sort of bigger problems.
We heard about the health information online. I think that’s information that has bigger ramifications to the industry, I think, than defining this is an ad, this is a free listing.
Jared Spool: Okay. There a study that came out this morning which – or, not this morning, earlier this week — from Penn State, where researchers took links and showed them to users. Sometimes showing those links in the sponsored ad section on the right, sometimes, showing them in the organic wing section on the left. But the same exact links with the same exact descriptions, they would just move their position, just to see if users were more likely to pick them in one or the other. And they found that, in fact, users were far more likely to pick the organic ads, the organic content than to pick the stuff in the sponsored section. They also looked at the top sponsorship slot. And when they moved things around, they found people shied away from advertising. They also found that people said, told them that they trusted those links less.
So, is part of this move to blend the paid advertisements, the paid placement into the organic stuff as part of what we’re seeing there basically an acknowledgement that users are not that interested in this and, therefore, the advertisers who pay for the searches are not getting their value, just because of how these are presented? Is there a chance that that’s what we’re seeing?
Cheryl Ho: I think from the examples that we’ve seen, in my view, they’re clearly separated. Moving from a red color to a gray color, it’s still identified as sponsored links. And I think the results of the study show that people know that these are paid for and they’re going to – and it’s true, most people, talk to people about where do they click, and a lot of people say, “I don’t even click on the paid links,” because they want to click on the organic results.
There are other views that, if I am actually looking to buy something, I’m going to click on the sponsored links. If I’m looking for information, I’m going to click on the organic listings. And I think that’s the sense that people have – sort of the consumer behavior that we see today.
So whether or not they’re differentiated or not, the fact that is that they are being differentiated.
Jared Spool: Okay. I want to turn right now to Dana Todd, who’s part of SEMPO, I’ll get that right eventually — we put the R in and now I have to take it out – SEMPO, which is the Search Engine Marketing Professionals Organization. The organization has a global membership of search engine marketers. Dana’s been working in this field since 1996. Dana, what’s your take on this?
Dana Todd: Well, I was going to ask – from what we’ve seen, actually, if you look at search engines like Yahoo! and AOL, who sort of are being accused of not being as obvious about their advertisement placement, there is a lower percentage of clicks going into the paid space. Google, because it separates out the ads so cleanly does get a lower click rate. So, from an advertiser perspective, yeah, your volume goes down. If ads are way over here – and it’s not even so much that they’re disclosed, it’s — you’re getting into eye-tracking behavior. Human eyes –and this goes way back to the old print days, why do you think newspaper columns are so skinny? It’s because your eyes just don’t go that far over; you’re lazy. So you stick right smack in the middle.
So it has less to do, in my mind, with a preferential – there may be some of that, of course, but largely, it’s because their eyes are like this on the screen.
Jared Spool: Okay. Dan?
Danny Sullivan: I just – a couple of comments. One, I don’t trust any of these statistics, because I used to be in political polling, and I could get 90% approval for killing bunny rabbits, or 90% against, depending on how I asked the question. And I assume that Consumer Reports asked the question differently than a company that sells advertising.
But the thing is that with the way that the ads are displayed, depending on how you do it, how prominently you disclose that it’s an advertisement, a more prominent disclosure will decrease probably the click volume, but increase the quality of the resulting traffic, as well. So it’s sort of offsetting. What all the search engines are trying to do is optimize their revenue, and what all the advertisers are trying to do is optimize their profit from it. So where exactly it fits perfectly, you’d have to study a lot of users, but I think more clear disclosures doesn’t hurt advertisers in direct proportion to the drop in volume.
Jared Spool: Okay. That makes sense. On the statistics thing, someone once told me that if you – someone I worked with who was a statistician said, well, you know, if you torture data long enough, it will confess to anything you want it to. Just always remembered that.
Why don’t you tell us who you are?
Jessie Stricchiola: Oh, Jessie Stricchiola. I am on the board of SEMPO and also run my company, Alchemist Media. So I had a quick question of the panel, actually, and I’d like to know by a raise of hands. Cheryl, you bring up a really interesting point that can help inform the discussion as it goes forward about the definition of what is advertising within search. And I hadn’t quite heard it put that way, but to the best of my understanding, it’s your opinion that defining paid or sponsored listings as ads is questionable at best, because the result that is served up, the resulting ad, is relevant to the user experience. It’s not like whatever analogy you gave, something unrelated to the content in a banner being delivered.
I find that definition problematic only because that assumes then that if someone opens a yellow pages book, because they searched on dog kennels and they found something, an ad for dog kennels, that that means that’s not an advertisement, just because it’s relevant to their search. I mean, really, at the most basic comparison.
So what I – I define it as advertising. I would like to know what the panel defines it, how many of you here define these ads as advertising, regardless of the fact that they’re content-targeted and relevant?
Cheryl Ho: I’ll make a quick comment on that. I don’t disagree that it’s not advertising, but there are so many forms of advertising. There are advertorials, then you hear voicevertising or dogvertising, I mean, this can be adversults, right? There’s some – it’s some hybrid. I’m not saying it’s not advertising at all. It is. I mean, what is the definition of advertising? It’s promoting a service or a product, and I would say that advertising shows up in organic results, as well. Be it paid or not.
Jessie Stricchiola: And I completely agree with you on that. I just want to know — because it’s obvious that the industry itself and external forces are pushing towards more disclosure and disclosure of what exactly is what we’re getting into — advertising? Paid advertisement? So I just wanted to get a sense of what the take on that term was.
Jared Spool: Okay. I think we’ll talk about more of these different aspects of advertising as the panel goes on. I want to turn now to Mara Hannula. You know, I had these all right before we started this. Who I mentioned is Senior Director of E-Commerce Marketing for Marriott International. She’s responsible for Marriott’s E-Commerce marketing efforts and in charge of Marriott’s search engine optimization program and their paid placement advertising.
So, is Marriott hurt if paid placement ads are clearly labeled as advertising?
Mara Hannula: I don’t – well, I guess to start, I’m speaking as a consumer and also as representative of Marriott, which is a company that’s founded on the principle of serving customers and we take that very seriously. So I think the comment about what’s more important is quality and relevance is really – I agree with that. But I don’t think quality and relevance go along with confusion. I think that’s the issue at stake here. So the issue goes beyond just disclosure. I’m really not as concerned with are the ads – I mean, I think it’s an issue, but a far less issue than some of these other things that we need to think about as a group.
And that has to do with how does the process work? What triggers an ad to appear? The issue of broad-match alone is so complicated. If I search on “Nike sneakers,” I’m very likely to get ads for Nike sneakers or for Puma sneakers or for stores that sell sneakers, but have nothing to do with Nike. Does the customer understand that? I think they may not understand that. That goes one level deeper than just are there ads on the search engines.
I think another critical issue that hasn’t really been discussed but needs some further thought around is the issue of who’s sponsoring these ads. So, for example, if a customer is searching on a specific brand, like Marriott, and they see Marriott’s ad in the title of an ad or in the text of an ad or in the url or in some combination thereof, it’s very reasonable to assume that Marriott sponsored that ad. But more often than not, Marriott did not sponsor that ad. And I have a number of examples, real examples, where you click over to a landing page and, again, the ad says Marriott, in this case it says, let’s say, 75% off room rates. You click over to a landing page, it says Marriott, it says 75% off. It looks like a Marriott site, it has a picture of our hotel on there. And you make a booking, which by the way, we have a look-no-further guarantee, so chances are you’re not going to get a lower price. Do you know enough to think, gee, this wasn’t Marriott who sponsored this ad? You may have made your booking and never know, thinking, gee, I got 75% off, that’s great. There’s an element of deception there.
And then I think the third issue, which sort of is related to that, around the issue of pricing, is what role, if any, do the search engines play in developing, or should they play in the development of policies and the enforcement of policies?
Jared Spool: So you mentioned you have examples. What’s an example?
Mara Hannula: Well, the one I just gave is a real example. Where Marriott–
Jared Spool: So we could bring it up here? So where should we go, Google?
Mara Hannula: Go to Google and you can type in Marriott – now, it was up this morning — Marriott Hotels.
Jared Spool: Okay.
Mara Hannula: And it said Marriott reservations in the url. I can’t see…
So I think the one that I was talking about was the third one right there. It says “Save up to” — it’s on the paid side — it says, “Save up to 75% Off.” And it says “Marriott” in the url. Online reservations.
So it would be reasonable, I think, as a customer to assume that that’s Marriott who sponsored that ad. If you searched through a hotel on this site, from when I’ve done it, you actually see the same rates that you do on Marriott’s Web site or our 800-number. And the customer has no way of knowing that that’s perhaps not Marriott sponsoring this and that’s confusing.
Jared Spool: Okay. Tell us who you are?
Laura Lippay: My name is Laura Lippay, I’ve been working as a consultant, but soon to work with CNET. And I completely agree with what you had to say. I think that this issue we’ve been talking about today has a lot to do with the relevance, which is what people are looking for. People are looking for something that’s relevant. They’re not looking at something, and the first thing on their mind is, is this paid or not paid? They’re saying, as they’re scanning through the results and saying, “Is this relevant?”
So in that respect, when you have something like this, like Google’s model, and you have your sponsored links on the side, everyone knows that that’s sponsored links – more people are going to use Google because they trust them more. So maybe you have your sponsored links clearly stated there and maybe advertisers don’t like that as much, but Google will have more people coming to them and so will people who model themselves after Google, so it seems.
So what I think maybe is part of the issue is exactly what you were talking about, is how do you get relevancy from these paid advertisements? How do you make sure that they’re relevant to the user, and that they’re not going to be mad about it when they click on it and it’s not Marriott. And Marriott’s not going to be mad about it also.
Mara Hannula: I just want to say one more thing on that. This is even more complicated, because it has a discount associated with it. And, as I understand it, Google has a policy today that says if your ad includes a price or discount or special offer, it must be clearly and accurately displayed one or two clicks from the landing page of that ad. Which is a very good policy to have, in fact.
But when you – and, again, we could do other searches — when you search on, I did a search on IBM and I found 68 or 50% off, I found it another of different times. You click over to a reputable site like Amazon and you don’t see that discount there. So in that case, I think the policy may be the right policy, but to the extent that the policy may not be enforced, the customer just doesn’t know to be aware. And it’s not that I think ads are bad and that customers don’t – if they’re relevant, I think they’ll work. I think it’s just a question of transparency and what role do we each want to play in it.
Jared Spool: So the question becomes, who’s looking out for this? Is this the search engine’s responsibility to police these ads and make sure these offers are fair? And if it’s not the search engine’s responsibility to do that – after all, it’s just a commercial business that has millions of customers – then whose responsibility is it to protect the consumer in that instance? And that I think is a big piece of the problem.
Mara Hannula: I think – sorry.
Jared Spool: That’s okay. What I want to do now is introduce Danny Sullivan, who’s the founder and editor of Search Engine Watch and he’s been covering these issues since 1996. Danny, what’s your take on this?
Danny Sullivan: Everything?
Jared Spool: Everything.
Danny Sullivan: Oh, it’s all good and bad. Let me touch on the thing that’s most fresh in my mind, which was the question of how do you know the relevancy? And I think it’s crucial to remember, and others have said this already, but one of the primary ways you determine whether or not the ad is going to be relevant is the fact that the advertiser’s paying for it. And if you click and you leave, they pay for it and they’ve made nothing off of that. And the more sophisticated the advertisers become, they’re tracking and they’re tracking everything that people are doing. So you click, you didn’t convert, you left. They know that ad didn’t convert them and it cost them a dollar, two dollars, three dollars. They’ve learned in short order that they can’t keep running those ads. So there’s sort of a natural evolution of relevancy to some of the ads.
But, having said that, you can also have ads that may be deceptive and still manage to convert someone. [unintelligible] because we said that we’d give a discount, we didn’t give it, but nobody really noticed and nobody really cared. And I don’t have a good answer to how you police all those sorts of things. There are so multitudes of issues that get involved with it.
Let’s take the pharmacy thing that was raised this morning. So the search engines aren’t carrying ads for unlicensed pharmacy, and that seemed to be primarily because the federal government suggested that they would take action against them if they didn’t do this. It didn’t seem to be motivated because the search engine just naturally said we don’t want to stop taking money from companies in Mexico and Canada. And the [unintelligible] I don’t think commented publicly on it, so we’re all assuming that that was the case. Or maybe it was the gambling ads that they stopped taking – well, one or the other. One of those two things.
But we don’t have the ads. We do have this where if you’re only a licensed pharmacy, you can do it. Now, go search for some – I mean, you can search for whatever you want, that should be a drug that you should only get from a licensed person and, if there are no ads showing up, that’ll be fine. You’re still going to get 10 free listings from companies in Mexico or Canada.
Search for “unlicensed prescriptions.” That’s what I’m – you know. Things you should only get – you shouldn’t be able to buy, go do a search, you can’t advertise for that. If you sell prescriptions to people without a – if you sell medicines without a prescription, you’re not supposed to be able to buy those ads on Google or Yahoo! any longer.
So the ads are gone, but the people who still have those are still listed in the regular results. Here we are. [Reviewing search results.]
Matt Cutts: No, it’s information.
Danny Sullivan: No, no, come on, Matt. Keep going, keep going.
Matt Cutts: It’s information! It’s information!
Danny Sullivan: Yeah, yeah, yeah, yeah.
Matt Cutts: [unintelligible] journals.org! And a [unintelligible].
Danny Sullivan: Keep going, I want to see the rest. Keep going, keep going.
Matt Cutts: Strawberry care! [unintelligible] MTA studies! [unintelligible]
Danny Sullivan: Okay, go back up. Search up, go [unintelligible].
Jared Spool: Maybe we should try Yahoo!
Danny Sullivan: But I can still go to USAMedicsonline and buy prescription drugs.
Speaker: There’s an ad.
Danny Sullivan: Yeah, but there’s an ad.
Speaker: Yeah, but we don’t know whether they–
Danny Sullivan: Kudos to Matt and I believe you, Matt. That was impressive. It would probably take me two or three minutes to come up with a query that would bring up some non-informational stuff. Because you can’t anticipate all the different things that are out there.
And he’s got a right to be proud if they’ve done it, because that’s what they’re trying to do is how do we skew these results or how do we tweak these results so that, if you’re doing the search, you’re getting the informational type of information. Which, by the way, when they do that, then all the other people on the Web start saying, oh, you must be evil, because you’re just trying to get us to buy ads now, you’ve forced us all out of the listings. Which is also something that happens.
But even if they could manage to do that on every single — well, they can’t do it on every single query, because you can still get up into other things where you just didn’t anticipate it or you didn’t get the algorithm twisted enough. Or at some point, somebody has managed to go over to some site, bought enough links from sites that were non-commercial in nature to convince you that they were informational and they weren’t. And this is an ever-changing game.
And I – part of me feels like some of the answer is just more human review of what goes on in the search results.
Jared Spool: Okay.
Danny Sullivan: We do have the human review of the ads; more human review of the regular listings as well may help.
Jared Spool: Okay. Say who you are.
Ken Shuman: Yeah, my name’s Ken Shuman and, Danny, you brought up a lot of great points and we’ve been hearing this “organic search results” to the left, people like the organic. But how organic is it if people understand how to get these fake links, spammers get in – what can be done about that?
Danny Sullivan: It’s not — there’s no level playing field. And that’s an important thing to understand. That it is not a level playing field. The unpaid results are not equal to being non-commercial. There are definitely commercial people in there. There are definitely people manipulating things. For good reasons and for bad reasons.
When you pick up your newspaper and you read New York Times, L.A. Times, San Francisco – I don’t care what it is, and you read that editorial column, that was not an unbiased article that you read. Absolutely not. Everybody’s got some kind of a balance. Or bias to them. It may have been a balanced article, it may have been a reporter’s best effort to make it balanced. But they did not have all the facts on hand. They may have gotten a press release from someone that triggered that they should write about this company.
Google’s a great example. Google’s got great products. Other people have existing products that have been out there beforehand; sometimes you never know about those, because Google gets the attention in that particular space sometimes. I’m not picking on them, but it’s the space that I write about.
And there are other people who may have those sorts of products that are great also, but you never heard about them, because they couldn’t get the time of day out of some reporter. That’s an off-world equivalent to what happens on-world.
And what you do about it, again, I don’t know. Diligently trying to, honestly trying to come up with good results, trying to get you the best answer, it’s still plagued by the fact that you never know exactly what someone’s going to search for. And even if you did the human review and you went through and you picked out the best results for depression, let’s say, what do you do when somebody types in, “I’m really depressed, what’s the best drug for that?” Did you human review all that? I mean, the queries go on.
You hear about this long tail, and I wrote once, search is a long tail, and it’s much – it’s been out there for ages, for anybody who knows search, we can’t anticipate all those sorts of things.
So I don’t know, for them, for the search engines, it’s an arms race, keep going. For the people who are doing it, perhaps over time, there’ll be more – I hesitate to say ethics, but perhaps there’ll be more of a sense of we ought to do something better or we ought to better understand that, if we aren’t the best site coming up, we shouldn’t be trying to promote ourselves for it. But that’s a difficult task, when you see things that you know aren’t the best site that are coming up, anyway. And then people feel like, well, what the heck.
Matt Cutts: So, to go back to your question, it’s certainly the case that people spend a lot of time trying to figure out how to show up in the organic search, because as we’ve seen, people are more likely to click on that, both savvy and novice users are more likely to trust stuff on that left-hand side. And the fact of the matter is, Google has a team of engineers that eats, breathes, sleeps, goes to bed thinking about how to make those left results as good as possible. So when somebody does a blind taste test unlicensed prescriptions and we bring up information and FDA and PDFs and things like that, that warms the cockles of my heart.
It’s certainly the case that people aren’t ever going to stop. I’ve seen people use hidden text on their Web page and they were talking about the power of prayer. Priests who were spamming for Catholicism. Usually, usually money is involved. And we do our very best to come up with robust, scalable ways – and every search engine does — to try to tackle those problems.
It’s relatively rare these days for you to type in your own name or some random search and get off-topic, porn kind of results. But that’s certainly also led to more of a proliferation where we saw on that video this morning, where the person said I want to type in a hotel’s home page and I want the hotel’s home page as a first result. To a first approximation, that’s kind of what quality at Google and other search engines is about. And he said, “I don’t want to wade through 50 people that will sell me the hotel.”
So I was kind of relieved, because Marriott was number one. Claremont, I tried last night, is number one. The Lafayette Park Hotel, where a lot of people stayed last night, is number one. So that’s good, we’re doing well on those searches.
But in any given month, about 50% of searches will be unique and so you can’t check every single search, you can’t check every single set of results. So you have to try to find ways that are scalable, that are robust. Whether it be – some people involve rating results by hand, some people lean more toward algorithms. A lot of different companies use all sorts of different things to try to keep that relevance high. But it’s not the sort of the thing that I think in the extremely near future you could say, “Okay, we’re done. Search results are perfect, let’s lock up the search engine and go home.
Jared Spool: I’m going now to Mike Moran, who, if he’s looking distinguished, it’s because he was just promoted as Distinguished Engineer at IBM. And he’s been in the search technology field for 20 years before the Web. Do you have the same problem that Mara has with Marriott, with people putting up ads pretending to give discounts off IBM mainframes, or whatever it is they sell? I mean, is that an issue that you guys are dealing with?
Mike Moran: We’ve dealt with it on occasion. Not to the degree that Mara does, but I think part of that is because the business we’re in, products and hotel reservations are different kind of businesses. There aren’t that many people that would have the confidence to plunk down thousands of dollars to someone unknown that they hope will ship it to them and hope they’ll actually have a warranty associated with the product. So I think that just the kind of nature of the business we’re in, it’s just less of a problem for us.
But we’ve had occasion where there have been things that we’re very concerned about showing up in Google or Yahoo! paid placement. We’ve been able to talk to them and get that worked out.
But just listening to the discussion here, when the first survey came out two years ago that said 40% of searchers know that there’s paid placement, I was actually happy that it was that high. And I didn’t look at this as some awful thing. I said this proves that in a very short amount of time, people are getting very savvy about that.
Suppose you did a survey that said, what percentage of viewers of the James Bond movie know why he’s not driving an Aston Martin? You think it’s 40% know that some car company paid them more money to use their car? I don’t think so, I don’t think it’s anywhere near that. If you went and interviewed those people coming out of the theater and you told them why he’s not driving an Aston Martin, would they be upset? And if you told them, look, Ian Fleming wrote Aston Martin in the book, I think they’d be upset.
Jared Spool: Was that a paid placement?
Mike Moran: In the book?
Jared Spool: In the book.
Mike Moran: Not in the book, but it was certainly a product placement in the movie. And so I think there’s all sorts of things that we deal with every day that are not disclosed and we have to become savvy about them.
But I think the thing that’s different with search — because there is a difference, it’s not the same as watching a movie — is the experience that we bring to search is not one of being entertained or one of even reading a magazine. But it’s more of going to the library.
So we’ve got a much different perception of what we think unbiased ought to be. Even though it’s free, and we know it’s free and we’re not paying for it, we kind of have that expectation of it’s a library, because we kind of feel like the library is free, too. Even though it’s kind of not free. Right? That’s government supported, it’s subsidized. And so unbiased in the library – and, by the way, I don’t think libraries are necessarily completely unbiased, either, there’s probably no such thing as being unbiased.
But maybe one of the things that can happen is that search engines can develop kind of a journalistic ethic. In journalism, you have paid advertising. It is the thing that keeps it going, even though it’s not free, or most magazines and newspapers aren’t free, even though some of them are. They’re very low-cost, they don’t come near paying for the price of the magazine, unless you’re Consumer Reports and you don’t accept advertising.
But most magazines don’t pay for themselves with subscriptions. But you expect that the articles are unbiased because there’s kind of a professional calling here that people take very seriously. And I think, when I hear Matt talk, that’s what he’s really talking about.
Maybe one of the problems is that we don’t have any words for that yet. Maybe we can’t call that journalistic ethics, maybe we need a different word for it. Maybe that’s a good word for it, I don’t know. But that’s the kind of thing that the search engines need to strive for, I think, because that’s a model that I think is scalable, it works in other places and you might be able to use that.
And so, yes, you’ll disclosure. Yes, you’ll do those kinds of things. And yes, it’s true, that when the PR person calls the report—
[End of Side]
–it’s not completely unbiased. Same is true that when someone hires a search engine consultant to help them put, to get their search rankings higher, it’s not completely unbiased.
Well, and that’s an interesting question, right? Because I think it was GM just pulled its advertising from major newspapers, like the L.A. Times because they were not getting favorable stories in those papers. So they went and they pulled their advertising dollar and decided they were no longer going to advertise with them anymore.
Mike Moran: And the place—and the company–
Jared Spool: Would the search engine companies today be able to withstand that type of reaction?
Mike Moran: Absolutely. Because guess which company, of all the ones you mentioned, is most likely to tell everybody that happened? The Los Angeles Times. Because they’re going to wear that as a badge of honor, to tell everybody how unbiased they are, how we are not besmirched by this kind of pressure, how this shows you why you should trust us. They’re the ones who are going to hold that up and have a clarion call over it. Do you think General Motors wants that out there? I don’t think so.
And so the same kind of thing can happen in the search business as well.
Jared Spool: Now, we’ve been talking a lot about paid placement. I want to talk about paid inclusion now. Paid inclusion is the act of paying a search engine to make sure your items get listed on the organic side of the listing, to make sure that it’s there.
We saw virtually no disclosure for paid inclusion. People, most people, are completely unaware that this is happening. Is this a problem?
Mike Moran: I think that – there’s definitely a perception problem with paid inclusion. And part of it comes from the business model.
For those of you who don’t know, paid placement is all those ads that you see where they say sponsored listings. Paid inclusion is something very different. Paid inclusion is when a company is paying money to a search engine so that their page will be included in the organic listings. And if you’re not – it sounds horrifying when you first say it, because, oh, nobody should be able to do that. Just understand, they’re not paying for a certain position, so they can’t pay to be number one in the organic listings. They can just pay to make sure that they’re in the index so that, if they match, they get there.
Now, why would a company do this? Because everybody else is in there for free. One of the reasons they do it is that many Web sites, especially large, corporate Web sites, sites that use a lot of expensive technology to display their Web pages, the search engines can’t easily spider their pages. They have a program that comes out and looks at all their pages and puts them in the search index. Some search engines are really good at getting that done, others aren’t.
Then there are some kinds of pages that the search engines really can’t or won’t spider and there’s all sorts of technical reasons for that that are really complicated, but they’re computers, so they don’t always do everything we want them to do. There are limits to them.
And so what happened as a result was companies said, “Okay, look, if we give you some money, will you put our pages in there, anyway?” And so a few years ago, many different search engines did this. The one that has never done it is Google. And what happens is that Google has always said, anytime you’re taking money, then you’re casting aspersions on whether you’re unbiased or not.
Now, other companies say, well, no, we’re really being unbiased. Now, Yahoo! is the one that mostly does paid inclusion today, and they actually not only take money to put the page in, but they take money every time the page is clicked. Which I think is a bad idea, because it gives them an incentive to muck with the rankings. Although I honestly don’t think they do. But it gives people that perception that it would be in their interest to do so, and I think it’s a bad idea that they do that.
But when you then go further and say should a company who has paid the search engine — because they can’t get into the index the free way — should their listing be identified as being paid in the organic listings? And I would say that, given the horror that people express about how awful paid listings are, I think you’re going to detrimentally affect that company for something that really isn’t their fault. Or it’s hard for them to control.
I’m not saying that you shouldn’t do it, I’m just saying that it isn’t a black and white thing. Disclosure isn’t something that’s pain-free. When you disclose things, you have to understand what’s the capacity of the people you’re disclosing it to to really understand and make a good judgment. One of the things that we’ve looked at here, and I think was brought up earlier, is that people talk about the organic listings as being money-free – they’re not. Would you like to know that Google gives higher rankings to pages that are on sites that are linked to by the Yahoo! directory that you pay $300 a year to be in? Is this a bad thing that they’re doing that? I don’t think so. But why do they do it? They do it because, to get into that Yahoo! directory, you have to really pass editorial muster. There’s a whole group of people that won’t put bad sites in there. So a link from there means you’re kind of a pretty good site, so they give you a higher ranking. That seems like a good idea. But companies that can’t afford the $300, they’re not in.
Jared Spool: I think it’s your turn, right?
Speaker: My turn?
Jared Spool: No, not quite. Her turn.
Speaker: You could fight.
Speaker: Two things I wonder — since Yahoo’s kind of taken over the paid inclusion thing — paid inclusion used to be a little bit different, now you have to do the thing through Yahoo! and pay-per-click. I kind of wonder if that’s working, because everyone that I’ve ever heard seem to think that’s a bad idea on both ends, except for Yahoo!
That’s one question. And the other thing is, actually, the way I see any kind of search engine listing, whether it’s organic or paid or paid inclusion or whatever it is, it’s advertising, no matter what. If you’re in marketing and you get your site on a Web site, whether you paid for it or whether you didn’t pay for it, you’re advertising yourself. Paid inclusion, as far as marking that as paid to people who are searching results, I don’t think there’s any reason to do that. It is a stigma to kind of say that somebody paid for that, even if it’s just to be included and not to be, not for your ranking. But I do think that, again, the way that Google is structured where you really tell that – and sometimes, a lot of times, actually, in searches that I do, I’ve found the paid results to be more relevant than – I can’t find what I’m looking for in my organic results.
So what about that? What if that’s more relevant sometimes than the organic results? And I’m not going to be mad that I’m going to have to click on a paid listing, because it’s what I’m looking for.
Jared Spool: I think that’s a good question.
Matt Cutts: Well, I can chime in with one point, which is I think Larry [Page] and Sergey [Brin], the founders of Google, sort of view the ads as potentially as useful as the search results. Because, unlike some other companies, where it’s just whoever wants to pay the most shows up number one. In Google advertising, it’s also how often an ad gets clicked on. And so typically, the more useful, the more interesting an ad is, the more it gets clicked on, the higher it will go. Even if the person isn’t willing to pay quite as much money.
So whenever we say “search results” at Google, we typically think about that left-hand side, the organic search results is what we call them. We don’t think about search results as being the advertisements on the right quite as much. But it’s certainly true is that what we treat it as is active search. And so if somebody types in “Nike sneakers” or whatever, we want to return some sort of shoe product. We don’t want to return an ad for a car or something that the users not at all interested in.
So in that sense, it is interesting, because Larry [Page] and Sergey [Brin] often view the ads as potentially just as useful as the search results and they want to optimize both sides. It’s just they want them also to be clearly marked, so people know what they’re clicking on.
Jared Spool: I’d like to introduce now Brian Bowman, who’s the Vice President of Marketing and Product Management for Infospace and responsible for the search and business directory parts of that. Are you thinking that inclusion – you don’t have inclusion on Infospace, right? Or you don’t know if you do, it depends on whether the feeds that you get have it. Infospace is meta-search, Dogpile –
Brian Bowman: If you can do me a favor, bring up Dogpile, just so I can visually show what it is.
So, if you guys are aware, we’re the only company that has contractual relationships with AOL, I’m sorry, Yahoo!, Google, Ask Jeeves and MSN. So we offer both paid ads from Yahoo! or Overture – sorry, Google or Overture. Type “discount prescriptions” as an example. And then you’ll see, we do something very different than what’s being discussed today. We blend our search results to build on what Matt was talking about.
Often paid ads are more relevant than paid ads or natural ads. It depends upon the intent of the consumer. So, if do me a favor, you can see that each ad on our page to the – we only have one in the list, is sponsored and it says it’s sponsored. But if you click on Google and Yahoo!, little yellow buttons, you’ll see – yeah, click on that, please. You’ll see what Google is returning specifically and what Yahoo! is returning, specifically. And what is yellow is what’s unique to that search engine.
Kind of the challenge we see from our vantage point of blending these search results is that only about 3% of page one search results are unique to a given search engine. So if you go to Missingpieces.dogpile.com, you can actually compare the page one results from Yahoo!, Google and Ask Jeeves.
And so when you type a query like “Civil War” or “Tylenol” or the “Michael Jackson trial,” who’s right? You’re going to get very different editorial views of search results. So there’s this common misperception in the industry that one particular search engine covers the Web and any search engine can keep it up-to-date in real time – it’s just not possible.
When you’re looking at esoteric queries like, “I would like a hotel room in Honolulu that offers daycare for children” – try and find that. It’s quite a challenge going through – I’m not talking about sports scores on ESPN or breaking news that are highly indexed and highly queried, but actual, real search queries.
And what’s important to understand is there’s an enormous difference between relevance and perceived relevance. Relevance is, when I’m alone at home and I’m at my computer and I’m trying to do something, whatever that is, and I’m actually going through and I’m working to find a solution. Often, paid ads, when relevant to me, are relevant. Versus to having a group of people sit down and say, Okay, go find out if there’s paid ads on the site or not actually performing those tasks. That’s the difference between perceived relevance and relevance. People may say it’s not relevant, because they’re not actually trying to perform the actual function.
But what’s fascinating to me is just how unique all the major search engines are. And if you look at just from a pure advertiser perspective, because we broke out both algorithmic ads and paid ads, it’s only about 2.5-3% of paid ads overlap between the engines. So, if you’re a marketer, there’s tremendous opportunities across the three engines, because they’re not covering advertising the same.
Jared Spool: Okay.
Barbara Coll: I have – first, thank you for launching that Missing Pieces thing [Missingpieces.dogpile.com]. It’s very cool. Second, paid inclusion works phenomenally. And, third, I don’t understand why anybody expects the engines to do anything for free. I don’t know where the expectation came from that Google results need to be free.
First, of all, Mara, I’ll put you on the spot for a minute. You spend – you have a title that has optimization in it, right? Where you do search marketing for Marriott? Because you spend a lot of money on ensuring that you show up ahead in the free listings above your affiliates and above the hotel brokers, the room brokers. That costs you a fortune. That you would rather not spend. Correct? You’d rather not spend it.
Mara Hannula: Well, we don’t spend nearly as much on natural [results] than we do–
Barbara Coll: Right. Absolutely.
Mara Hannula: But we do spend money on natural to optimize it. Otherwise, we wouldn’t appear, perhaps.
Barbara Coll: Exactly. Therefore, it’s not free. And I don’t understand why anybody thinks it should free.
Mara Hannula: I don’t think that was the issue, so to speak. I mean, I thought the issue that we were talking about was around is there confusion or not.
Barbara Coll: No, no, I’m more concerned with why are we even having a discussion as to whether paid inclusion is good or not, when the expectation shouldn’t – I don’t think the expectation should be there’s any confusion – it’s all ads. Like she said. It’s all ads.You know, that’s sort of where I–
Matt Cutts: Well, I’ll take the differing point of view on that.
Barbara Coll: Well, of course, I expect it.
Matt Cutts: I think there’s interesting arguments for pay for inclusion. For example, if you just had a product that just launched, a search engine might take a week, two weeks, a year or two — it might take a month or two months or several months to find that page. So if you’ve just launched something, you might want to be covered more quickly.
Likewise, search engines are not perfect, and they don’t always crawl the site as well as they can, as you heard earlier. So those are a couple of reasons why people might pay to have a search engine crawl them. At Google, we’ve always had the viewpoint that people will trust us more are – like editorial results on the left and clearly marked ads on the right, and people can decide what they want to do. Anything that involves money, pay-for-inclusion, hand-coding search results — all those sort of things are fundamentally a conflict of interest, in our viewpoint. And I think the market has borne out that a lot of users, that model resonates with them.