Interview with Google’s Nick Fox and Diane Tang on Ad Quality Scoring

I had the chance to talk to Nick Fox and Diane Tang from Google’s Ad Quality team about quality scoring and how it impacts the user experience. Excerpts from the article along with additional commentary will be in Friday’s Just Behave column, but here is the full transcript.

Gord: What I wanted to talk about a little bit was just how the  quality, particularly in the top sponsored ads, impacts user experience and talk a little about relevancy. Just to set the stage, one of the things I talked about at SES in Toronto was just the fact that as far as a Canadian user goes, because the Canadian Ad market isn’t as mature as the American one, we’re not seeing the same acceptance of those sponsored ads at the top.  Just because you’re not seeing the brands that you would expect to see for a lot of the queries. You’re not seeing a lot of trusted vendors in that space. They just have not adopted search the same way they have in the States.  What we’ve seen in some of our user studies is a greater tendency to avoid that real estate … or at least to quickly scan it and then move down.  So, that’s the angle I really want to take here.  Just how important it is ad quality and ad relevance to impacting that user experience and then also talking about one thing I’ve always noticed in the number of our user studies. Of all the engines, Google seems to be the most stringent on what it takes to be a qualified ad. To get promoted from the right rails to the top sponsored ads. So that sets a broad framework of what I wanted to talk about today.

Nick: Let me give you a quick overview of who I am and who Diane is and what we work on and then we’ll jump into the topics that you’ve raised.  So what Diane and I work on is called Ad Quality and it is essentially everything about how we decide which ads to show on Google and our partners and what they should look like, how much we charge for them and all those types of things. How the auction works…everything from soup to nuts.  If you ask us what our goal is…our goals is to make sure our users love our ads. If you ask Larry Page what our goal is…it’s to make our ads as good as our search results. So it’s a heavy focus on making sure that our users are happy and that our users are getting back what they want out of our ads.  We sort of think of ourselves as among the first that work on the average product for Google. We represent the user, to make sure the user is getting what they really need.  We’re very similar to what we do on the search quality side, making sure that search results are very good.

I think a lot of the things you’ve picked up on are very accurate. In terms of the focus on top ad quality..in general, the focus on quality..I think what you picked up on in your various  reports as well as the study in Canada are pretty accurate and pretty much what drives what we are working on here.  The big concern that I would have, the main motivation for why I think ad quality is important is as a company we need to make sure users continue to trust our ads.  If users don’t trust our ads, they will stop looking at the ads, and once they stop looking at the ads they’ll stop clicking on the ads and all is lost. So what we need to make sure we are doing in long run  is that the users believe that the ads will provide them what they are looking for and they will continue looking at the ads as valuable real estate and to continue to trust that.
So that is what we are going for. I think as we look at the competitors landscape as well, we see a lot of what you see. We certainly have historically, and continue to do so, have much more of a focus on the quality of the ads. Making sure we’re not doing things where we trade off the user experience against revenue. We all have the ability to show more ads or worse ads, but we take a very stringent approach, as you’ve noticed, to making sure we only show the best ads that we believe the user will actually get something out of. If the user’s not going to get something out of the ad, we don’t show the ad. Otherwise the user is going to be less likely to consider ads in the future.

Diane: It’s worth pointing out that basically what we’re saying is that we are taking a very long term view towards making sure our users are happy with our ads and it’s really about making them trust what we give them.

Gord: One thing I’ve noticed in all my conversations whether they’re with Marissa or Matt or you, the first thing that everyone always says at Google is the focus around the user experience. The fact that the user needs to walk away satisfied with their experience. When we’re talking about the search results page, that focuses very specifically on what we’ve called in our reports the “area of greatest promise”. That upper left orientation on the search results page and making sure that whatever is appearing in that area had better be the most relevant result possible for the user.  In conversations with other engines I hear things like balanced ecosystems and communities that include both users and advertisers. I’ve always been struck by the focus at Google and I’ve always been a strong believer that corporations need sacred cows, these untouchable driving principles that everyone can rally around.  Is that what we’re talking about here with Google?

Nick: I think it is.  I think it comes from the top and it comes from the roots. If we were doing a proposal to Larry and Sergey and Eric where we’re saying, “Hey, let’s show a bunch of low quality ads”  the first question they’re going to ask is “Is this the right thing for the user?”  And if the answer is no, we get kicked out of the room and that’s the end of conversation. So you get that from the top and it permeates all the way through. You hear it when you speak to Marissa and Matt and us. It permeates the conversations we have here as well.  It’s not just external when we talk about the user; it’s what the conversation is internally as well. It just exudes through the company because it’s just part of what we think. I wouldn’t say that there isn’t a focus on the advertiser too, it’s just that our belief is that the way you get that balance is by focusing on the user, and as long as the user’s happy, the user’s clicking on the ad, and as long as the user’s clicking on the ad, the advertiser’s getting leads and everything works. If you focus on the advertiser’s in the short term, maybe the advertisers will be happy in the short term, but in the long term that doesn’t work. That used to be a hard message to get across. It used to be the case that advertiser’s didn’t really get that. And one of the most rewarding things for me is that the advertisers see that, they get that. Some of the stuff we do in the world of ad quality is frustrating to advertisers because in some cases we’re preventing their ads from running in cases where they’d like it to run. We’ve seen that the advertiser community is actually more receptive to that recently because they understand why we’re doing it and they understand that in the long term, they’re benefiting from it as well. I think that you are seeing that there is a difference in approach between us and our competitors. That we believe the ecosystem thrives if you focus on the users first.

Gord: I’d like to focus on what, to me, what’s a pretty significant performance delta between right rail and top sponsored. We’ve seen the scan patterns put top sponsored directly in the primary scanning path of users where right rail is more of a side bar that may be considered after the primary results are scanned. With whatever you can share, can you tell me a little about what’s behind that promotion from right rail to top sponsored?

Nick: Yes, it’s based on two things.  One is the primary element is the quality of the ad. The highest quality ads get shown on the top. The lower quality ads get shown on the right hand side. We block off the top ads from the top of the auction, if you really believe those are truly excellent ads…

Diane: It’s worth pointing out that we never break auction order…

Nick: One of the things that’s sacred here is making sure that the advertiser’s have the incentive. In an auction, you want to make sure that the folks who win the auction are the ones who actually did win the auction. You can’t give the prize away to the person who didn’t win the auction. The primary element in that function is the quality of the ad. Another element of function is what the advertiser’s going to pay for that ad. Which, in some ways, is also a measure of quality. We’ve seen that in most cases, where the advertiser’s willing to pay more, it’s more of a commercial topic. The query itself is more commercial, therefore users are more likely to be interested in ads. So we typically see that queries that have high revenue ads, ads that are likely to generate a lot of revenue for Google are also the queries where the ads are also most relevant to the user, so the user is more likely to be happy as well. So it’s those two factors that go into it. But it is a very high threshold. I don’t’ want to get into specific numbers, but the fraction of queries that actually show these promoted ads is very small.

Gord: One thing we’ve noticed is, actually in an eye tracking study we did on Google China, there where the search market is far less mature, you very, very seldom see those ads being promoted to top sponsored. So I would imagine that that’s got to be a factor. Is the same threshold applied across all the markets or does it vary, does the quality threshold vary from market to market?

Nick:  I don’t want to get too much into the specifics of that kind of detail. We do certainly take an approach in market that we believe is most effective for that market. Handling everything at a global level doesn’t really make a lot of sense because in some cases you have micro markets that, or, in the case of China, a large market, where it makes sense to tailor our approach to what makes sense for that market…what users from that market are looking for, what the maturity of that market is. A market that has a different level of search quality, for example, it might make sense to take a different approach in how we think about ads as well. So that’s what I want to say there. But you’re right, in a market like China that’s less mature and at the early stage of it’s development, you do see fewer ads at the top of the page, there are just fewer ads there that we believe are good enough to show at the top of the page. Contrast that with a country like the U.S. or the U.K., where these markets are very mature and have the high quality ads we feel comfortable showing at the top, we show top ads.

Diane: But market maturity is just one area we look at. There’s also user sophistication with the internet and other key factors. We have to take all this into account to really decide what the approach is on a market basis.

Gord: One of the questions that always comes up every time I sit on a panel that has anything to do with quality scoring is what’s in an ad that might generate a click through is not necessarily what will generate a quality visitor when you carry it forward into conversion. For instance you can entice someone to click through but they may not convert and, of course, if you’re enticing them to click through you’re going to benefit from the quality scoring algorithm.  How do we correct that in the future?

Nick:  I think there are two things. One is, in general, an ad’s that’s being honest, and gets a high click rate from being honest,  is essentially a very relevant ad and therefore gets a high click through rate. We’ll typically see that that ad also has a high conversion rate. In cases where the advertiser’s not being dishonest, the high click through rate is generally correlated with a high conversion rate. And it’s simply because that ad is more relevant, it’s more relevant in terms of getting the user to click on that ad in the first place, it’s also more relevant in delivering what that user is looking for once they actually got to the landing page. So you see a good correlation there.

There are cases where advertisers can do things where they’re misleading in their ad text and create an incentive for a user to click on their ad and then not be able to deliver, so the advertiser could say “great deals on iPods” and then they sell iPod cases or something. In that case, the high click through rate is unlikely to be correlated with a high conversion rate because the users are going to be disappointed when they actually end up on the page. The good thing for us is that the conversion rate typically gets reflected in the amount that the advertiser’s actually willing to pay, so that’s one of the reasons why the advertiser’s bid is a relatively decent metric of the quality, for example in this ipod cases case, because that conversion rates likely to be low, the advertiser’s not likely to bid as much for that. The click just isn’t worth as much to them, therefore they’ll bid less and end up getting a lower rank as a result of that. So, in many cases, this doesn’t end up being a problem because that just sort of falls out of the ranking formula. It’s a little bit convoluted.

Gord: Just to restate it to make sure I’ve got it here. You’re saying that if somebody is being dishonest, ultimately the return they’re getting on that will dictate that they have to drop their bid amount, so it will correct itself. If they’re not getting the returns on the back end, they’re not going to pay the same on the front end and ultimately it will just find it will just find it’s proper place.

Nick: What an advertiser should probably be thinking most about is mostly ROI per click…it’s actually ROI per impression. From the ad that’s likely to generate the most value for the user, and therefore the most value to Google as well as the most value to the advertiser, all aligned in a very nice way, is the ad that’s the most likely to generate the most ROI per impression. And because of our ranking formula, those are the ads that are most likely to show up at the top of the auction. And the ones that aren’t fall out. So the advertiser should care click through rate, but they shouldn’t care about click through rate exclusively to the extent that that results in a low conversion rate and a low ROI per click for them.

Gord: We talked a little bit about ads being promoted to the top sponsored and over the past three or four years, you have experimented a little bit with the number of ads that you show up there. When we did our first eye tracking study, usually we didn’t see any more than two ads, and that increased to three shortly after. Have you found the right balance with what appears above organic results as far as sponsored results?

Diane: I would say that it’s one of those things where the user base is
constantly shifting, the market is constantly shifting. It’s something that we definitely reevaluate frequently. It was definitely a very thought through decision to move to three, and we show three actually very rarely. We seriously consider that when we show three, is it in the best interest for the user? There’s a lot of evaluation of the entire page at that point and not even just the ads, whether or not it was the right thing. We’re very careful to make sure that we’re constantly at the right balance. It’s definitely something that we look at.

Gord: One of the things we’ve noticed in our eye tracking studies is that there’s a tendency on the part of users to “break off” results in consideration sets and the magic number seems to be around four, so what we’ve seen is even if they’re open to looking at sponsored ads, they want to include at least the number one organic result as well, as kind of a baseline for reference. They want to be able to flip back and forth and say, “Okay, that’s the organic result, that how relevant I feel that is. If one of the sponsored ads is more relevant, than fine, I’ll click on it.” It seems like that’s a good number for the user to be able to concentrate on at one time, quickly and then make their decision based on that consideration set that would usually include one or two sponsored ads and at least one organic listing, and where the highest relevancy is. Does that match what you guys have found as well?

Nick: I don’t think we’ve looked at it in the way of consideration sets, along those lines. I think that’s consistent with the outcomes that we’ve had and maybe some of the thought process that lead us to our outcome. The net effect is the same outcome. One of the things that we are careful about is trying to make sure that you don’t want to create an experience where you show no organic results on the page, you know, or at least above the fold on the page. You want to make sure that the user is going to be able to make that decision, regarding what they want to click on and if you just serve the user with one type of result you’re not really helping the user make that type of decision. What we care more about is what the user sees in the above the fold real estate, not quite so much the full result. And probably relatively consistent on certain sets of screen resolutions.

Gord: One of the things that Marissa said when I talked to her a few days ago was that as Google moves into Universal Search results and we’re starting to see different types of results appear on the page, including in some cases images or videos, that opens the door to potentially looking at different presentations of advertising content as well. How does that impact your quality scoring and ultimately how does that impact the user?

Nick: We need to see. I don’t think we know yet. Ultimately it would be our team deciding whether to do that or not, so fortunately we don’t have to worry too much about hooking up the quality score because we would design a quality score that would make sense for it. The team that focuses on what we call Ad UI, that’s the team that’s looking at …it’s sub group within that, that’s the team that essentially thinks about what should the ads actually look like?

Diane: And what information can we present that’s most useful to the user?

Nick: So in some cases, that information may be an image, in some cases that information may be a video. We need to make sure in doing this that we’re not just showing video ads, because video happens to be catchy. We want to make sure that we’re showing video ads because the video is what actually contains the content that’s actually useful for the user. With Universal Search we found that video search results, for example, can contain that information, so it’s likely that their paid results set could be the same as well. Again, just as in text ads, we’d need to make sure that whatever we do there is user driven rather than anything else and that the users are actually happy with it. There would be a lot of user experimentation that would happen before anything was launched along those lines.

Diane: You can track our blogs as well. All of our experiments show up at some point there.

Gord: Right. Talking a little bit about personalization, you started off by saying that Larry and Sergey have dictated that the ads should be more relevant than the organic results in an ideal situation and just as a point of interest, in our second eye tracking study, when we looked at the success rate of click throughs as far as people actually clicking through to a site that appeared to deliver what they were looking for, for commercial tasks, it was in fact the top sponsored ads that had the highest average success rate of all the links on the page. When we’re looking at Personalization, one of the things that, again, Marissa said is we don’t want our organic results and our sponsored results to be too far out of sync. Although personalization is rolling out on the organic side right now, it would make sense, if that can significantly improve the relevancy to the user, for that to eventually fold into the sponsored results as well. So again, that might be something that would potentially impact quality scoring in the future, right?

Nick: Yes. So we have been looking at some.. I’m not sure if the right word is personalization or some sort of user based or task based…what the right word is..changes to how we think about ads. We have made changes to try to get a sense of what the user’s trying to do right now. Whether they’re, for example, in a commercial mind set and alter how we do ads somewhat based on that type of an understanding of the user’s current task. We’ve done much less with trying to..we’ve done nothing really…with trying to build profiles of the user and trying to understand who the user is and whether the user is a man or woman or a 45 year old or a 25 year old. We haven’t seen that that’s particularly useful for us. You don’t want to personalize users into a corner, you don’t want to create a profile of them that’s not actually reflective of whom they are. We don’t want to freak the user out. If you have a qualified user you could risk alienating that user. So we’ve been very hesitant to move in that direction and in general, we think that there’s a lot more we can that doesn’t require profiles down that path.

Diane: You can think of personalization in a couple of different ways, right? It can manifest itself in regards to the results you actually show. It can also be more about how many ads or even the presentation of those ads with regards to actual information. Those sorts of things. There are many possible directions that can be more fruitful than, like Nick points out, profiling.

Gord: Right, right.

Nick: For example, one of the things that you could theoretically do is, as you know, we changed the background color of our top ads from blue to yellow, because we found that yellow works better in general. You might find that for certain users, green is better, you might find that for certain users, blue is actually better. Those types of things, where you’re able to change things based on what users are responding to, is more appealing to us than these broad user classification types of things. It seems somewhat sketchy.

Gord: It was funny. Just before those interview, I was actually talking to Michael Ferguson at Ask.com and one of the things he mentioned that I thought was quite interesting was a different take on personalization. It may get to the point where it’s not just using personalization for the sake of disambiguating intent and improving relevancy, it might actually be using personalization to present results or advertising messages in the form that’s most preferred by the user. So some may prefer video ads. Some may prefer text ads and they may prefer shorter text ads or longer text ads. And I just thought that that was really interesting. Looking at personalization to actually customize how the results are being presented to you. In what format.

Nick: Yes.

Gord. One last question. You’ve talked before about quality scoring and how it impacts two different things. Whether it’s the minimum bid price or whether it’s actually position on the page. And the fact that there’s more factors, generally, in the “softer” or “fuzzier” minimum bid algorithm than there is in the “hard” algorithm that determines position on the page. And ideally you would like to see more factors included in all of it. Where is Google at on that line right now?

Nick: There are probably two things. One is that when setting the minimum bid, we have much less information available to us. We don’t know what the specific query is that the user issued. We don’t know what time of day it is. We know very little about the context of what the user is actually trying to do. We don’t know what property that user’s on. There’s a whole lot that we don’t know. What we need to do when we set a minimum bid is much coarser. We just need to be able to say, what do we think this keyword is, what do we think the quality of the ad is, does the keyword meet the objective of the landing page and make a judgment based on that. But we don’t have the ability to be more nuanced in terms of actually taking into account the context of how the ad is likely to actually show up. There’s always going to be a difference in terms of what we can actually use when we set the minimum bid versus what we use at auction time to set the position. The other piece of it though is there are certain pieces that only affect the minimum bid. Let me give you an example. Landing page quality normally impacts the minimum bid but it doesn’t impact your ranking. The reason for that is mostly from the standpoint of our decision to launch the product and what we thought was the most expedient way to improve the landing page quality of our ads rather than what we think will be the long term design of the system. So I’d expect things like that, where signals like landing page quality should impact not only the minimum CPC but also rank which ads show at the top of the page and things like that as well. That’s where you’ll see more convergence. But there’s always going to be context that we can get at query time to use for the auction than we can for minimum CPC.

Ask: The Reasoning Behind Ask 3D

Last week in my interview with Jakob Nielsen, he called Ask’s 3D label “stupid”. Just to refresh your memory, here’s how the exchange went:

Gord: Like Ask is experimenting with right now with their 3D search. They’re actually breaking it up into 3 columns, and using the right rail and the left rail to show non-web based results.

Jakob: Exactly, except I really want to say that it’s 2 dimensional, it’s not 3 dimensional.

Gord: But that’s what they’re calling it.

Jakob: Yes I know, but that’s a stupid word. I don’t want to give them any credit for that. It’s 2 dimensional. It’s evolutionary in the sense that search results have been 1 dimensional, which is linear, just scroll down the page, and so potentially 2 dimensional (they can call it three but it is two) that is the big step, doing something differently and that may take off and more search engines may do that if it turns out to work well.

My friend Michael Ferguson at Ask (who has his own interview coming up soon) sent me a quick email with the reasoning behind the label:

The 3D label came from the 3 dimensions of search we folded onto one page: query expression in the left rail, results in the center, and content on the right (vs. the one dimension of returning solely results).

Interview with Jakob Nielsen on the Future of the SERP (and other stuff)

jakob-nielsen_cropped.jpg.400x400_q95_crop_upscaleI recently had the opportunity to talk to Jakob Nielsen for a series I’m doing for Search Engine Land about what the search results page will look like in 2010.  Jakob is called a “controversial guru of Web design” in Wikipedia (Jakob gets his own shots in at Wikipedia in this interview) because of his strongly held views on the use of graphics and flash in web design. I have a tremendous amount of respect for Jakob, even though we don’t agree on everything, because of his no frills, common sense approach to the user experience. And so I thought it was quite appropriate I sound him out on his feelings about the evolution of the search interface, now that with Universal search and Ask’s 3D Search we seem to be seeing more innovation in this area in the last 6 months than we’ve seen for the last 10 years. Jakob is not as optimistic about the pace of change as I am, but the conversation was fascinating. We touched on Universal Search, personalization, banner blindness on the SERP and scanning of the web in China, amongst other things. Usability geeks..enjoy!

Gord: For today I only really have one question, although I’m sure there be lots of branch offs from it. It revolves around what the search engine results page may look like in 2010.  I thought you would be a great person to lend your insight on that.

Jakob: Ok, sure.

Gord: So why don’t we just start? Obviously there are some things that are happening now with personalization and universal search results. Let’s just open this up. What do you think we’ll be seeing on a search results page in 3 years?

Jakob: I don’t think there will be that big a change because 3 years is not that long a time. I think if you look back three years at 2004, there was not really that much difference from what there is today.  I think if you look back ten years there still isn’t that much difference.  I actually just took a look at some old screen shots in preparation before this call at some various search engines like Infoseek and Excite and those guys that were around at that time, and Google’s Beta release, and the truth is that they were pretty similar to what we have today as well.  The main difference, the main innovation seems to have been to abandon banner ads, which we all know now really do not work, and replace them with the text ads, and of course that affected the appearance of the page.  And of course now the text ads are driven by the key words, but in terms of the appearance of the page, they have been very static, very similar for 10 years.  I think that’s quite likely to continue. You could speculate the possible changes. Then I think there are three different big things that could happen.

One of them that will not make any difference to the appearance and that is a different prioritization scheme. Of course, the big thing that has happened in the last 10 years was a change from an information retrieval oriented relevance ranking to being more of a popularity relevance ranking. And I think we can see a change maybe being a more of a usefulness relevance ranking. I think there is a tendency now for a lot of not very useful results to be dredged up that happen to be very popular, like Wikipedia and various blogs. They’re not going to be very useful or substantial to people who are trying to solve problems. So I think that with counting links and all of that, there may be a change and we may go into a more behavioral judgment as to which sites actually solve people’s problems, and they will tend to be more highly ranked.

But of course from the user perspective, that’s not going to look any different. It’s just going to be that the top one is going to be the one that the various search engines, by what ever means they think of, will judge to be the best and that’s what people will tend to click first, and then the second one and so on. That behavior will stay the same, and the appearance will be the same, but the sorting might be different. That I think is actually very likely to happen

Gord: So, as you say, those will be the relevancy changes at the back end. You’re not seeing the paradigm of the primarily text based interface with 10 organic results and  8-9 sponsored results where they are, you don’t see that changing much in the next 3 years?

Jakob: No.  I think you can speculate on possible changes to this as well. There could be small changes, there could be big changes.  I don’t think big changes. The small changes are, potentially, a change from the one dimensional linear layout to more of a two dimensional layout with different types of information, presented in different parts of the page so you could have more of a newspaper metaphor in terms of the layout. I’m not sure if that’s going to happen.  It’s a huge dominant user behavior to scan a linear list and so this attempt to put other things on the side, to tamper with the true layout, the true design of the page, to move from it being just a list, it’s going to be difficult, but I think it’s a possibility.  There’s a lot of things, types of information that the search engines are crunching on, and one approach is to unify them all into one list based on it’s best guess as to relevance or importance or whatever, and that is what I think is most likely to happen.  But it could also be that they decide to split it up, and say, well, out here to the right we’ll put shopping results, and out here to the left we’ll put news results, and down here at the bottom we’ll put pictures, and so forth, and I think that’s a possibility.

Gord: Like Ask is experimenting with right now with their 3D search. They’re actually breaking it up into 3 columns, and using the right rail and the left rail to show non-web based results.

Jakob: Exactly, except I really want to say that it’s 2 dimensional, it’s not 3 dimensional.

Gord: But that’s what they’re calling it.

Jakob: Yes I know, but that’s a stupid word. I don’t want to give them any credit for that. It’s 2 dimensional. It’s evolutionary in the sense that search results have been 1 dimensional, which is linear, just scroll down the page, and so potentially 2 dimensional (they can call it three but it is two) that is the big step, doing something differently and that may take off and more search engines may do that if it turns out to work well.  But I think it’s more likely that they will work on ways on integrating all these different sources into a linear list. But those are two alternative possibilities, and it depends on how well they are able to produce a single sorted list of all these different data sources.  Can they really guess people’s intent that well?

All this stuff..all this talk about personalization, that is incredibly hard to do. Partly because it’s not just personalization, based on a user model, which is hard enough already. You have to guess that this person prefers this style of content and so on.  But furthermore, you have to guess as to what this person’s “in this minute” interest is and that is almost impossible to do. I’m not too optimistic on the ability to do that.  In many ways I think the web provides self personalization, you know, self service personalization. I show you my navigational scheme of things you can do on my site and you pick the one you want today, and the job of the web designer is to, first of all, design choices that adequately meet common user needs, and secondly, simply explain these choices so people can make the right ones for them.  And that’s what most sites do very poorly. Both of those two steps are done very poorly on most corporate websites. But when it’s done well, that leads to people being able to click – click and they have what they want, because they know what they want, and its very difficult for the computer to guess what they want in this minute.

Gord:  When we bring it back to the search paradigm, giving people that kind of control to be able to determine the type of content that’s most relevant to them requires them interacting with the page in some way.

Jakob: Yes, exactly, and that’s actually my third possible change. My first one was changing to the ranking scheme; the second one was the potentially changing to two dimensional layouts. The third one is to add more tools to the search interface to provide query reformulation and query refinement options. I’m also very skeptical about this, because this has been tried a lot of times and it has always failed.  If you go back and look at old screen shots (you probably have more than I have) of all of the different search engines that have been out there over the last 15 years or so, there have been a lot of attempts to do things like this. I think Microsoft had one where you could prioritize one thing more, prioritize another thing more. There was another slider paradigm. I know that Infoseek, many, many years ago, had alternative query terms you could do just one click and you could search on them, which was very simple. Yet most people didn’t even do that.

People are basically lazy, and this makes sense.  The basic information foraging theory, which is, I think, the one theory that basically explains why the web is the way it is, says that people want to expend minimal effort to gain their benefits.  And this is an evolutionary point that has come about because the people, or the creatures, who don’t exert themselves, are the ones most likely to survive when there are bad times or a crisis of some kind. So people are inherently lazy and don’t want to exert themselves. Picking from a set of choices is one of the least effortful interaction styles which is why this point and click interaction in general seems to work very well. Where as tweaking sliders, operating pull down menus and all that stuff, that is just more work.

Gord: Right.

Jakob: But of course, this depends on whether we can make these tools useful enough, because it’s not that people will never exert themselves.  People do, after all, still get out of bed in the morning, so people will do something if the effort is deemed worthwhile.  But it just has to be the case that if you tweak the slider you get remarkably better results for your current needs.  And it has to be really easy to understand. I think this has been a problem for many of these ideas. They made sense to the search engine experts, but for the average user they had no idea about what would happen if they tweaked these various search settings and so people tended to not do them.

Gord: Right. When you look at where Google appears to be going, it seems like they’ve made the decision, “we’ll keep the functionality transparent in the background, we’ll use our algorithms and our science to try to improve the relevancy”, where as someone like Ask might be more likely to offer more functionality and more controls on the page. So if Google is going the other way, they seem to be saying that personalization is what they’re betting on to make that search experience better.  You’re not too optimistic that that will happen without some sort of interaction on the part of the user?

Jakob: Not, at least, in a small number of years. I think if you look very far ahead, you know 10, 20, 30 years or whatever, then I think there can be a lot of things happening in terms of natural language understanding and making the computer more clever than it is now. If we get to that level then it may be possible to have the computer better guess at what each person needs without the person having to say anything, but I think right now, it is very difficult.  The main attempt at personalization so far on the web is Amazon.com. They know so much about the user because they know what you’ve bought which is a stronger signal of interest than if you had just searched for something.  You search for a lot of things that you may never actually want, but actually paying money; that’s a very, very strong signal of interest.  Take myself, for example. I’m a very loyal shopper of Amazon. I’ve bought several hundred things from them and despite that they rarely recommend (successfully)…sometimes they actually recommend things I like but things I already have. I just didn’t buy it from them so they don’t know I have it. But it’s very, very rare that they recommend something where I say, “Oh yes, I really want that”. So I actually buy it from them.  And that’s despite the (fact that the) economic incentive is extreme, recommending things that people will buy. And they know what people have bought. Despite that and despite their work on this now for already 10 years (it’s always been one of their main dreams is to personalize shopping) they still don’t have it very well done. What they have done very well is this “just in time” relevance or “cross sell” as it’s normally called. So when you are on one book on one page, or one product in general, they will say, here are 5 other ones that are very similar to the one you’re looking at now. But that’s not saying, in general, I’m predicting that these 5 books will be of interest to you. They’re saying, “Given that you’re looking at this book, here are 5 other books that are similar, and therefore, the lead that you’re interested in these 5 books comes from your looking at that first book, not from them predicting or having a more elaborate theory about what I like.

Gord: Right.

Jakob: What “I like” tends not to be very useful.

Gord: Interesting. Jakob, I want to be considerate of your time but I do have one more question I’d love to run by you.  As the search results move towards more types of images, we’re already seeing more images showing up on the actual search results page for a lot of searches. Soon we could be seeing video and different types of information presented on the page. First of all, how will that impact our scanning patterns?  We’ve both done eye scanning research on search engine results, so we know there is very distinct patterns that we see.  Second of all, Marissa Mayer in a statement not that long ago seemed to backpedal a bit about the fact that Google would never put display ads back on a search results page, seeming to open a door for non text ads.  Would you mind commenting on those two things?

Jakob: Well they’re actually quite related.  If they put up display ads, then they will start training people to exhibit more banner blindness, which will also cause them to not look at other types of multimedia on the page. So as long as the page is very clean and the only ads are the text ads that are keyword driven, then I think that putting pictures and probably even videos on there actually work well.  The problem of course is they are inherently a more two dimensional media form, and video is 3 dimensional, because it’s two dimensional – graphic, and the third dimension is time, so they become more difficult to process in this linear type of scanned document “down the page” type of pattern.  But on the other hand people can process images faster, with just one fixation and you can “grok” a lot of what’s in an image, so I think that if they can keep the pages clean, then it will be incorporated in peoples scanning pattern a little bit more. “Oh this can give me a quick idea of what this is all about and what type of information I can expect”.  This of course assumes as well one more thing which is that they can actually select good pictures.

Gord: Right.

Jakob: I would be kind of conservative until higher tweaking with these algorithms, you know, what threshold should you cross before you put an image up.  I would really say tweak it such so that you only put it up when you’re really sure that it’s a highly relevant good image.  If there starts becoming that there are too many images, then we start seeing the obstacle course behavior. People scan around the images, as they do on a lot of corporate websites, where the images tend to be stock photos of glamour models that are irrelevant to what the user’s there for.  And then people involve behavior where they look around the images which is very contrary to first principals of perceptual psychology type of predicting which would be that the images would be attractive. Images turn out to be repelling if people start feeling like they are irrelevant. It’s a similar effect to banner blindness. If there’s any type of design element that people start perceiving as being irrelevant to their needs, then they will start to avoid that design element.

Gord: So, they could be running the risk of banner blindness, by incorporating those images if they’re not absolutely relevant…

Jakob: Exactly.

Gord: …to the query. Ok thank you so much.  Just out of interest have you done a lot of usability work with Chinese?

Jakob: Some. I actually read the article you had on your site. We haven’t done eye tracking studies, but we did some studies when we were in Hong Kong recently, and to that level the findings were very much the same. In terms of pdf was bad and how people go though shopping carts. So a lot of the transactional behavior, the interaction behavior, is very, very similar.

Gord: It was interesting to see how they were interacting with the search results page.  We’re still trying to figure out what some of those interactions meant

Jakob: I think it’s interesting. It can possibly be that the alphabet or character set is less scannable, but it is very hard to say because when you’re a foreigner, these characters look very blocky, and it looks very much like a lot of very similar scribbles.  But on the other hand, it could very well be the same, that people who don’t speak English would view a set of English words like a lot of little speck marks on the page, and yet words in English or in European languages are highly scannable because they have these shapes.

Gord: Right.

Jakob: So I think this is where more research is really called for to really find out.  But I think it’s possible, you know the hypothesis is that it’s just less scannable because the actual graphical or visual appearance of the words just don’t make the words pop as much.

Gord: There seems to be some conditioning effects as well and intent plays a huge part.  There’s a lot of moving pieces with that and we’re just trying to sort out. The relevancy of the results is a huge issue because the relevancy in China is really not that good so…

Jakob: It seems like it would have a lot to do with experience and amount of information.  If you compare back with uses of search in the 80’s, for example, before the web started, that was also a much more thorough reading of search results because people didn’t do search very well. Most people never did it actually, and when you did do it you would search through a very small set of information, and you had to carefully consider each probability. Then, as WebCrawler and Excite and AltaVista and people started, users got more used to scanning, they got more used to filtering out lots of junk. So the paradigm has completely changed from “find everything about my question” to “protect myself against overload of information”.  That paradigm shift requires you to have lived in a lot of information for awhile.

Gord: I was actually talking to the Chinese engineering team down at Yahoo! and that’s one thing I said. If you look at how the Chinese are using the internet, it’s very similar to North America in 99 or 2000. There’s a lot of searching for entertainment files and MP3s. They’re not using it for business and completing tasks nearly as much. It’s an entertainment medium for them, and that will impact how their browsing things like search results. It’ll be interesting to watch as that market matures and as users get more experienced, if that scanning pattern condenses and tightens up a lot

Jakob: Exactly. And I would certainly predict it would. There could be a language difference, basically a character set as we just discussed, but I think the basic information foraging theory is still a universal truth. People have to protect themselves against information overload, if you have information overload. As long as you’re not accustomed to that scenario, then you don’t evolve those behaviors. But once you get it… I think a lot of those people have lived in an environment where there’s not a lot of information.  Only one state television channel and so forth and gradually they’re getting satellite television and they’re getting millions of websites. But gradually they are getting many places where they can shop for given things, but that’s going to be an evolution.

Gord: The other thing we saw was that there was a really quick scan right to the bottom of the page, within 5 seconds, just to determine how relevant these results were, were these legitimate results? And then there was a secondary pass though where they went back to the top and then started going through. So they’re very wary of what’s presented on the page, and I think part of it is lack of trust in the information source and part of it is the amount of spam on the results page.

Jakob: Oh, yes, yes.

Gord: Great thanks very much for your time Jakob.

Jakob: Oh and thank you!

Interview with Ask’s Michael Ferguson

I recently had the opportunity to chat with one of my favorite usability people, Michael Ferguson at Ask.com. You can find excerpts of the interview, along with commentary, on Search Engine Land in this week’s Just Behave column. Some of Michael’s comments are particularly timely now, given Google’s announcement of Universal search.

Gord: How does Ask.com approach the search user experience and in big terms, what is your general philosophy?

Michael: A lot of what we do is, to some extent, informed by core search needs but also by our relevant market share, understanding that people have often experienced other engines before they come to us, not necessarily in that session but generally on the web. People have at least done a few searches on Google and Yahoo, so they have some context coming from those search experiences. So often, we’re taking what we’ve learned from best practices from competitors and others and then, on top of that, trying to add a lot of product experience and relevance experiences that are differentiated. Of course, we’re coming from this longer history of the company where we’ve had various user experiences over the time that we’ve been around. We’ve marketed around natural language, in the late 90’s and answered people’s questions at the top of the page, but in the last year and a half or so, we’ve rebranded and really focused on getting the word out to the end users that we are a keyword search engine, an everyday search engine.

A lot of the things that we’ve done with users have been to try to, implicitly, if not explicitly, inform users that are coming to the site you can use it very much like you can use any other kind of search engine you’ve been on before. Or, if they’re current users and people are coming back to the site, to let them know that the range of experiences and the type of information we bring back to them has greatly expanded. So that’s pretty much it. It’s informed by the context of not just a sense of pure search and information retrieval and all the research that’s gone on that in the last 35 or 40 years but also the dynamics of the experiences that we’ve had before and people’s previous experiences with Ask. Then, an acknowledgement that they’ve often searched on other sites and looked for information.

Gord: You brought up a number of topics that I’d like to touch on, each in sequence. You mentioned that in a lot of cases, they’re coming to Ask and they’ve used Google or Yahoo or they’ve used another engine as one of their primary search tools. Does Ask’s role as a supplemental engine or an alternative engine give you a little more latitude? You can add things from a functionality point of view to really differentiate yourselves. I actually just did a search and see that you, at least on my computer here, have made the move to incorporate some of the things that you were testing on AskX into the main site. Maybe we’ll start there. Is that an ongoing test? Am I just part of a beta test on that or this rollover complete now?

Michael: We’re still in testing with that and it will roll out. We have decided because of a lot of the user experience metrics that we’re getting from the beta test that we’re going to go for it. We have decided to move the full experience over to the AskX experience. Of course, there are variants to that, but the basic theme of, in a smart way, bringing together results from different search verticals and wrapping those around the core organic results (as well as) a sponsored experience. So that will happen sometime this year. We don’t know exactly when, but just a couple of days ago, we really decided we’ve seen enough and we’re pretty excited about that.

Google has a really great user experience going, and Yahoo does too, but they have so many different levers that move so much revenue and traffic and experience metrics that I think it’s harder for them to take chances and to move things around and get buy-offs at a bureaucratic level. To some extent, we see ourselves as having permission and a responsibility to really innovate on the user experience. It’s definitely a good time for us because we have such great support from IAC and they’re very much invested in us improving the user experience and getting more traffic and getting frequency and taking market share and they’re ready to very much invest in that. So we don’t need to cram the page with sponsored links and things like that. It’s mostly a transitional time when we’re getting people to reconsider the brand and the search engine as a full keyword based, everyday search engine that has lots to offer. I’m talking to people all the time about Ask and there’s definitely still people that say, “Hey, last night, it came up with my buddies at the bar, this trivia question about the Los Angeles Lakers, 1966 to 1972 (and I went to Ask and asked a question)”. Then there are other people that see us as evolving beyond that but still really surprised that we haven’t had image search.  Now with AskX we’ll have preview search and there’s lots of other stuff coming along now. So yes, it’s a great place to be. I love working with it. There are so many things that, in an informed way, we can take chances on, relative to our competitors.

Gord: So does this mean that the main site becomes more of an active site? Are you being more upfront with the testing on Ask.com rather than on AskX.com?

Michael: Well, I think the general sense of what we’re going to do is that, at some point this year, the AskX experience will, at least at a wireframe level, become the default experience and, of course, we have a lot of next generation “after that” stuff queued up that we’re thinking about and we’re actively testing right now but not in any live sense.  So potentially, things will slide in behind the move of the full interface going out and then AskX will remain a sandbox for another instance of, hopefully, new and really useful and differentiated search experience coming after that. A general thing that we’re going to try to do, instead of having 15 or 18 different product managers and engineering teams working on all these different facets of information retrieval and services, we’re going to stay search focused and just have one sandbox area where people go in and see multiple facets of what we’re thinking about.

Gord: Let’s talk about the sponsored ads for a bit. I notice that for a couple of searches that I’ve done while we’ve been talking that they’ve definitely been dialed down as far as the presence of sponsored on the page. I’m only seeing top sponsored appear, so you’re using the right rail to add additional search value or information value, whether it be suggested searches or on a local search, where it brought me back the current weather and time. So what’s the current strategy on Ask as far as presentation of sponsored results and the amount of real estate devoted to them?

Michael: Just to fit along with the logic of Eye Tracking II (Enquiro’s second eye tracking study), those ads are not a delineated part of the user experience for the end user and they’re relevance and their frequency can color the perception of the rest of the page and especially the organic listings below them. Right now, as I said, we’re very much focusing on improved user experience and building frequency and retention of customers, which all the companies are, I’m sure. But we’re really being, basically, cautious with the ads and getting them there when they’re appropriate and, as best we can, adjust them over time, so that when they’re there, they’re going to valuable for the user and for the vendor.

Gord: That’s a fairly significant evolution in thinking about what the results page looks like from say, two years ago, with Ask. Is that purely a function of IAC knowing that this is a long term game and it begins with market share and after that comes the monetization opportunities?

Michael: Actually, I think way before we got acquired by IAC we knew that. We test like other engines would. We test lots of different ad configurations and presentations and things like that but definitely you want to balance that. Way before we got acquired we realized that there’s one thing that’s kind of fun about making the quarter and blowing through it a little bit and then there’s another thing about eroding customers. And definitely there’s a lifetime value that can be gained by giving people what you know is a better user experience over time, so once we did become part of the IAC family, we brought them up to speed with the results that we were finding that were pointing to taking that road and they’ve very much been in support of it. And, of course, their revenue is spread amongst a lot of different pieces of online and offline business so their ability to absorb it is probably more flexible than ours was as a stand alone company.

Gord: That brings me to my next question, which is, with all the different properties that IAC has and their deep penetration into some of the vertical areas, you had talked about the opportunity to bring some of that value to the search results page. What are we looking at as far as that goes? Are we going to see more and more information pulled from other IAC into the main AskX interface?

Michael: Maybe the most powerful thing about the internet is that you as an individual now have a very empowered position relative to other producers of information, other businesses where you can consume a bunch of different points of view. You have a bunch of different opportunities to do business and get the lowest price and read reviews that the company itself hasn’t sanctioned, or anything like that.  You have access to your peer network and to your social networks. Search, like the internet, becomes, and it necessarily needs to be, a proxy for that neutral, unbiased view of all the information that’s available. This probably gets a little bit into what may or not may work with something like Google’s search history. Users over time have said again and again, “Don’t hide anything from me or don’t over think what you may think I might want. Give me all of the best stuff, use your algorithms to rank all that, but if I get the sense that anything’s biased or people are paying for this, then I’m not going to trust you and I’m going to go somewhere else where I can get that sense of empowerment again.”

As I’ve sat in user experience research over time, I’ve seen people..and I know this isn’t true of Google and I know it isn’t true of Ask right now with the  retraction from paid inclusion…but you ask users why they think this came up first on Google, maybe with a navigational query like Honda or Honda Civic and Honda comes up first. They’ll say, “Oh, Honda paid for that.” So even with the engines that aren’t doing paid inclusion, there’s still this kind of wariness that consumers have of just generally somebody on the internet, somewhere, behind the curtains, trying to take advantage of them or steer them in some way. So as soon as we got acquired by IAC, we have made it very much part of their perception of this and their culture. Their product management point of view is that you can’t sacrifice that neutrality. You can’t load a bunch of IAC stuff all over the place. The relationship with IAC does give us access to proprietary databases that we can do lots of deep dives in and get lots of rich information out  that can help the user in their instance of their search needs that other companies wouldn’t be able to get access to, while maintaining access to everything else.

The way we approached AskCity was a great example of this. We had leveraged a lot of CitySearch data but at the same time, we know that when people go out and want to see reviews, they want to see reviews from AOL Neighborhoods, they want to see reviews from Yelp they want to see reviews from all these other points of view too. So we go and scrape all those and fold them into the CitySearch stuff. We give access to all those results that come up on AskCity. If they’re, for instance, at a restaurant, you can get Open Table reviews and you can get movie reservations through Fandango and other stuff like that. Those companies have nothing to do with IAC. Those decisions were borne from user needs and from us looking as individuals in particular urban areas, and saying “Hey, what would I want to come up?” We know from previous experience from AOL that the walled garden thing doesn’t work. It’s just not what people expect from search and not what they expect from the internet, so that lesson’s been learned. I don’t know how much it would be different if we had some dominant market share over search, but that’s even more reason for us to be appealing to as wide a population as possible. That’s my philosophy right now.

Gord: I guess the other thing that every major engine is struggling with right now is in this quest to disambiguate intent, where is the trade-off with user control? Like you said, just show me a lot of the best stuff and I’ll decide where I want to drill down and I’ll change the query based on what I’m seeing to filter down to what I want. In talking to Marissa at Google and their moves towards personalization and introducing web history, I  think for anyone who understands how search engines work, it’s not that hard to see the benefits of personalization but from a user perspective there does seem to be some significant push back against that. Some users are saying, “I don’t want a lot of things happening in the background that are not transparent to me. I want to stay in control.” How is Ask approaching that?

Michael: The other major thing that’s going on right now is that we have fully revamped how we’re taking this. We developed the Direct Hit late 90’s technology. And then the Teoma technology we acquired. And really, it’s not that we’re taking those to the next level, we got all of that stuff together and over the past three years, we’ve been saying, “Okay, what do we have and what’s unique and differentiated?” There’s a lot of great user behavior data that Direct Hit understands.  We have a whole variety of things there and that’s unlocked, that’s across all the people coming in and out over time but not any personally identifiable type of stuff. And then there’s Teoma, which is good at seeing communities on the web, expertise within the communities and how communities relate. So right now, even though we have personalization stuff and My Stuff and other things that are coming up, we’re investing a lot more in the next version of the algorithm and the infrastructure for us to grow called Edison. And we started talking about that a week ago since A.G. (Apostolos Gerasoulis) mentioned it. Across a lot of user data it understands a lot about the context from the user intention side and because we’re constantly capturing the topology of the web and it’s communities and how they’re related, we then match the intention and the map of the web as it stands and the  blogosphere as it stands and other domains as they stand. Our Zoom product, which is now on the left under the search box in the AskX experience and it’s on the right on the live site, is the big area that we’re going to more passively offer people different paths.

For example, just like with AskX, you search for U2, it’s going to bring up news, and product results, and video results and images, and a Smart Answer at the top of the page. It’s also going to know that there’s U2 as the entity, the music band and therefore search the blogosphere but just search within music blogs. So what it’s doing, over time, is trying to give a personalized experience that’s informed by lots of behavior and trying to capture the structure of the web, basically. So that’s where we are there.

There’s a book that came out in early 1999 called Net Worth, which you might want to read. I almost want to revisit it myself now. It’s a Harvard Business School book that Marc Singer and John Hagel came out with. It talked about infomediaries and it imagined this future where there’d be these trusted brands and companies. They were thinking along the lines of American Express or some other concurrent banking entity at the time, but these infomediaries would have outside vendors come to them and they would entrust all their information, as much as they wanted to, they could control that, both online and offline.  You were talking in your latest blog post about understanding in the consideration phase where somebody is and presenting, potentially, websites that they hadn’t seen yet or ones that they might like at that point in the car purchase behavior. But the way that they were imagining it was that there would be a credit card that might show that someone had been taking trips from the San Francisco Bay area to the Tahoe region at a certain time of year and had maybe met with real estate agents up there and things like that. But these infomediaries, on top of not just web history but even offline stuff, would be a broker for all that information and there would be this nice marketplace where someone could come and say, “I want to pay $250 to talk to this person right now with this specific message”. So it seems that Google is doing a lot of that, especially with the DoubleClick acquisition. But I’m just wondering about the other side of it, keeping the end user aware of and empowered over that information and where it’s at. So Net Worth is a neat book to check out because the way they were describing it, the end user, even to the broker, would seep out exactly what they wanted to seep out at any given time. It wouldn’t be this passive recording device thing that’s silently taping. My experience so far of using the Google Toolbar that’s allowing the collection of history, is that it’s ambiguous to me about how much of my behavior is getting taken up by that system and used. We’ll see where it goes but right now we don’t have strong plans to do anything with that for search.

Gord: It’s going to be really interesting because, up to now, the tool bar was collecting data but there was no transparency into what it was collecting, and now that they’ve done that, we’ll see what the user response is to that. Now that they can go into their web history and have that initial shock of realizing how much Google actually does know about them.

One other question, and this is kind of a sidelight, but it’s always something that I’ve been interested in. Now that you have the search box along the left side there and it gives search suggestions as you’re typing, have you done any tracking to see how that’s altered your query logs? Have you noticed any trends in people searching differently now that you’re suggesting possible searches to them as they’re typing?

Michael: There are two broad things that are encouraging to us. One is that over time, the natural language queries are down tremendously. Our queries, because we promoted in the late nineties this “ask a question” thing, tended to be longer and more phrase based, more natural language based.  That’s really gone down and is approaching what we would consider normal for an every day search engine profile as far as the queries. And we really think that this zooming stuff has really helped that because it’s often keyword based. You will sometimes see some natural language stuff in there. There are communities on the web that are informing us that there’s an interest in this topic that’s related to the basic topic so it is helping change the user behavior on Ask.

And the other result of that is as people use it more for everyday keyword based search engine, the topics or the different categories of queries that people see are normalizing out too. Less and less they’re reference type stuff and more and more they’re transactional type queries, so that’s a good thing. And that’s just been happening as we rebranded and we presented Zoom.

And then with the AskX experience, we are definitely seeing that even more because of the fact that they’re just in proximity to the search box. We always knew that these suggestions should ideally be close to the search box so that people understand fully what we’re trying to offer them. For instance, on the current site, we do see users that will sometimes type a query in the search box on top and because they’re used to seeing ads on the right rail on so many other sites and because they don’t necessarily know what narrow and expand your search is they think those are just titles to other results or websites. It’s a relatively small portion. Most people get what it is, but there was that liability there. Now in the AskX experience, it’s close and visually grouped with the search box. It’s definitely getting used more and guiding queries and people seem even more comfortable putting general terms in. We’ve made it that you can just arrow down to the one and hit return. It’s definitely driving the queries differently.

Gord: I’ve always liked what you guys have done on the search page. I think it’s some of the most innovative stuff with a major search property that I see out there and I think that there’s definitely a good place for that kind of initiative. So let me wrap up by asking, if you had your way, in two years, what part would Ask be playing in the total search landscape?

Michael: We’d definitely have significantly more than 10% market share. My point of view, from dealing with the user experience, is that I’ve been proud of the work that we’ve done and I really think that we’ve been very focused and innovative with a very talented team here and we’re really hoping that as we look at the rest of the year and we put out Edison and the AskX experience, that we become recognized for taking chances and presenting the user experience in a differentiated way that people have to respond to us in the market and start adopting some of the things that we’re doing. Because of the amount of revenue that Microsoft, Yahoo and Google are dealing with on the search side, they often get a lot of press but our hope is really to take share and to hopefully have a user experience that inform and improve the user experience of our competitors.

Gord: Thank you for your time Michael.

Matt Cutts: Personalization and the Future of SEO

mattcutts555I had the chance to interview Matt Cutts this week about personalization and it’s impact on the SEO industry. Excerpts from the interview and some additional commentary are in my Just Behave column on Search Engine Land today. As promised, here is the full transcript of the interview:

Gord: We’ve been awhile setting this up, and actually, this came from a discussion we had some time ago about geo-targeting of results in Canada, and we’re going to get to that a bit later. With this recent move by Google to move towards more personalization of the search results page, there’s some negative feedback and, to me, it seems to be coming from the SEO community. What’s your take on that?

Matt: I think that it’s natural that some people would be worried about change, but some of the best SEO’s are the SEO’s that are able to adapt, that are able to look down the road 4 or 5 years and say, “What are the big trends going to be?” and adjust for those trends in advance, so that when a search engine does make a change which you think is inevitable or will eventually happen, they’ll be in a good position. Personalization is one of those things where if you look down the road a few years, having a search engine that is willing to give you better results because it can know a little bit more about what your interests are, that’s a clear win for users, and so it’s something that SEO’s can probably predict that they’ll need to prepare for. At the same time, any time there’s a change, I understand that people need some time to adjust to that and need some time to think, “How is this going to affect me? How is this going to affect the industry? And what can I do to benefit from it?”

Gord: It seems to me, having a background in SEO, that the single biggest thing with personalization is the lack of a “test bed”, the lack of something to refer to when you’re doing your reverse engineering. You can’t look at a page of search results any more and say “that’s going to be the same page of test results that everyone’s seeing“. Given that, , more and more, we’re going to be seeing less of universal search results, is this the nail in the coffin for shady SEO tactics?

Matt: I wouldn’t say that it’s necessarily the nail in the coffin, but it’s clearly a call to action, where there’s a fork in the road and people can think hard about whether they’re optimizing for users or whether they’re optimizing primarily for search engines. And the sort of people who have been doing “new” SEO, or whatever you want to call it, that’s social media optimization, link bait, things that are interesting to people and attract word of mouth and buzz, those sorts of sites naturally attract visitors, attract repeat visitors, attract back links, attract lots of discussion, those sorts of sites are going to benefit as the world goes forward. At the same time, if you do choose to go to the other fork, towards the black hat side of things, you know you’re going to be working harder and the return is going to be a little less. And so over time, I think, the balance of what to work on does shift toward working for the user, taking these white hat techniques and looking for the sites and changes you can implement that will be to the most benefit to your user.

Gord: It would seem to be that there’s one sector of the industry that’s going to be hit harder by this, and I think it was Greg Boser or Todd Friesen who said, “You don’t take a knife to a gun fight.” So when you’re looking at the competitive categories, like the affiliates, where you don’t have that same site equity, you don’t have that same presence on the web to work with, that’s where it’s going to get hit, right?

Matt: I think one area that will change a lot, for example, is local stuff. Already, you don’t do a search for football and get the same results in the U.K. as you do in the U.S. So there are already a lot of things that return different search results based on country, and expect that trend to continue. It is, however, also the case that in highly commercial or highly spammed areas, if you are able to return more relevant, more personalized results, it gets a little harder to optimize, because the obstacles are such that you’re trying to show up on a lot of different searches rather than just one set of search engine result pages, so it does tilt the balance a little bit, yes.

Gord: I had a question about localization of search results, and I think being from Canada we’re perhaps a little bit more aware of it. How aware are American SEO’s that this is the case, that if  they’re targeting markets outside the U.S., they may not be seeing the same results that you’re seeing in the U.S.

Matt: I think that many SEO’s are relatively aware, but I’ve certainly talked to a few people who didn’t realize that if you do a search from the U.K., or from Canada, or from India, or from almost any country, you can get different results, instead of just the standard American results. And it’s definitely something that’s a huge benefit. If you’re in the United Kingdom and you type the query newspapers, you don’t want to get, necessarily, the L.A. Times or a local paper in Seattle, the Post-Intelligencer. Something like that. So I think it’s definitely started down that trend, and, over time, personalization will help a lot of people realize that it’s not just a generic set of results, or a vanilla set of results. You have to be thinking about how you’re going to show up in all of these markets, and personalization and localization complement each other in that regard.

Gord: Now one difference between localization and personalization is that personalization has the option of a toggle, you can toggle it on and off. Localization doesn’t have that same toggle, so as a Canadian, sometimes I may not want my results localized. Where does that put the user?

Matt: It’s interesting, because you have to gauge…and you talked to Marissa a couple times already, and from that you probably got a feel for the difficulty in making those decisions about just how much functionality to expose, in terms of toggles and advanced user preferences and stuff like that. So what we try to do is tackle the most common case and make that very simple. And a lot of the times, the functionality is such that you don’t even necessarily want someone that’s coming in from the U.K. to be able to search as if they’re coming in from Africa because it just makes things a lot more complicated. So, over time, I’d say we’re probably open to lots of different ways of allowing people to search. For example, you can select different countries for the advertisements. There’s a GL parameter I believe, where you can actually say, “now, show the ads as if I were searching from Canada. Okay, now I’m going to switch to Mexico”. Stuff like that. And that’s been very helpful, because if you giving Google money to buy ads, you want to be able to check and see what those ads would look like, in different regions. For search we haven’t historically made that as easy. It’s something that we’d probably be open to, but again, it’s one of those things where probably SEO’s are a lot more interested,you’re your regular user isn’t quite as interested.

Gord: And that gets to the ongoing problem. SEO’s have one perspective, users have another and arguably, yes, localization is good for the user, but for an SEO that deals with a lot of Canadian companies where the U.S. is their primary market. They’re looking at hitting that U.S. market. I guess this restricts them to making it look like their sites actually reside in the U.S. to get around it. So again, we’re trying to poke holes in the functionality, rather than live with it.

Matt: Well, one thing that should be possible is to indicate some sort of preference, or some sort of origin of location where you can indicate where you are. Historically Google has been ahead of the other search engines at the time by not just using the top level domain, so .ca, but also the I.P. address. So you can have .com hosted in Canada and that’s worked very well for many, many years. But we do continue to get feedback that people would like more flexibility, more options, so it’s a matter of deciding how many people that would help and just allocating resources on those types of things.

Gord: So we talked about personalization, we talked about localization. Are there other factors that are coloring the search results we should be aware of as we’re trying to consider all these aspects?

Matt: Once you’ve sort of “broken the mould” with different results for different countries, after that it’s good for people to move beyond the idea of a monolithic set of search results. If we had the ability to say someone is searching for Palo Alto or someone is searching for Kirkland or Redmond and give them local newspapers, truly local newspapers, that would be a good win for users as well.  So over time, I would expect search results to serve a broader and broader array of services.  The idea of a monolithic set of search results for a generic term will probably start to fade away, and you already see people expect that if I do a search and somebody else does the search, they can get slightly different answers. I expect that over time people will expect that more and more, and they’ll have that in the back of their heads.

Gord: Let’s take that “over time” and drill down a little more.  One of the things it was interesting for me when I was talking to Marissa with the fact that the Kaltix acquisition was made four years ago and it’s really taken four years for that technology to really show up in the search results.  Obviously a cautious approach to it.  And even with that we’re talking a couple of results being lifted into the top 10 and we’re talking one in five searches.  Also Marissa wasn’t exactly sure about this so I’ll clarify this with you.  She believed that it would never replace the number one organic result.

Matt: I believe that’s correct.  I’d have to double check to make sure.

Gord: So that’s a fairly tentative step in the direction of personalization, and you said over time we can expect this to continue to ship to be more of an individual experience.  Are we talking months, are we talking years, are we talking tomorrow?

Matt (chuckling): It’s usually not our policy to comment on exactly when stuff might roll out in the future, but personalization is an important trend and the ability to make search results better through personalization is really exciting to us here at Google.  I think if you look backwards over time, a lot of the reason why we might not have been able to personalize before was because Google was very much a “you come to the front page, you do a search, you get the results and you’re gone” type of model.  And there really weren’t that many opportunities to have a single sign on or some sort of Google account, where we could actually learn or know a little bit more about you to make your results more relevant.  So I think part of it involves getting all of the different ways of having an account together, so you can have personalized news, which rolled out a while ago, you could have a personalized homepage and those things give people a reason to sign in to Google.  Once you’re signed in to Google that helps us a lot more, by having your search history and the ability to offer personalization.  So at least looking backwards, I think some of the amount of time was just getting people ready to have a Google account and not just show up in Google, do a search and leave.

Gord: So part of it is that transition from a tool you use to more of a community you are engaged in.

Matt: Yes

Gord: That’s moving closer to your competition, notably Yahoo and Microsoft.  Google’s done very well as a tool.  Is this just the inevitable progression?

Matt: I think one nice thing is that Google adapts very well to what users want, and also the industry marketplace.  And so when our primary competition was a pure search engine, whether it be AltaVista or AlltheWeb or HotBot or Inktomi, then pure search mattered very much.  Search is still a part of everything we do.  It’s at the core of all the information that we organize and yet competing against sites like Yahoo and Microsoft involves a different set of strategies than competing against just a search engine for example.  So I think competition is very good for users, because it makes all of us work hard and it keeps us on our toes.  The one strength that Google has is that we do adapt and we look at the marketplace and we say, “What do we need to deliver next for our users to help them out and to encourage them to be more loyal to Google?”

Gord: So for your job, where you’re looking at the quality of the index and policing it, how does personalization change your job?

Matt: To some degree, it makes it easier, because it’s not one monolithic set of search results anymore.  But let me flip that around and say how we can make it easier for SEO’s as well.  I’m a computer graphics person, so if you go back to a concept called digital half toning, it’s this process where you have nothing but black and white, yet you are able to approximate different shades of gray. And if you look at the existing set of search results, a lot of people before had a very black or white mentality.  I’m ranking, or maybe I’m ranking number one or are not in the search results at all.  And that’s a very harsh step function, in terms of you not ranking where you think you should be, and maybe you’re not getting very much traffic at all.  If you are ranking number one, or very highly, you’re a very happy person.  And yet that monolithic set of search results may not serve users the best.  So now as we see that spread and soften, more people can show up at number one but for a smaller volume of queries.  And so individual users are happier because they’re getting more relevant search results and yet it’s not a winner take all mentality for SEOs anymore.  You can be the number one ranking set of results for your niche, whether it be a certain demographic or a certain locality, or something like that.  And I think that’s healthier overall, rather than having just a few people that are doing very well, you end up with a lot more SEO, and a lot more users who are happy and that’s softens the effect quite a bit.

Gord: What you’re talking about is a pretty fundamental shift in thinking on the part of a lot of SEOs…

Matt: yes

Gord: … a lot of SEOs are almost more engineers right now, where they’re looking at the algorithm and trying to figure out how to best it.  You’re asking them to become a lot of things, more marketing, PR, content developers, and know more about the user, more about user behavior online.  These are very different skill sets and often don’t reside in the same body.  What is this going to do to the SEO industry?

Matt: I think the SEO’s that adapt well to change an optimized for users are going to be in relatively good shape, because they’re trying to produce sites that are really pleasing and helpful to users.  It’s definitely the case that if all you care about is an algorithm than the situation grows more complicated for you with personalization.  But it’s also an opportunity for people to take a fresh look at how they do SEO.  So give you a quick example: we always say, don’t just chase after a trophy phrase.  There are so many people who think if I ranked number one for my trophy phase I win and my life will be good.  When, in fact, numerous people demonstrated that if you chase after the long tail and make a good site that can match many many different user’s queries you might end up with more traffic than if you had that trophy phrase.  So already the smart SEO, looking down the road, knows that it’s not just the head of the tail, it’s the long part of the tail and with personalization and the changes in how SEO will work, it will just push people further along the spectrum, towards looking at “it’s not just looking at a number one result for one query, how do we make it across a lot of queries.  What value do I deliver?  Am I looking at my server logs to find queries that I should be targeting?  And not just search engines, how do I target different parts of the search engine?  Like the local part of Google, the maps part of Google.  How do I target Google notebook and the other properties and how do I show up well across the entire portfolio of search properties?”  And that’s a healthy transition period that will push people towards delivering better value for their users and that’s better for everybody.

Gord: I get that and I’m an SEO.  My challenge comes in getting my clients, who in a lot of cases did their own SEO or worked with another SEO firm before they came to us and are used to that trophy phrase ranking.  How do we get them to get?  Because I see that being a challenge with a lot of SEOs. They will understand that, but getting the client to understand it could be a different matter

Matt: Sometimes I think you might have to do a demonstration like sign them into personalized search, do a query, sign them out, do query and show them, these are very different sets of results.  And sometimes the demonstration can be very visceral, you know, it can drive home the point that it’s not just going to be this one trophy phrase. People are going to have to think and look at the entire horizon of the space.

Gord: In Google there’s a very definite church versus state divide and traditionally the relationship with the advertiser was almost exclusively on one side of that divide.  But this could mark a fairly fundamental shift, and it will impact your advertisers, so as part of that community, will Google be doing anything to help those advertisers understand the organic part of their visibility on Google?  Will you be doing the same demonstration you just telling us we should be doing?

Matt: I think Google is always trying to communicate with the outside community, both with webmasters and advertisers.  So it’s really exciting to see some of the different techniques that we’ve used, everything from webinars to training materials to making videos available.  I would definitely say that every part of Google is going to keep their eyes open on how to best communicate how to stay on top of changes, because nobody wants anybody outside of Google to be unprepared for personalization or improvements in any of our technologies.

Katie (Katie Watson, Google PR representative who was sitting in on the interview) Something to actually cite there is that I know we recently just opened up our webmaster blog to outside comments, so that’s a good example of gradually moving forward to communicate even better.

Matt: You were couching the question in terms of advertisers, but if you look at the general story of webmaster communication and assume that that’s the leading edge, it’s pretty safe to assume that those smart ideas are percolating throughout the company and we’re trying to figure out all the different ways to communicate more.

Gord: So that’s the canary in the coal mine. Whatever’s happening in the webmaster community will act as a testbed for communication?

Matt: Exactly.

Gord: There is a debate raging right now about “is SEO rocket science”?  (Matt begins laughing) So what does personalization means for that debate?  Does it become more complicated?  You said it becomes easier in some ways and I countered that by saying that may be, but is also spreading out in a lot of different directions. Is there still a place for the pure SEO consultant out there?

Matt: I think there still is a place for you for a pure SEO consultant but it’s also true that over time those consultants have to keep adding to their skill set.  A few years ago no one would have even thought about the word Ajax and now people have to think about Ajax or Flash and how do I handle some of these new interfaces to still make sites crawlable?  So I definitely think there will still be places for consulting and improving crawlability of sites and advice on keywords and personalization will add some wrinkles to that, but I have faith that, over time we’ll see the benefit to users and if you make good site for your users, you will naturally benefit as a result.  Some people spend a lot of time looking at data centers and data center IP addresses and if people want to have that as a hobby they’re welcome to it but a lot of people don’t do that anymore and they’re just worried about making good results and yet, everything still comes out pretty well for them.

Gord: Some time ago I wrote a column along that line and said that, in many ways, the white hat SEO has helped clean up the Black hat side of the street because they enabled those good site to be crawled, to show up in the index and to assume their rightful place in the results.  It would seem to mean that personalization is just going to drive that process faster.

Matt: I think it will.  It’s making Black Hat tougher to do.  I think it’s interesting, it was designed primarily to improve the relevance for users but as a side effect, it definitely changes the game a lot more if you’re on the Black hat side of things then if you’re on the white hat side of things

Gord: Matt, I think that wraps things up for me..

Matt: Thanks, that was fun.

Webpronews Video: Who Said What?

I happened to be browsing through Webpronews on the weekend and saw one of their new video news updates. The clips are well produced, professional looking and even have their own attractive newscaster, Nicole Eggers. One I happened to pick, however, left a little to be desired on the accuracy front. As you’re probably aware, I just did a series of interviews with the top usability people at each of the three engines for Search Engine Land and a couple weeks ago I did a recap talking about the differences I saw between each of their philosophical approaches. The blurb on the video appeared to be on the same topic so I decided to give it a watch. If it, Webpronews indicated that search expert Danny Sullivan had talked to each of the three usability people at the engines and had come to the following conclusions:

  • That relevancy was almost a religion for Google
  • Yahoo had a heightened sensitivity to the needs of their advertising community
  • Microsoft was still finding their competitive niche

Huh? That’s exactly what my recap said. They even pulled a few quotes from it and attributed them to Danny. I quickly e-mailed Danny to see if we were doing some kind of weird Cyrano de Bergerac thing but Danny was apparently as out of the loop on this as I was. Anyway a quick e-mail to Webpronews seems to have got it straightened out. They’ve pulled the clip and apparently they’re redoing it.

Not that I mind being mistaken for Danny, but I just hate to be putting words in his mouth. By the way, does anyone else feel like they’re being scolded by Nicole? Again, not that I mind.

Marissa Mayer Interview on Personalization

marissa-mayer-7882_cnet100_620x433Below is the full transcript of the interview with Marissa Mayer on personalization of search results. For commentary, see the Just Behave column on Searchengineland.

Gord: It’s a little more than two weeks ago since Google made the announcement that personalization would become more of a default standard for more users on Google.  Why did you move towards making that call?

Marissa: We’ve had a very impressive suite of personalized products for awhile now: personalized homepage, search history, the personalized webpage and we haven’t had them integrated, which I think has made it somewhat confusing for users. A lot of people didn’t know if they had signed up for search history or personalized search; whether or not it was on.  What we really wanted to do was move to a signed in version of Google and a signed out version of Google.  So if you’re signed in you have access to the personalized home page, the personalized search results and search history.  You know all three of those are working for you when you’re signed in.  And if you’re signed out, meaning that you don’t see an email in the upper right hand corner that personalized search isn’t turned on.  If anything, it’s a cleaning up of the user model, to make it clearer to users what services they’re using them and when they’re using them.

Gord: But some of the criticism actually runs counter to that.  One of the criticisms is that it used to be clearer, as far as the user went, when you were signed in and when you are signed out.  There were more indicators on the Google results page whether you were getting personalized results or not.  Some of those have seemed to disappear, so personalized results have become more of a default now, rather than an option that’s available to the user.

Marissa: If you think about it as default-on when you’re signed in, I think that it’s still as clear on the search results page.  We removed the “turn off the personalized search results” link, but you still see very clearly up in the upper right-hand corner whether or not you’re signed in, your e-mail address appears, and that’s your clue Google has personalized you and that’s why that e-mail address is there.  I do think, based on our user studies and our own usage at Google, that we’ve made the model clearer.  We were actually ended up at the stage with our personalized product earlier this year where, at one point, Eric (Schmidt) asked “am I using personalized search?”  And the team’s answer as to whether or not he was currently using it was so complicated that even he couldn’t follow it.  You’d have to go to “my account”, see whether or not he was signed up for personalized search, make sure that your toggle hadn’t been turned off or on, and there was no way to just glance at the search results page and easily tell whether or not it was invoked.  So now it’s very easy, if you see your username and e-mail address up in the upper left-hand corner, you’re getting personalized results and if you don’t, you’re not.  So effectively there are two parallel universes of Google, per se.  One if you’re signed out where you see the classic homepage and the classic search results and one where you’re signed in, where you get the personalized home page and…you’ll be able to toggle back and forth, of course…and then the personalized search results page and the search history becomes coupled with all that because that’s how we personalize your search.

Gord: So, to sum up, it’s fair to say that really the search experience hasn’t changed that dramatically, it’s just cleaning up the user experience about whether you’re signed in or signed out and that’s been the primary change.

Marissa: That’s right.  Before you could be signed in and be using one of the three products or two of the three products but not all and, of course, because people like to experiment with a new product, they forget whether they signed up for personalized search.  Had they signed up for search history?  This just makes it cleaner.  If you’re signed in you’re using and/or have access to all three, if you’re signed out, you’re on the anonymous version of Google that doesn’t have personalization.

Gord: We can say that it cleans up the user experience because it makes it easier to you know when you’re signed in or signed out, but having done the eye tracking studies, we know that where the e-mail address shows is in a location that’s not prominently scanned as part of the page.  Do the changes mean that more people are going to be looking at personalized search results, just because we’ve made that more of a default opt in and we’ve moved the signals that you’re signed in a little bit out of the scanned area of the page.  Once people fixate on their task they are looking further down the page.  This should mean at a lot more people are looking at personalized search results than previously.

Marissa: Actually, I don’t think it will change the volume of personalized search all that much, not based on what we’ve seen on our logs and usage.  It makes it cleaner to understand whether or not you’re using it and I do think that over time, what it does is it pushes the envelope of search more such that you expect personalized results by default.  And we think that the search engines in the future will become better for a lot of different reasons, but one of the reasons will be that we understand the user better.  And so when we think about how we can advance towards that search engine of the future that we’re building, part of that will be personalization.  I do think that when we look five years out, 10 years out, users will have an expectation of better results.  One of the reasons that they have that expectation is that search engines will have become more personalized.  I think that in the future, working with the search engine that understands something about you will become the expectation.  But you’re right in that we believe that for users that are signed in, who find value in the personalized search results, over time as those users know they are signed in and that there search history is being kept track of, that their search results are being personalized, and they don’t need to look at every single search task to see whether or not they are signed in because that’s what their expectation is and they’re expecting personalized results.  So I do think we won’t see a drastic increase of volume right now of the use of personalized search but that it will hopefully change the user’s disposition over time to become more comfortable that personalization is a benefit for them and it’s something they come to expect.

Gord: There are a number of aspects of that question that I’d like to get into, and leave behind the question of whether you’re signed in or signed out of personalized search, but I have one question before we move on.  We’ve been talking a lot about existing users. The other change was where people were creating a new Google account and they got personalized search and search history by default.  The opt-out box is tucked into an area where most users would go right past it.  The placement of that opt-out box seems to indicate that Google would much rather have people opting into personalized search.

Marissa: I think that falls in with the philosophy that I just outlined. We believe that the search engines of the future will be personalized and that it will offer users better results.  And the way for us to get that benefit to our users is to try and have as many users signed up for personalized search as possible.  And so certainly we’re offering it to all of our users, and we’re going to be reasonably aggressive about getting them to try it out. Of course, we try to make sure they’re well-educated about how to turn it off if that’s what they prefer to do.

Gord: When this announcement came out I saw it as a pretty significant announcement for Google because it lays the foundation for the future.  I would think from Google’s perspective the challenge would be knowing what personalized search could be 5 to 10 years down the road,  what it would mean for the user experience and how do you start adding that incrementally to the user experience in the meantime?  From Google’s side, you have invested in algorithmic work to categorize content online. I would think the challenge would be just as significant to introduce the technology required to disambiguate intent and get to know more about users. You’re not going to hit that out of the park on the first pitch. That’s going to be a continuing trial and error process.  How do you maintain a fairly consistent user experience as you start to introduce personalization without negatively impacting that user experience?

Marissa: I will say that there are a lot of challenges there and a lot of this is something that’s going to be a pragmatic evolution for us.  You have to know that this is not a new development for us. We’ve been working on personalized search now for almost 4 years. It goes back to the Kaltix acquisition. So we’ve been working on it for awhile and our standards are really high.  We only want to offer personalized search if it offers a huge amount of end user benefit.  So we’re very comfortable and confident in the relevance seen from those technologies in order to offer them at all, let alone have them veered more towards the results, as we’re doing today.  We acquired a very talented team in March of 2003 from Kaltix.  It was a group of three students from Stanford doing their Ph.D, headed up by a guy named Sep Kamvar, who is the fellow who cosigned the post with me to the blog. Sep and his team did a lot of PageRank style work at Stanford.  Interestingly enough, one of the papers they produced was on how to compute PageRank faster.  They wrote this paper about how to compute page rank faster and it caused a huge media roil around the web because everyone said there are these students at Stanford who created an even faster version of Google.  Because the press obviously doesn’t understand search engines and thinks that we actually do the PageRank calculation on the fly on each query, as opposed to pre-computing it.  Their advance was actually significant not because it helps you prepare an index faster, which is what the press thought was significant.  Interestingly enough, the reason they were interested in building a faster version of PageRank was because what they wanted to do was be able to build a PageRank for each user.  So, based on seed data on which pages were important to you, and what pages you seemed to visit often, re-computing PageRank values based on that. PageRank as an algorithm is very sensitive to the seed pages.  And so, what they were doing, was that they had figured out a way to sort by host and as a result of sorting by host, be able to compute PageRank in a much more computationally efficient way to make it feasible to compute a PageRank per user, or as a vector of values that are different from the base PageRank.  The reason we were really interested in them was: one, because they really grasped and cogged all of Google’s technology really easily; and, two, because we really felt they were on the cutting edge of how personalization would be done on the web, and they were capable of looking at things like a searcher’s history and their past clicks, their past searches, the websites that matter to them, and ultimately building a vector of PageRank that can be used to enhance the search results.

We acquired them in 2003 and we’ve worked for some time since to outfit our production system to be capable of doing that computation and holding a vector for each user in parallel to the base computation.  We’ve been very responsible in the way that we’ve personalized Search Labs and we also did what we called Site Flavored Search on Labs where you can put a search box on your page and that is geared towards a page of interests that you’ve selected. So if you have a site about baseball you can say you want to base it on these three of your favorite baseball sites and have a search box that has a PageRank that’s veered in that direction for baseball queries.

So, the Kaltix team has been really successful at integrating all these Google technologies and taking this piece of theoretical research and ultimately bringing it to life on the Web.  And as it’s growing stronger and stronger and our confidence around the Kaltix technology grew, we’ve been putting it forward more and more.  We started off on Labs through a sign-up process, then we transitioned it over to Google.com and now we are in effect leaning towards a model where for people who use Google.com and have a Google account, they get personalized search basically by default.  If you look at the historical reviews of the Kaltix work it’s gotten pretty rave reviews.  The users that have noticed it and have been using it for a long time, like Danny (Sullivan), they’ll say that they think it’s one of the biggest advances to relevance that they’ve seen in the past three years.

Gord: So when you the Kaltix technology working over and above the base algorithm, obviously that’s going to be as good as the signals you’re picking up on the individual.  And right now the signals are past sites they visited, perhaps what they put on their personalized homepage and sites that they’ve bookmarked. But obviously the data that you can include to help create that on-the-fly, individual index improves as you get more signals to watch.  In our previous interview you said one thing that was really interesting to you was looking at the context of the task you are engaged in, for example, if you’re composing an e-mail in Gmail. So is contextual relevance another factor to look at.  Are those things that could potentially be rolled into this in the future?

Marissa: I think so.  I think that overall, we really feel that personalized search is something that holds a lot of promise, and we’re not exactly sure of the signals that will yield the best results.  We know that search history, your clicks and your searches together provide a really rich set of signals but it’s possible that some of the other data that Google gathers could also be useful. It’s a matter of understanding how.  There’s an interesting trade off around personalized search for the user which is, as you point out, the more signals that you have and the more data you have about the user, the better it gets.  It’s a hard sell sometimes, we’re asking them to sign up for a service where we begin to collect data in the form of search history yet they don’t see the benefits of that, at least in its fullest form, for some time.  It’s one of those things that we think about and struggle with. And that’s one reason why we’re trying to enter a model where search history and personalized search are, in fact, more expected.  And I should also note that as we look at reading some of the signals across different services we will obviously abide by the posted privacy policies.  So there are certain services where we’ve made it very clear we won’t cross correlate data. For example on Gmail, we’ve made it very clear that we won’t cross correlate that data with searches without being very, very explicit with the end user.  You don’t have to worry about things like that.

Gord: One of the points of concern seems to be how smart will that algorithm get and do we lose control?  For example, when we’re exploring new territory online and we’re trying to find answers we’ve refine our results based on our search experience.  So, at the beginning, we use very generic terms that cast a very wide net and then we narrow our search queries as we go. Somebody said to me, “Well, if we become better searchers, does that decrease the need for personalization?”  Do we lose some control in that?  Do we lose the ability to say “No, I want to see everything, and I will decide how I narrow or filter that query.  I don’t want Google filtering that query on the front end”?

Marissa: I think it really depends on how forcefully we’re putting forth personalization.  And right now we might be very forceful in getting people to sign up to it, or at least more forceful than we were. The actual implementation of personalized search is that as many as two pages of content, that are personalized to you, could be lifted onto the first page and I believe they never displace the first result, in our current substantiation, because that’s a level of relevance that we feel comfortable with.  So right now, at least eight of the results on your first page will be generic, vanilla Google results for that query and only up to two of them will be results from the personalized algorithm.  We’re introducing it in a fairly limited form for exactly the reason that you point out.  And I think if we tend to veer towards a model where there are more results that are personalized, we would have ways of making it clearer: “Do you want to explore this topic as a novice or with the personalization in place?” So the user will be able to toggle in a different filter form.  I think the other thing to remember is, even when personalization happens and lifts those two results onto the page, for most users it happens one out of every five times.  When you think about it, 20% of the queries are much better by doing that, but for 80% of the queries, people are, in fact, exploring topics that are unknown to them and we can tell from their search history that they haven’t searched for anything in this sphere before. There’s no other search like it. They’ve never clicked on any results that are related to this topic, and, as a result, we actually don’t change their query set at all because we know that they need the basic Google results.  The search history is valuable not only because it can help personalize the results but they’re also valuable because we can tell when not to.

Gord: There’s two parts to that: one is the intelligence of the algorithm to know when to push personalization and when not to push personalization, and two, as you said, right now this is only impacting one out of five searches where you may have a couple of new results being introduced into the top 10 as a result of personalization.  But that’s got to be a moving target.  As you become more confident in the technology and that it’s adding to the user experience, personalization will creep higher and higher up the fold and increasingly take over more of the search results page, right?

Marissa: Possibly.  I think that’s one of many things that could possibly happen, and I think that’s a pretty aggressive stance.  I look at our evolution and our foray into personalization, where we’re sitting here three or four years in, with some base technology that several years old already and it still has been very slight in a way that we have it interact with the user experience.  Mostly because we think that base Google is pretty good.  As it becomes more aggressive, certainly I would be pushing for an understanding of the ability of the user to know that these results are, in fact, coming from my personalization and not background and if I want to filter them out and get back to basics, that that would be possible.  One thing that we’ve struggled with is if we should actually mark the results are entering the page as a result of personalization but because team is currently and frequently doing experiments, we didn’t want to settle on a particular model or marker at this exact moment.

Gord: The challenge there is as you roll more personal results into the results page and get feedback from some users that they would want more control over what on the page is personalized and the degree of personalization and introduce more filters or more sophisticated toggles, it complicates the user experience. And as we know, that user experience needs to be very simple. Is it a delicate balance of how much control you give the user versus how much do you impact the 95% of the searches that are just a few seconds in duration and have to be really simple to do?

Marissa: There are two thoughts there.  One, even if we introduce them to filtering on the results page, it wouldn’t be any more complicated than what you had two weeks ago, so we already have that filter.  Two, we put the user first, and people have varying opinions about whether their search results page is too complicated, but the same people who designed that user experience will be the people who will be tackling this for Google, so I think you can expect results of a similar style and direction.

Gord: In the last few weeks, Google has introduced some new functionality, related searches and refine search suggestions, that are appearing at the bottom of the page for a number of searches.  To me that would seem to be a prime area that could be impacted by personalization opportunities that are coming.  As you make suggestions about other queries that you could be using, using that personalization data to refine those. Is that something you’re considering? And how long before personalization starts impacting the ads that are being presented to you on a search results page?

Marissa: Refinement is an interesting but a neophyte technology from our perspective.  We are finally now just beginning to develop some refining technologies that we believe in enough to use on the search results page.  A lot of people have been doing it for a lot longer. When you look at the overall utility, probably 1 to 5% of people will click those query refinements on any given search, where most users, probably more than two thirds of users, end up using one of our results. So in terms of utility and value that is delivered to the end user, the search results themselves and personalizing those are an order of magnitude more impactful then personalizing a query refinement.  So part of it is a question of, it’s such a new technology that we really haven’t looked at how we can make personalization make it work more effectively.  But the other thing is on a “bang for the buck” basis, personalizing those search results get us a lot more.

And as to ads, I think there are some easy ways to personalize ads that we’ve known for some time, but we’ve chosen at this point to focus on personalizing the search results because we wanted to make sure to delivered the end-user value on that, because that’s our focus, before we look at personalizing ads

Gord: So, no immediate plans for the personalization of ads?

Marissa: That’s right

Gord: Thank you so much for your time Marissa.

Webpronews Interview Now Posted

Had a chance to chat with Mike Macdonald from Webpronews at the recent Search Engine Strategies conference in Chicago. The video has just been posted on the site.

We chatted about SEMPO Institute and our latest eye tracking study.

Interview with Shuman Ghosemajumder about Click Fraud

Had a chance to chat with Shuman Ghosemajumder regarding click fraud. Shuman is Google’s point person on the click fraud issue. This follows up on the post Andy Beal made on MarketingPilgrim earlier this week. Most of what we chatted about was in my Search Insider column this week. However, not all of it made it into the column, as there is a cut off which I routinely ignore (thanks to MediaPost editor Phyllis Fine for keeping me in line).

Here’s some tidbits that didn’t make it into the column:

First of all, I wanted to take the media to task for crying the sky is falling around this issue. I know that’s what journalists do, but the portrayal of the click fraud issue has been very one sided to this point. That’s why I wrote the column. I think it’s important we get balancing viewpoints. In the absence of numbers universally regarded as accurate, one has to poll the extremes and guess that the true answer lies somewhere in the middle. Up to this point, all we’ve heard are the negative estimates, and these are based on some studies with methodolgy that’s questionable at best (i.e. the Outsell study)

Secondly, I believe it’s unfair that everyone seems to be taking aim at Google, and to a lesser extent, Yahoo on this issue. I know they’re easy targets, because the targets are so damned big, but when the real numbers finally do come out, I’d bet my 89 Mazda 626 (the car that just won’t die!) that it’s the 2nd and 3rd tier networks that are the hotbeds of click fraud.

I dealt with it briefly in the column, but one of the main sources of misrepresentation seems to be this question of what click fraud is. For me, the definition is pretty simple, fraudulent clicks that leave the advertiser financially impacted. But when it comes to most of the media portrayals, there are a number of clicks that get lumped together under the label “click fraud”, the majority of which don’t meet this definition. And Google’s point of contention with reports of click fraud that come from the media and various 3rd party fraud detection tools comes from this aggregation of questionable numbers. There’s no distinction made between actual fraud, the clicks that cost the advertiser, and attempted fraud, the ones that got caught. And often more benign clicks, i.e. multiple legitimate clicks coming from the same IP address, get mistakenly labelled as click fraud.

Another positive move by Google was the inclusion of invalid clicks in the advertiser’s reporting dashboard. Every move that Google makes towards greater transparency is a very positive one. And the best know Google evangelist for communication, Matt Cutts, indicated so on a blog post. By the way, Shuman also has a blog, where he goes into greater depth on this issue.

I can only imagine how frustrating this must be for Shuman and the Google Click Fraud team. They sit and listen to numbers be bandied about in the 15% plus range, knowing from first hand experience that the real number is likely much much lower (in the column, using assumptions that are probably on the high side, the actual amount of click fraud that an advertiser would have to challenge Google on is less than 0.18%). Yet, their tongues are tied, both by Google’s legal and corporate communications department.

Why is the media targeting click fraud and trying to scare the hell out of advertisers? In no other industry I can think of are reporters more prone to mix and match numbers without regard for accuracy. They do it, and get away with it, because there are no independent and reliable numbers to look at.

There are a number of reasons. Google is in the vanguard of disruptive change agents that are shaking the very ground of marketing. It’s somewhat defensive to look for an Achilles heel, and right now, click fraud seems to fit the bill. Google in particular is boldly stating they want to change everything. That scares people.

Part of it is that there is still a lot of people that would love to see Google be knocked down a few pegs. Much as we rever success, wildly successful companies or individuals generate jealousy and suspicion. Our society gets a nasty little thrill when the mighty fall.

But perhaps the biggest reason is the very strength of search and online marketing: it’s accountability. Nothing else is as measurable. So when something appears to be eating away at the cost effectiveness, we tend to go all forensic on it and analyze the hell out of it. Could you imagine the mainsteam press making a big deal out of a .18% hole in the accountability in television advertising, or radio, or print? Even a 10 to 15% hole? Of course not, because much bigger holes than that are accepted every day as being inherent in the channel. But search and online ad networks are apparently fair game.

Is click fraud happening? Absolutely. And if you switch the lens a bit, there are some sophisticated click fraud operations that are making a killing. In a response to my column, Chris Nielsen had this excellent observation:

The problem is not overt clicking on ads, competitors clicking on ads, or double-clicking on ads. The problem is with large-scale concerted efforts that are massive enough to to have enough variety of IP address, user agents, etc. and pose as “valid” user click activity.

Of course this activity varies some with the bid price of the clicks, but it’s really the old idea of stealing a penny from a million people. If anyone notices, who’s really going to care? The problem is that in some areas, there are hundreds or thousands of people stealing pennys, and it is noticible and it is a problem. The only real indication is the lack of bona fide conversions, and that’s hard to say for sure if it’s fraud or real factors with the marketing or web site.

But it comes down to which lens you look through. Do you look at those looking to profit from click fraud, some of them doing it very well? Or do you look at the scope of the problem over the big picture? The problem I have with the BusinessWeek report is that the reporting is trying to do both at the same time, and you can’t get a clear picture by doing so.

I just wanted to wrap up this post by mentioning some other initiatives on this front that Google is pursuing which didn’t make it into the original column. Obviously they’re working on proprietary techniques to filter out click fraud, but they’re also trying to attack the problem on an industry wide basis as well. They’re working with the IAB Click Measurement working group, in which SEMPO is also involved. And they’re calling for stringent and scientific independent auditing standards, so when we throw around terms like click fraud, we’re all dealing with a common reference framework. By the way, I also asked Shuman about impression fraud. We didn’t go into a lot of depth on the issue, but they feel they’re equally on top of that as well.

A Conversation with Ask’s CEO, Jim Lanzone

First published May 18, 2006 in Mediapost’s Search Insider

Last week, I had the chance to spend some time talking to Ask’s new CEO, Jim Lanzone. The first thing that become very clear is that Lanzone is tired of his company’s being compared to Google, MSN and Yahoo. I immediately slipped into the trap of asking how Ask intends to fight the big G and the two other contenders. It was obviously a question that he has heard all too often in the past. “Let’s begin by resetting the framework for the question,” Lanzone replied. “We don’t want to climb Everest right now. We’re not planning on knocking out Google. Our goal is to take our 20 million users, who are currently using us twice a month, and bump that up to four times a month. That doubles our market share,” he said.

“Search is not a zero sum gain,” Lanzone continued. “Americans use about 3.2 engines a month. We want to get on the list. We want to offer an alternative to Google.” As he pointed out, it’s much easier to post impressive market share growth percentages when you’re starting with a relatively small slice of the pie. “For Google to move the needle even a few points, they have to attract huge numbers of new searches,” he said. “We can achieve huge growth just by getting people to use us a few more times each month.”

Focusing on the User

Ask has come out of the gate strongly since the rebrand and the removal of Jeeves. The focus has been squarely on improving the search experience for its existing user. “We want to put the right tool in the right place at the right time,” said Lanzone. “We want to be waiting for the user when they need us.”

Distancing Ask from the helpful butler has proven itself to be difficult. “People are much more aware of us as Ask Jeeves than Ask. It will take time,” he said. For right or wrong, the efforts of Ask Jeeves to brand themselves as a natural query engine have stuck. A long-time member of Ask’s usability team, Michael Ferguson, pointed out the challenges of trying to change how people search. “For years, we’ve been telling people to ask us a question. We’ve perhaps been a little too successful in encouraging them. I once asked a lady what was the last thing she searched for on (then) Ask Jeeves. She was disappointed in the results she got when she typed in ‘Midnight basketball programs are more successful in LA than in St. Louis. Why is that?'”

I had written some time ago that I believe we’ve become used to truncating our search intent into a few words. Apparently, Ask now agrees. Rather than trying to accommodate natural language, Ask now takes a more standard approach to query construction.

Tools When You Need Them, Where You Need Them

Even the physical layout of Ask’s new home page tried to ease the transition from the once ubiquitous butler. Ferguson said, “We put the new tool palette where Jeeves used to be. It’s about the same height and size, so visually it has the same balance.”

The palette is a big part of the new usability focus of Ask. Lanzone explained, “We have some great tools, and we wanted to move them more upfront for the user. We didn’t want to hide them with tabs, which no one clicks on. With the palette, it’s right there, waiting for them.”

Based on initial numbers, people are using the tools more than ever before. Usage on most of them has doubled, with some, like image and map search, posting far higher gains.

Another change was a rethinking of how to use the real estate of the search results page. “The ads from the right rail are gone,” said Lanzone. “We’ve replaced that with something searchers can use, the ‘Narrow’ and ‘Broaden’ your search suggestions. Rather than 1 or 2 percent click-throughs on ads, we’re getting 30 percent click throughs on those suggestions. We know that search is an iterative process, so why not help make it easier by helping the user get to the right search faster?”

Another cleaned up area of the SERP is the top-sponsored ads. It used to be that organic results were pushed right off the page by far too many sponsored listings. Top-sponsored listings have since been restricted to three, in line with other major engines, with the rest shown at the bottom of the page. This used to be one of my pet peeves with Jeeves, and apparently I had plenty of sympathizers in the Ask usability team. “We were a public company, responding to demand for profits,” Lanzone explains. “A lot of us never agreed with that.” Kudos to parent company IAC for eventually listening to the champions of the user experience.

Not Bolder, Just Better

I wrapped up by telling Lanzone that I believed Ask was in a good position to become the bold innovator in search. Unlike Google and Yahoo, it isn’t solely dependent on a huge revenue stream from search, so it can afford to take some risks in testing new interfaces and developments. He replied, “You know, we’ve never considered what we’ve done, or plan to do, as being bold. It just has to be right for the user. All the changes we’ve made were done because in our testing, it felt right. We’re not chasing technology for its own sake. We have a laser focus on the user experience. We just want to do search right.”

I’m a huge believer in focused strategy and feet on the ground, practical user-centricity. They’re two commodities that are in drastically short supply in the current heated search space. If Ask sticks to its guns, it just may just get a shot at the big dogs in search.