Category Archives: Search Behaviors

The Other Long Tail of Search

smallcoverIn 2006 Wired Editor Chris Anderson released The Long Tail, and suddenly we were finding long tails in everything. The swoop of the power law distribution curve was burned on our consciousness, and search was no exception. Suddenly, the hot new strategy was to move into the long tail of search, those thousands of key phrases that individually may only bring a handful of visitors, but in aggregate, can bring more than the head phrases. At Enquiro, we were no exception. But lately, I haven’t heard as much about long tail campaigns. And I think it’s because our thinking was a little flawed.

One of the key elements of a long tail marketplace is that there can’t be a scarcity bottleneck. Shelf space has to be unlimited, production, distribution, etc. The economics of supporting a long tail have to make sense. There has to be little to no overhead in making a vast selection available to the market. The ideal example is digital entertainment. Once an MP3 is encoded, the only overhead for the distributor is a couple of megabytes of storage capacity.

That’s just not true in search management. It takes time to set up a campaign for a keyword. It takes time to organically optimize for it. There’s significant campaign management overhead associated with search. From a resource perspective, search marketing is expensive, and as anyone who’s tried to recruit a search marketer can tell you, there is no abundance mentality at play here. Sure there’s thousands of keyphrases that may potentially bring one or two visitors a month, and if you add them up, it would be hundreds of thousands of potential leads, but we just can’t move from the head to the tail because we don’t have the time.

Engines tried to open up the bottle neck by offering broad match, but results from broad match campaigns are dramatically less than spectacular. In many cases, we needed to go back to the more granular control that comes with exact match to keep performance levels where we needed them to be.

But in perusing some of our spreadsheets from past studies, and with Anderson’s Long Tail curve fresh in my mind, I found another long tail in search. In several studies, we’ve tracked click throughs by position. Look what happens when you take the most popular click through position, the number one organic spot, and then work down by position:

clickthrough tail

As you can see, we have another long tail. Now, due to the scale of the graph, it looks like the tail goes to zero. This isn’t the case, but by the time we get to the second page of results, we tend to hit percentages that are fractions of 1%. As long tails go, this is skinnier than most, showing the power of position on the first page of results.

But now let’s combine the two long tails. If you take the head being a #1 organic ranking for your most popular phrase, and work down from there, the results get quite dramatic. For the sake of this example, I’ll move down one position for each keyword. Here’s a table showing the numbers:

Keyword Volume Position X Click Through % Clicks
#1 46500 Org 1 26.35 12252
#2 38450 Org 2 13.05 5017
#3 32500 Org 3 10.2 3315
#4 23400 Spon 1 8.75 2047
#5 15750 Org 4 5.35 842
#6 12450 Spon 2 4.125 513
#7 8750 Org 5 3.875 339
#8 5250 Org 6 2.875 150
#9 4325 Side 1 2.5 108
#10 2750 Org 7 1.875 51

Now, let’s plot that on a graph:

clickthrough tail2

Given that search is a resource hog in terms of manpower, I argue that we’d be far better working our way up the tail rather than down it. And by moving up the tail, I mean the click through by position tail. By moving up the page, even a position or two, with relatively popular keywords, you can leverage the compounding effect of the two curves and dramatically improve your campaign performance.

Ask Beginning to Break Through

For quite some time, I’ve been wondering if my user “Spidey-Sense” was wonky. From everything I saw about the Ask 3-D interface, it should have been gaining marketshare. Also, for all my preaching about build a better user experience and you’ll reap the rewards, Ask’s reaping appeared to be a little on the grim side, lingering at about 3.5% of the market. But finally, according to a recent post by Bill Tancer over at Hitwise, my instincts seem to be back on track. Take a look at this graph:

ask

Ask is finally making a move. And their “share of search” has moved up from 3.49% of executed searches in August to 4.32% of searches in October, a bump of 23.7%. That’s huge. Bill wonders if it has anything to do with the ads Ask is running. I suspect it has a lot more to do with a great interface and some user generated buzz that’s beginning to catch some ears. Michael Ferguson and his team did exactly what they needed to do, shake things up by thinking about what users want.

Ask’s strategy has always been to be your first second choice. They don’t ever expect to knock Google out of the lead, but what they want to do is be the place you turn when you find Google just isn’t cutting it. So their move to 3D made a lot of sense. For certain types of searches, notably entertainment or discovery searches, users want something more than Google’s spartan, click and get out interface. They want a stickier, richer, more visual appearance. They want Ask 3D. We found the interface tested pretty well in our recent Search:2010 Eye Tracking study for entertainment based searches.

In fact, Marissa Mayer at Google paid Michael and his team the ultimate compliment when she mentioned the likelihood of Google moving to more of a portal, encyclopedia type format sometime in the future. So..that would make Google more like..Ask!

I will be watching with interest Ask’s marketshare numbers over the next 6 months. Again according to Hitwise, the jump regains all the marketshare they’ve lost in the last year, and puts them a lot closer to the current number 3, Microsoft, who is sitting just 3 and a half points ahead at 7.83%. Microsoft has been on a continuous slide for the past year, dropping 3 full points. Yahoo seems perpetually stuck between 22 and 23%. Google has captured most of the fallout, adding those 3 points to their marketshare numbers. But Google’s .5% drop in the last month seems to have gone directly to Ask, showing that the “First Second Choice” strategy might be paying off. Like Jim Lanzone said to me once, “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,” At the time Lanzone made the comment, Ask was sitting with about 2.5% marketshare. If you look at the table below, Ask has just about hit their goal.

 

Percentage of U.S. Searches Among Leading Search Engine Providers
Domain

Sept-07

Aug-07

Sept-06
http://www.google.com

63.55%

63.98%

60.93%
search.yahoo.com

22.55%

22.87%

22.29%
search.msn.com

7.83%*

7.98%*

10.87%*
http://www.ask.com

4.32%

3.49%

4.28%
Note: Data is based on four week rolling periods (ending 9/29/07, 9/01/07; 9/30/2006) from the Hitwise sample of 10 million US Internet users.

* – includes executed searches on Live.com and MSN Search.
Source: Hitwise

But I don’t think Ask is going to stop there. Within 6 months, you’re going to be reading stories all over the web about how Ask bumped Microsoft out of the #3 spot. It will be David vs Goliath, or in this case, Barry (Diller) vs Bill (Gates). Ask is on a roll, and thanks to Bill Tancer’s revisiting of the numbers, I have regained enough confidence to say, “mark my words”.

Search and the Digital CPG Shelf

First published October 25, 2007 in Mediapost’s Search Insider

Last April, James Lamberti from comScore, Randy Peterson from Procter and Gamble and I (representing SEMPO) grabbed a relatively quiet corner at the New York Hilton to talk about a potential research project. Here was our wish list:

–    A study that tied together online activity to offline purchase behavior
–    A study that identified the impact of search in a category not typically one that would be identified with search marketing
–    A study that would attempt to quantify the leveraged impact of search with brand advocates

Search and CPG: Are You Kidding?

As you can see, these were pretty lofty targets to shoot for. Choosing the product category was done at that table. What is the last category you would think of as generating significant search activity? Consumer packaged goods. After all, aren’t these either replenishment purchases, where we keep buying the same brand over and over, or a non-considered purchase, where we’re not really concerned with doing much research? Why would we need to turn to a search engine to figure out which toothpaste to buy, or which would be the right chips for Sunday’s ball game? We reasoned that CPG had the “Sinatra” Factor going for it: If search can make it here, it can make it anywhere.

To be honest, we really didn’t know what to expect, but comScore, together with a lot of help from Yahoo and Procter and Gamble, managed to come up with a workable study design. SEMPO jumped on board as a co-sponsor and we put the study out in the field. This week, with numbers crunched and charts drawn, the results were released in a study labeled The Digital Shelf. After several months of holding our collective breaths, we were about to see if people had already locked CPG brands into their brains, eliminating the need to search for product information.

Apparently not.

People went online for CPG information — in fact, to a significantly higher degree than even our most optimistic predictions.  Over a 3-month period, comScore recorded over 150 million visits to CPG websites in four categories: Food Products, Personal Care Products, Baby Products and Household Products. Those are numbers no marketer should ignore. But even more significantly, search drove significant portions of that traffic, from 23% of all visits in Household Products to 60% in Baby Products.

It’s not just automotive or travel that drive search traffic. We search for recipes, how to get the stains out of carpets, the most eco-friendly disposable diaper and yes, even the nutritional information for potato chips. We search, a lot!

And our searching sets us apart as a consumer segment. Searchers tend to be more interested in product information, comparing against competitors and what they need to make a purchase decision. Non-searchers are more interested in special offers and coupons.

Searchers spend more, about 20% more  — in all the categories in the study. In fact, in the Baby Care Category alone, people searching for information and eventually purchasing could result in almost $12 billion in sales.

Search = Opportunity

But perhaps the most compelling finding was this: People search because they’re comparing alternatives. This means they’re not locked into a brand. They could very well be your competitor’s customer right now. Non-searchers are more likely to go directly to a site because they do have a brand preference. They’re just looking for a bargain on that brand. The study found that 36% of searchers had recently switched their brand, compared to 29% of non-searchers. And, interestingly, searchers are less motivated by price. Only 27% of searchers switched because of price, compared to 38% of non-searchers.

So, the study delivered on our original wish list, and then some. It showed that search is a significant influencing factor in the most unlikely product category of all, the stuff on your pantry shelf or under the sink in your bathroom. In fact, I have yet to see a study done on any product category where search didn’t blow the doors off the competition in its effectiveness in connecting with customers. So perhaps the biggest question left unanswered by the study is this: Why are all those branding dollars still going everywhere but search?

On Your Search Menu Tonight

First published October 4, 2007 in Mediapost’s Search Insider

This week Yahoo unveiled a new feature. It doesn’t really change the search game that much in terms of competitive functionality. If anything, it’s another case of Yahoo catching up with the competition. But it may have dramatic implications from a user’s point of view. To illustrate that point further I’d like to share a couple of stories with you.

The feature is called Search Assist. You type your query in, and Yahoo provides a list under the query box with a number of possible ways you could complete the query. This follows in the footsteps of Google’s search suggestions in its toolbar. Currently, Google doesn’t offer this functionality within the standard Google query box, at least in North America. Ask also offers this feature.

Because Yahoo is late to the game, the company had the opportunity to up the functionality a little bit. For example, the suggestions that come from Yahoo can include the word you’re typing anywhere in the suggested query phrase. Google uses straight stemming, so the word you’re typing is always at the beginning of the suggested phrases. Yahoo also seems to be pulling from a larger inventory of suggested phrases. The few test queries I did brought back substantially more suggestions than did Google.

It’s not so much the functionality of this feature that intrigues me; it’s how it could affect the way we search. I personally have found that I come to rely on this feature in the Google toolbar more and more. Rather than structuring a complete query in my mind, I type the first few letters of the root word in and see what Google offers me. It leads me to select query phrases that I probably never would have thought of myself.

Some time ago I wrote that contrary to popular belief, we’ve actually become quite adept at paring our queries down to the essential words. It’s not that we don’t know how to launch an advanced query; it’s that most times, we don’t need to. This becomes even truer with search suggestions. All we have to do is think of one word, and the search engine will serve us a menu of potential queries. It reduces the effort required from the searcher, but let me tell you a story about how this might impact a company’s reputation online.

I Wouldn’t Recommend That Choice

Some time ago I got a voicemail from an equity firm. The woman who left a message was brash, a little abrasive and left a rather cryptic message, insisting that I had to phone her right back. Now, since I’m in the search game, getting calls from venture capitalists and investment bankers is nothing really new. But I’d never quite heard this tone from one of these prospecting calls before. So, I did as I usually do in these cases and decided to do a little more research on the search engines to determine whether I was actually going to return this call or not. I did my quick 30-second reputation check.

Normally, I would just type in the name of the firm and see what came up in the top 10 results. Usually, if there’s strong negative content out there, it’s worth paying attention to and it tends to collect enough search equity to break the top 10. This time, I didn’t even have to get as far as the results page. The minute I started typing the company name into my Google toolbar, the suggestions Google was providing me told the entire story: “company” scam, “company” fraud and “company” lawsuits. Of the top eight suggestions, over half of them were negative in nature. Not great odds for success. Needless to say, I never returned the call.

If these search suggestions are going to significantly alter our search patterns, we should be aware of what’s coming up in those suggestions for our branded terms. Type your company name into Yahoo or Google’s toolbar and see what variations are being served to you. Some of them may not be that appetizing.

Would You Prefer Szechuan?

My belief is that users are increasingly going to use this to structure their queries. It moves search one step closer to be coming a true discovery engine. One of the overwhelming characteristics of search user behavior is that we’re basically lazy. We want to expand a minimal amount of effort but in return, we expect a significant degree of relevance. Search suggestions allow us to enter a minimum of keystrokes and the search engine obliges us with a full menu of options.

This brings me to my other story. Earlier this year we did some eye-tracking research on how Chinese citizens interact with the search engines Baidu and Google China. After we released the preliminary results of the study, I had a chance to talk to a Google engineer who worked on the search engine. In China, Google does provide real-time search suggestions right from the query box. The company found that it’s significantly more work to type a query in Mandarin than it is in most Western languages. Using a keyboard for input in China is, at best, a compromise. So Google found that because of the amount of work required to enter a query, the average query length was quite short in China, giving a substantially reduced degree of relevancy. In fact, many Chinese users would type in the bare minimum required and then would scroll to the bottom of the page, where Google showed other suggested queries. Then, the user would just click on one of these links. Hardly the efficient searching behavior Google was shooting for. After introducing real-time search suggestions for the query box, Google found the average length of query increased dramatically and supposedly, so did the level of user satisfaction.

Search query suggestions are just one additional way we’ll see our search behavior change significantly over the next year or two. Little changes, like a list of suggested queries or the inclusion of more types of content in our results pages, will have some profound effects. And when search is the ubiquitous online activity it is, it doesn’t take a very big rock to create some significant and far-reaching ripples.

Personalization Catches the User’s Eye

First published September 13, 2007 in Mediapost’s Search Insider

Last week, I looked at the impact the inclusion of graphics on the search results page might have on user behavior, based on our most recent eye tracking report. This week, we look at the impact that personalization might bring.

One of the biggest hurdles is that personalization, as currently implemented by Google, is a pretty tentative representation of what personalization will become. It only impacts a few listings on a few searches, and the signals driving personalization are limited at this point. Personalization is currently a test bed that Google is working on, but Sep Kamvar and his team have the full weight of Google behind them, so expect some significant advances in a hurry. In fact, my suspicion is that there’s a lot being held in reserve by Google, waiting for user sensitivity around the privacy issue to lessen a bit. We didn’t really expect to see the current flavor of personalization alter user behavior that much, because it’s not really making that much of a difference on the relevancy of the results for most users.

But if we look forward a year or so, it’s safe to assume that personalization would become a more powerful influencer of user behavior. So, for our test, we manually pushed the envelope of personalization a bit. We divided up the study into two separate sessions around one task (an unrestricted opportunity to find out more about the iPhone) and used the click data from the first session to help us personalize the data for the search experience in the second session. We used past sites visited to help us first of all determine what the intent of the user might be (research, looking for news, looking to buy) and secondly to tailor the personalized results to provide the natural next step in their online research. We showed these results in organic positions 3, 4 and 5 on the page, leaving base Google results in the top two organic spots so we could compare.

Stronger Scent

The results were quite interesting. In the nonpersonalized results pages, taken straight from Google (in signed out mode) we saw 18.91% of the time spent looking at the page happened in these three results, 20.57% of the eye fixations happened here, and 15% of the clicks were on Organic listings 3, 4 and 5. The majority of the activity was much further up the page, in the typical top heavy Golden Triangle configuration.

But on our personalized results, participants spent 40.4% of their time on these three results, 40.95% of the fixations were on them, and they captured a full 55.56% of the clicks. Obviously, from the user’s point of view, we did a successful job of connecting intent and content with these listings, providing greater relevance and stronger information scent. We manually accomplished exactly what Google wants to do with the personalization algorithm.

Scanning Heading South

Something else happened that was quite interesting. Last week I shared how the inclusion of a graphic changed our “F” shaped scanning patterns into more of an “E” shape, with the middle arm of the “E” aligned with the graphic. We scan that first, and then scan above and below. When we created our personalized test results pages, we (being unaware of this behavioral variation at the time) coincidentally included a universal graphic result in the number 2 organic position, as this is what we were finding on Google.

When we combined the fact that users started scanning on the graphic, then looked above and below to see where they wanted to scan next with the greater relevance and information scent of the personalized results, we saw a very significant relocation of scanning activity, moving down from the top of the Golden Triangle.

One of the things that distinguished Google in our previous eye tracking comparisons with Yahoo and Microsoft was its success of keeping the majority of scanning activity high on the page, whether those top results were organic or sponsored.

Top of page relevance has been a religion at Google. More aggressive presentation of sponsored ads (Yahoo) or lower quality and relevance thresholds of those ads (Microsoft) meant that on these engines (at least as of early 2006) users scanned deeper and were more likely to move past the top of the page in their quest for the most relevant results. Google always kept scan activity high and to the left.

But ironically, as Google experiments with improving the organic results set, both through the inclusion of universal results and more personalization, their biggest challenge may be in making sure sponsored results aren’t left in the dust. Top of page scanning is ideal user behavior that also happens to offer a big win for advertisers. As results pages are increasingly in flux, it will be important to ensure that scanning doesn’t move too far from the upper left corner, at least as long as we still have a linear, 1 dimensional top to bottom list of results.

An Image Can Change Everything for the Searcher

First published September 6, 2007 in Mediapost’s Search Insider

For the many of you who responded to last week’s column about Nona Yolanda, I just want to take a few seconds to let you know that she passed away the evening of Sept. 3, having fought for five days more than doctors gave her. She was in the presence of her family right until the end. We printed off your comments and well wishes and posted them on the hospital door. It was somewhat surprising but very gratifying for my wife’s family to know that Nona’s story touched hearts around the world. Thank you. – G.H.

The world of the search results page is changing quickly, which means that we’re going to have to apply new rules for user behavior. This week, I’d like to look at some results from a recent eye tracking study we did about how we interact with search when graphic elements start to appear on the page. We also tested for the inclusion of personalized results. There’s a lot of ground to cover, so I’ll start off with Universal Search this week, and cover personalization and the future of search next week.

Warning: Graphic Depictions Ahead

You can’t get much more basic than the search results page we’ve all grown to know in the past decade. The 10 blue organic links and, more recently, the top and side sponsored ads have defined the interface. It’s been all text, ordered in a linear top to bottom format. The only sliver of real estate that saw any variation was the vertical results, sandwiched between top sponsored and top organic. So it was little wonder that we saw a consistent scan pattern emerge, which we labeled the Golden Triangle. It was created by an “F”-shaped scan pattern, where we scanned down the left hand side, looking for information scent, and then scanned across when we found it.

But that design paradigm is in the middle of change. The first and most significant of these will be the inclusion of different types of results on the same page, blended into the main results set. Google’s label is Universal Search, Ask’s is 3D Search and Yahoo’s is Omni Search. Whatever you choose to call it, it defines a whole new ball game for the user.

Starting at the Top…

In the classic pattern, users began at the top left corner because there was no real reason not to. We saw the page, our eyes swung up to the top left and then we started our “F”- shaped scans from there. Therefore, our interactions with the page were very top-heavy. The variable in this was the relevance of the top sponsored ads. If the engine maintained relevance by only showing top sponsored when they were highly relevant (i.e. Google) to the query, we scanned them. If the engine bowed to the pressures of monetization and showed the ads even when they might not be highly relevant to the query (we saw more examples of this on Yahoo and Microsoft) users tended to move down quickly and the Golden Triangle stretched much further down the page. It was a mild form of search banner blindness. The one thing that remained consistent was the upper left starting point.

But things change, at least for now, when you start mixing results into the equation. If the number 2 or 3 organic return is a blended one, with a thumbnail graphic, we assume the different presentation must mean the result is unique in some way. The graphic proves to be a power attractor for the eye, especially if it’s a relevant graphic. It’s information scent that can be immediately “grokked” (to use Jakob Nielsen’s parlance) and this often drew the eye quickly down, making this the new entry point for scanning. This reduces the top to bottom bias (or totally eliminates it), making the blended result the first one scanned. Also, we saw a much more deliberate scanning of this listing.

Give Me an F, Give Me an E…

Another common behavior we identified is the creation of a consideration set, by choosing three or four listings to scan before either choosing the most relevant one or selecting another consideration set. In the pre-blended results set, this consideration set was usually the top three or four results. But in blended results, it’s usually the image result being the first result scanned, and then the results immediately above and below it. Rather than an “F”-shaped scan, this changes the pattern to an “E”-shaped scan, with the middle arm of the “E” focused on the graphic result.

The implications are interesting to consider. The engines and marketers have come to accept the top to bottom behavior as one of the few dominant behavioral characteristics, and it has given us a foundation on which to build our positioning strategy. But if the inclusion of a graphic result suddenly moves the scanning starting point, we have to consider our best user interception opportunities on a case-by-case basis.

Next week, I’ll look at further findings.

Search Engine Results: 2010 – Marissa Mayer Interview

marissa-mayer-7882_cnet100_620x433Just getting back in the groove after SES San Jose. You may have caught some of my sessions or heard we have released a white paper looking at the future of search and with some eye tracking on personalized and universal search results. We don’t have the final version up yet, but it should be available later this week. The sneak preview got rave reviews in SJ.

Anyway, I interviewed a number of influencers in the space, and I’ll be posting the full transcripts here on my blog over the next week. I already posted Jakob Nielsen’s interview. Today I’ll be posting Marissa Mayer’s, who did a keynote at SES SJ. It makes for interesting reading. Also, I’ll be running excerpts and additional commentary on Just Behave on Search Engine Land. The first half ran a couple weeks ago. Look for more (and a more regular blog schedule) coming out over the next few weeks. Summer’s over and it’s back to work.

Here’s my chat with Marissa:

Gord: I guess I have one big question that will probably break out into a few smaller questions.  What I wanted to do for Search Engine Land is speculate on what the search engine results page might look like to the user in three years time.  With some of the emerging things like personalization and universal search results and some the things that are happening with the other engines: Ask with their 3D Search, which is their flavor of Universal, it seems to me that we might be at a point for the first time in a long time the results that we’re seeing may have a significant amount of flux over the next 3 years.  I wanted to talk to a few people in the industry about their thoughts of what we might be seeing 3 years down the road.  So that’s the big over-arching question I’m posing.

Marissa: Sure, Minority Report on search result pages…Well, I’d like to say it’s going to be like that but I think that’s a little further out.  There are some really fascinating technologies that I don’t know if you’ve seen..some work being done by a guy named Jeff Han?

Gord: No.

Marissa: So I ran into Jeff Han both of the past years at TED. Basically he was doing multi-touch before they did it on the iPhone on a giant wall sized screen, so it actually does look a lot like Minority Report. It was this big space where you could interact, you could annotate, you could do all those things.  But let me talk first about what I see happening as some trends that are going to drive change.

One is that we are seeing more and more broadband usage and I think in three years everyone will be on very fast connections, so a lot more to choose from and  a lot more data without taking a large latency hit.  The other thing we’re seeing is different mediums, audio, video.  They used to not work.  If you remember getting back a year ago, everytime you clicked on an audio file or a movie file, it would be, like, ‘thunk’?  It needs a plug in, or “thunk”, it doesn’t work.  Now we’re coming into some standardized formats and players that are either browser or technology independent enough, or are integrated enough that they are actually going to work.  And also we’re seeing users having more and more storage on their end.  And those are the sort of 3 computer science trends that are things that are going to change things.  I also think that people are becoming more and more inclined to annotate and interact with the web. It started with bloggers, and then it moved to mash ups, and now people are really starting to take a lot more ownership over their participation on the web and they want to annotate things, they want to mark it up.

So I think when you add these things together it means there’s a couple of things.  One, we will be able to have much more rich interaction with the search results pages. There might be layers of search results pages: take my results and show them on a map, take my results and show them to me on a timeline.  It’s basically the ability to interact in a really fast way, and take the results you have and see them in a new light.  So I think that that kind of interaction will be possible pretty easily and pretty likely.  I think it will be, hopefully, a layout that’s a little bit less linear and text based, even than our search results today and ultimately uses what I call the ‘sea of whiteness’ more in the middle of the page, and lays out in a more information dense way all the information from videos to audio reels to text, and so on and so forth.  So if you imagine the results page, instead of being long and linear, and having ten results on the page that you can scroll through to having ten very heterogeneous results, where we show each of those results in a form that really suits their medium, and in a more condensed format.  A couple of years ago we did a very interesting experiment here on the UI team where we took three or 4 different designs where the problem was artificially constrained.  It was above the fold Google.  If you needed to say everything that Google needed to say above the fold, how would you lay it out?  And some came in with two columns, but I think two columns is really hard when it was linear and text based.  When you started seeing some diagrams, some video, some news, some charts, you might actually have a page that looks and feels more like an interactive encyclopedia.

Gord: So, we’re almost going from a more linear presentation of results, very text based, to almost more of a portal presentation, but a personalized portal presentation.

Marissa: Right and I think as people, one, are getting more bandwidth and two, as they’re more savvy with how they look at more information, think of it this way, as more of serial access versus random access.  One of my pet peeves is broadcast news, where I really don’t like televised news anymore.  I like newspapers, and I like reading online because when I’m online or with newspapers, I have random access.  I can jump to whatever I’m most interested in.  And when you’re sitting there watching broadcast news you have to take it in the order, at the pace and at the speed that they are feeding it to you.  And yes, they try to make it better by having the little tickers at the bottom, but you can’t just jump in to what you’re interested in.  You can only read one piece of text at a time, and it’s hard to survey and scan and hone in on one type of medium or another when it’s all one medium.  So certainly there is some random access happening with the search results today.  I think as the results formats becomes much more heterogeneous, we’re going to have a more condensed presentation that allows for better random access.  Above the fold being really full of content, some text, some audio, some video, maybe even playing in place, and you see what grabs your attention, and pulls you in.   But it’s almost like random access on the front page of the New York Times, where am I more drawn to the picture, or the chart, or this piece of content down here?  What am I drawn to?

Gord: Right.  If you’re looking at different types of stimuli across the page, I guess what you’re saying is, as long as all that content is relevant to the query you can scan it more efficiently than you could with the standardized text based scanning, linear scanning, that we’re seeing now

Marissa: That’s right.

Gord: Ok.

Marissa: So the eyes follow and they just read and scan in a linear order, where when you start interweaving charts and pictures and text, people’s eyes can jump around more, and they can gravitate towards the medium that they understand best.

Gord: So, this is where Ask is going right now with their 3D search, where it’s broken it into 3 columns and they’re mixing images and text and different things.  So I guess what we’re looking at is taking it to the next extreme, making it a richer, more interactive experience, right?

Marissa: Rather than having three rote columns, it would actually be more organic.

Gord: So more dynamic.  And it mixes and matches the format based on the types of material it’s bringing back.

Marissa: Well, to keep hounding on the analogy of the front page of the New York Times.  It’s not like the New York Times…I mean they have basically the same layout each time, but it’s not like they have a column that only has this kind of content, and if it doesn’t fill the column, too bad.  They have a basic format that they change as it suits the information.

Gord: So in that kind of format, how much control does the user have? How much functionality do you put in the hands of the user?

Marissa: I think that, back to my third point, I think that people will be annotating search results pages and web pages a lot.  They’re going to be rating them, they’re going to be reviewing them.  They’re going to be marking them up, saying  “I want to come back to this one later”.  So we have some remedial forms of this in terms of Notebook now, but I imagine that we’re going to make notes right on the pages later.  People are going to be able to say I want to add a note here; I want to scribble something there, and you’ll be able to do that.  So I think the presentation is going to be largely based on our perceived notion of relevance, which of course leverages the user, in the ways they interact with the page, and look at what they do and that helps inform us as to what we should do.  So there is some UI user interaction, but the majority of user interaction will be on keeping that information and making it consumable in the best possible way.

Gord: Ok, and then if, like you said, if you go one step further, and provide multiple layers, so you could say, ok, plot my search results, if it’s a local search, plot my search results on a map.  There’s different ways to, at the user’s request, present that information, and they can have different layers that they can superimpose them on.

Marissa: So what I’m sort of imagining is that in the first basic search, you’re presented with a really rich general overview page, that interweaves all these different mediums, and on that page you have a few basic controls, so you could say, look, what really matters to me is the time dimension, or what really matters to me is the location dimension.  So do you want to see it on a timeline, do you want to see it on a map?

Gord: Ok, so taking a step further than what you do with your news results, or your blog search results, so you can sort them a couple of different ways, but then taking that and increasing the functionality so it’s a richer experience.

Marissa: It’s a richer experience. What’s nice about timeline and date as we’re currently experimenting with them on Google Experimental is not only do they allow you to sort differently, they allow you to visualize your results differently.  So if you see your results on a map, you can see the loci, so you can see this location is important to this query, and this location is really important to that query.  And when you look at it in time line you can see, “wow, this is a really hot topic for that decade”.  They just help you visualize the nut of information across all the results in these fundamentally different ways that ‘sorts’ kind of get at. But it’s really allowing that richer presentation and that overview of results on the meta level that helps you see it.

Gord: Ok.  I had a chance to talk to Jakob Nielsen about this on Friday, and he doesn’t believe that we’re going to be able to see much of a difference in the search results in 3 years.  He just doesn’t think that that can be accomplished in that time period.  What you’re talking about is a pretty drastic change from what we’re seeing today, and the search results that we’re seeing today haven’t changed that much in the last 10 years, as far as what the user is seeing.  You’re really feeling that this is possible?

Marissa: It’s interesting, you know, I pay am lot of attention to how the results look.  And I do think that change happens slowly over time and that there are little spurts of acceleration.  We at Google certainly saw a little accelerated push during May when we launched Universal Search.  I’m of the view that maybe its 3 years out, maybe it’s 5 years out, maybe it’s 10 years out.  I’m a big subscriber to the slogan that people tend to overestimate the short term and underestimate the long term.  My analogy to this is that when I was 5, I remember watching the Jetson’s and being, like, this rocks!  When I’m thirty there are flying cars!  Right?  And here I am, I’m 32 and we don’t even have a good flying car prototype, and yet the world has totally changed in ways that nobody expected because of the internet and computing.  In ways that in the 1980s no one even saw it coming.  Because personal computers were barely out, let alone the internet.  It’s interesting.  We do our off site in August. I do an offsite with my team where we do Google two years out. There it’s really interesting to see how people think about it.  I take all the prime members on my team, so they’re the senior engineers, and everybody has homework.  They have to do a homepage and a results page of Google, and this year it’ll be Google 2009.

Gord: Oh Cool!

Marissa: Six months out, it’s really easy because if we’re working on it, because if it’s going to launch in 6 months and it’s big enough that you would notice, we’re working on it right now and we know it’s coming.  And five years or ten years out we start getting into the bigger picture things like what I’m talking to you about.  When the little precursors that get us ready for those advances happen between now and then that’s what’s shifting.   So I’m giving you the big picture so you can start understanding what some of the mini steps that might happen in the next 3 years, to get us ready for that, would be.  The two to three year timeframe is painful. Everybody at my offsite said, “this timeframe sucks!” So it’s just far enough out that we don’t have great visibility, will mobile devices be something that’s a really big new factor in three years?  Maybe, maybe not.  Some of the things are making fast progress now may even take a big leap, right, like it was from 1994 to 97 on the internet.  Or if you think about G-mail and Maps, like AJAX applications..you wouldn’t have foreseen those in 2002 or 2003.  So, two or three years is a really painful time frame because some things are radically different, but probably in different ways than you would expect.  You have very low visibility in our industry to that time frame.  So I actually find it easier to talk about the six month timeframe, or the ten year timeframe.  So I’m giving you the ten year picture knowing that it’s not like the unveiling of a statue, where you can just take the sheet, snatch it off and go, “Voila there it is”.  If you look at the changes we’ve made over time at Google search they’ve always been “getting this ready, getting this ready”.  So the changes are very slow and feel like they’re very incremental.  But then you look at them in summation over 18 months or two years, you’re like, “you know, nothing felt really big along the way, but they are fundamentally different today”.

Gord: One last question.  So we’re looking at this much richer search experience where it’s more dynamic and fluid and there are different types of content being presented on the page.  Does advertising or the marketing message get mixed into that overall bucket, and does this open the door to significantly different types of presentation of the advertising message on the search results page?

Marissa: I think that there will be different types of advertising on the search results page.  As you know, my theory is always that the ads should match the search results.  So if you have text results, you have text ads, and if you have image results, you have image ads.  So as the page becomes richer, the ads also need to become richer, just so that they look alive and match the page.  That said, trust is a fundamental premise of search.  Search is a learning activity.  You think of Google and Ask and these other search engines as teachers.  As an end user the only reason learning and teaching works, the only way it works, is when you trust your teacher.  You know you’re getting the best information because it’s the best information, not because they have an agenda to mislead you or to make more money or to push you somewhere because of their own agenda.  So while I do think the ads will look different, they will look different in format, or they may look different in placement, I think our commitment to calling out very strongly where we have a monetary incentive and we may be biased will remain.  Our one promise on our search results page, and I think that will stand, is that we clearly mark the ads.  It’s very important to us that the users know what the ads are because it’s the disclosure of that bias, that ultimately builds the trust which is paramount to search

Gord: Ok.  Great to see you’re a keynote at San Jose in August.

Marissa: Should be fun.  This whole topic has me kind of jazzed up so maybe I’ll talk about that.

Search Engine Results: 2010 – Interview with Danny Sullivan

Danny-SullivanHere’s another in the series of the Search:2010 transcripts, this one of my chat with Search Engine Land Editor Danny Sullivan:

Gord: The big question that I’m asking is how much change are we going to see on the search engine results page over the next three years.  What impact are things like universal search and personalization and some of the other things we’re seeing come out, how much of that is going to impact the actual interface the user is going to see.  Maybe let’s just start there.

Danny: I love the whole series to begin with because then I thought, Gosh, I never really sat down and tried to plot out how I would do it, and I wish I had had the time to do that before we talked (laughs).  But it would be nice to have a contest or something for the people who are in the space to say I think this is the way we should do it or where it should go.
But the thing at the top of my head that I expect or I assume that we’re going to get is… I think they’re going to get a lot more intelligent at giving you more from a particular database when they know you’re doing a specific a kind of search.  It’s not necessarily an interface change, but then again it is.  This is the thing I talked about when I was saying about when the London Car Bombing attempts happened, and I’m searching for “London Bombings”.  When you see a spike in certain words you ought to know that there’s a reason behind that spike.  It’s going to be news driven probably, so why are you giving me 10 search results? Why don’t you give me 10 news results?  And saying I’ve also got stuff from across the web, or I’ve got other things that are showing up in that regard.  And that hasn’t changed.  I‘d like to see them get that.   I’d like to see them figure out some intelligent manner to maybe get to that point.  Part of what could come along with that too is that as we start displaying more vertical results the search interface itself could change.  So I think the most dramatic change in how we present search results, really, has come off of local.  And people go “wow, these maps are really cool!” Well of course they’re really cool, they’re presenting information on a map which makes sense when we’re talking about local information.  You want things displayed in that kind of manner.  It doesn’t make sense to take all web search results and put them on a map. You could do it, but it doesn’t communicate additional information for you that’s probably irrelevant and that needs to be presented in a visual manner.  If you think about the other kinds of search that you tend to do, Blog search for instance, it may be that there’s going to be a more chronological display. We saw them do with news archive where they would do a search and they would tell you this happened within these years at this time.  Right now when I do a Google blog search, by default it shows me ‘most relevant’.  But sometimes I want to know what the most recent thing is, and what’s the most recent thing that’s also the most relevant thing right? So perhaps when I do a Search, a Google blog search, I can see something running down the left hand side that says “last hour” and within the last hour you show me the most relevant things in the last hour, the last 4 hours, and then the last day.  And you could present it that way, almost sort of a timeline metaphor. I’m sure there are probably things you could do with shading and other stuff to go along with that.  Image search…Live has done some interesting things now where they’ve made it much less textual, and much more stuff that you’re hovering over, that you can interact with it in that regard.  An I don’t know, it might be that with book search and those other kinds of things that there’ll be other kinds of metaphors that come into place that you can do when you know you are going to present most of the information just from those sorts of resources.  With Video search… I think we’ve already seen a lot of the thing with video search is just giving you the display and being able to play the videos directly.  Rather than having to leave the site because it just doesn’t make sense to have to leave the site in that regard.

Gord: When I was talking to Marissa, she saw a lot more mash ups with search functionality, and you talked about having maps and that with local search making sense, but its almost like you take the search functionality and you layer that over different types of interfaces that make sense, given the type of information your interacting with.

Danny: Right.

Gord: One thing I talked about with a few different people is ‘how much functionality do you put in the hands of the user?’ how much needs to be transparent? How hard are we willing to work with a page of search results?

Danny: By default, not a lot, you know if you’re just doing a general search, I don’t think that putting a whole lot of functionality is going to help you. You could put a lot of options there but historically we haven’t seen people use those things, and I think that’s because they just want to do their searches. They want you to just naturally get the right kind of information that’s there and a lot of the time if they give you that direct answer you don’t need to do a lot of manipulation.  It’s a different thing I think when you get into some very vertical, very task orientated kinds of searches, where you’re saying, ‘I don’t just need the quick answer, I don’t just need to browse and see all the things that are out there, but actually, I’m trying to drill down on this subject in a particular way’.  And local tends to be a great example. ‘Now you’ve given me all the results that match the zip code, but really I would like to narrow it down into a neighborhood, so how can I do that?’  Or a shopping search.  ‘I have a lot of results but now I want to buy something, so now I need to know who has it in inventory? Now I really need to know who has it cheapest? And I need to know who’s the most trusted merchant?’ Then I think the searcher is going to be willing to do more work on the search and make use of more of the options that you give to them.

Gord: Like you say, if you’re putting users directly into an experience where they are closer to the information that they were looking for, there’s probably a greater likelihood that they’re willing to meet you half way, by doing a little extra work to refine that if you give them tools that are appropriate to the types of results they are seeing.  So if it’s shopping search, filtering that by price, or by brand.  That’s common functionality with a shopping search engine and maybe we’ll see that get in to some of the other verticals. But I guess the big question is, in the next three years are the major engines going to gain enough confidence that they’ll be providing a deeper vertical experience as the default, rather than as an invisible tab or a visible tab.

Danny: I still tend to think that the way that they are going to give a deeper vertical experience is the visible tab idea, which is you know, that you are not going to be overtly asked to do it, it is just going to do it for you, and then give you options to get out of it, if it was the wrong choice. So, both Ask, and Google, which are getting all the attention right now, for universal search, you know, blended search if you wanna find a generic term for it that, doesn’t favor one service over the other.  The other term is federated search and I’ve always hated that because it always felt like something from that, you know, came out of the Star Trek Enterprise (laugh). No, I want Klingon search! (laugh) I think that in both of those cases you do the search and the default still is web.  And Ask will say, over here on the side we have some other results. Yes, universal search is inserting an item here or an item there but in most of the cases it still looks like web search, right? They still, really feel like OneBoxes. I haven’t had a universal search happen to me yet that I’ve come along and I’ve thought ‘that really was something I couldn’t have got just from searching the web’ except when I’ve gotten a map.  That’s come in when they’ve shown the map, and that is that kind of dramatic change, and I think at some point they will get to that point, that kind of dramatic change where you just search for “plumbers” and a zip code.  I’m so confident of it I’m just going to give you Google local. I’m not just going to insert a map and give you 7 more web listings that are down there. I’m going to give you a whole bunch of listings and I’m going to change the whole interface on you and if you’re going ‘well, this isn’t what I want’, then I’m going to be able to give you some options if you want to escape out of it.  I like what Ask does, in the sense that it’s easy to escape out of that thing because you just look off to the side and there’s web search over here, there’s other stuff over there.  I think it’s harder for Google to do that when they try to blend it all together. The difficulty remains as to whether people will actually notice that stuff off to the side, and make use of it.

Gord: That was actually something that Jacob Nielsen brought up. He said the whole paradigm of the linear scan down the page is such a dominant user behavior, that we’ve got so used to, you know engines like Ask can experiment with a different layout where they’re going two dimensional, but will the users be able to scan that efficiently?

Danny: I’ve been using this Boeing versus Airbus analogy when I’m trying to explain to people the differences between what Google is doing and what Ask is doing.  Boeing is going, ‘Well, we’ll build small fast energy-efficient jets’ and Airbus is saying ‘We’ll build big huge jets, and we’ll move more people so you’ll be able to do less flights’.  And when I look at the blended search, Google’s approach is, well, we’ve got to stay linear, we’ve got to keep it all in there. That’s where people are expecting the stuff and so we’re going to go that way.  Ask’s approach is we’re going to be putting it all over the place on the page and we’ve got this split, really nice interface.  And I agree with them. And of course Walt Mossberg wrote that review where he said ‘oh they’re so much nicer, they look so much cleaner’, and that’s great, except that he’s a sophisticated person, I’m a sophisticated person, you’re a sophisticated person, we search all the time.  We look at that sort of stuff. A typical person might just ignore it; it might continue to be eye candy that they don’t even notice. And that is the big huge gamble that is going on between these two sorts of players and then, yet again, it might not be a gamble because when you talk to Jim Lanzone, he says ‘My testing tells me this is what our people do’. Well, his people might be different from the Google people. Google has got a lot more new people that come over there that are like, ‘I just want to do a search, show me some things, where’s the text links? I’m done’. So I tend to look perhaps more kindly on what Google is doing, than some people who try to measure them up against Ask because I understand that they deal with a lot more people than Ask, and they have to be much more conservative than what Ask is doing.  And I think that what’s going to happen is those two are going to approach closer together.  The advantage, of course, Jim has over at Ask, is that he doesn’t have to put ads in that column so he’s got a whole column he can make use of, and it is useful, and it is a nice sort of place to tuck it in there. If you really want to talk about search interfaces, what will be really fun to envision is what happens when Ajax starts coming along and doing other things. Can I start putting the sponsored search results where they are hovering above other results? Is there other issues that come with that?  There may be some confusion as to why I’m getting this and why I’m getting that. Can I pop up a map as I hover over a result? I could deliver you a standard set of search results and I can also deliver you local results on top of a particular type of picture.  If I move my mouse along it, I could show you a preview of what you get in local and you might go “Oh wow, there’s a whole map there”. I want to jump off in that direction.  That would be really fun to see that type of stuff come along there, but I’m just not seeing anything come out of it.  What we typically have had when people have played with the interface is, these really WYSIWYG things like, ‘well we’ll fly you though the results, or we’ll group them’.  None of which is really something that you’d need, that added to the choices, “do I want to go vertical, do I not want to go vertical?”

Gord: When we start talking about the fact that the search results page could be a lot more dynamic and interactive, of course the big question is what does that do for monetization of the page?  One of the things that Jakob (Nielsen) talked about was banner blindness.  Do people start cutting out sections of the page?  We talked a little about that.  How do you make sure that the advertising doesn’t get lost on the page when there’s just a lot more visual information in there to assimilate?

Danny: Well I think a variety of things that are going to start happening there.  For example, Google doesn’t do paid inclusion, right, but Google has partnerships with YouTube and they have these channels, and they’re going to be sharing revenue from these channels with other people. So when they start including that stuff up, perhaps they are getting paid off of that.  They didn’t pay to put it in the index but, because they are better able to promote their video channels, more people are going over there, and they’re making money off of that as a destination.  So in some ways, they can afford to have their video results start becoming more relevant because they don’t have to worry about if you didn’t click on the ad from the initial search result, they sort of lost you.  In terms of how the other ads might go, I guess the concern might be if the natural results are getting better and better why would anyone click on the ads anyway?  Maybe people will reassess the paid results and some people will come through and say that paid search results are a form of search data base as well.  So we’re going to call them classifieds or we’re going to call them ads, we’re going to move them right into the linear display.  You know there’ll be issues, because at least in the US, you have the FCC guidelines that say that you should really keep them segregated.  So if you don’t highlight them or blend them in some way, you might run into some regulatory problems.  But then again, maybe those rules might start to change as the search innovation starts to change, and go with it from there.  I don’t know, the search engines might come up with other things.  You know we’re getting toolbars that are appearing more on all of our things. Google might start thinking, ‘Well, let’s put ads back onto that toolbar’.  We used to have those sorts of things, and everyone seems to catch on, but they might come back, and that might be another way that some of the players, especially somebody like Google, might make money beyond just putting the ad on the search result page.

Gord: In the next three years, are we going to get to the point where search starts to become less of a destination activity like the way it is now, and the functionality  sits underneath more of Web 2.0 or semantic web or whatever you want to call it.  It almost becomes a mash up of functionality that underlies other types of sites. Are we going to stop going to a Google or a Yahoo as much to launch a distinct search as we do now?

Danny: You know people have been saying that for at least 3 or 4 years now, especially with Microsoft. ‘Oh you’re not even going to go there, you’re going to do it from your desktop.’  Vista, which I have yet to actually use.  I’ve got the laptop and I’m about to start playing with it! Apparently, it’s supposed to be even more integrated than it was with XP.  But I still tend to think, you know what? We do stuff in our browsers.  I know widgets are growing and I know there’s more stuff that’s just drawing stuff into your computer as well, but we still tend to do stuff in our browser.  I still see search as something where I’m going to go to a search engine and do the search.  With the exception of toolbars. I think we’re going to do a lot more searching through toolbars.  Tool bars are everywhere; it’s really rare for me to start a search where I’m actually not doing it from the toolbar.  I just have a toolbar that sits up there, and I don’t need to be at the search engine itself.  But I still want the results displayed in my browser.  Because I think most of the stuff I’m going to have to deal with is going to be in my browser as well.  So it doesn’t really help to be able to search from Microsoft Word, right?  Because I don’t want all these sites in a little window within Word. I’m probably going to have to read what they say, so I’m probably going to have to go there.  I think that will change though if I have a media player, then I think it makes much more sense for me, and you can already do this with some media players, where you can do searches, and have the results flow back in.  iTunes is a classic example. iTunes is basically a music search engine.  Sure, it’s limited to the music and the podcasts that are within iTunes, but it doesn’t really make any sense for me to go to the Apple website. Although, interestingly, here’s an example where Apple is just a terrible failure.  They’ve got all this stuff out there, they’ve got stuff that perhaps you might be interested in even if you don’t use their software and there’s just no way to get to it on the web.  The last time I looked you really had to do the searches in iTunes.  So they’re missing out on being a destination for those people who say ‘I’m not going to use iTunes’  or ‘I don’t have iTunes’ or ‘I’m on a different version.’ I don’t know if you’ve downloaded it recently but it takes forever and it’s just a pain.

Gord: I think that covers off the main questions I wanted to cover off in this.  Is there anything else as far as search in the next three years that you wanted to comment on?

Danny: You know, it’s hard because if you’d asked me that three years ago, would I have told you, ‘watch for the growth of verticals and watch for the growth of blended search’, (laughs) right?  I’ve been thinking really hard because, I’m like, ‘Gosh, now what am I going to talk about because they’re doing both of those things’. I think personalized search is going to continue to get strong.  I do think that Google is onto something with their personalized search results.  I don’t think that they’re going to cause you to be in an Amazon situation where you’re continuing to be recommended stuff you’re no longer interested in.  I think that people are misunderstanding how sophisticated it can be.  I think that the next big trend is that, ironically from what I just said to you, search is going to start jumping into devices.  And everything is going to have a search box.  But it will be appropriate.  My iPod itself will have a search capability within it.  And the iPhone, to some degree, maybe is going to be that look at how it’s happening already. But I’ll be able to search, access, and get information appropriate to that device within it.  Windows Media Center, when I first got that in 2005, I said, this is amazing, because it’s basically got TV search built into it.  I do the search and then of course, it allows me to subscribe to the program, and records the program, and knows when the next ones are coming up.  And it makes so much more sense for that search to be in that device than it did for me to have it elsewhere.  I use it all the time, when I want to know when a programs on, I don’t have to find where the TV listings are on the web, I just walk over to my computer and do a search from within the Media Center player.  So I think we’re going to have many more devices that are internet enabled, and there’s going to be reasons why you want to do searches with them, to find stuff for them in particular.  That’s going to be the new future of search and search growth will come into it.  And in terms of what that means to the search marketer, I think it’s going to be crucial to understand that these are going to be new growth areas, because those searches when they start are going to be fairly rudimentary. It’s going to be back in the days of, OK, they’re probably going to be driven off of meta data, so you got to make sure you have your title, and your description and making sure the item that your searching for is relevant.

Gord: So obviously all that leads itself to the question of mobile search, and will mobile search be more useful by 2010?

Danny: Sure, but it’s going to be more useful because it’s not going to be mobile search.  It’s just the device is going to catch up and be more desktop-like.  I have a Windows mobile phone at the moment, and I have downloaded some of the applets like Live Search and Google Maps, and those can be handy for me to use, but for the most part, if I want to do a search, I fire up the web browser, I look for what I’m looking for, the screen is fairly large, and I can see what I wanted to find.  And I think that you’re going to find that the devices are going to continue to be small and yet gain larger screens, and have the ability for you to better do direct input. So if you want to do search, you can do a search. It’s not like you’re going to need to have to have something that’s designed for the mobile device that only shows mobile pages.  I think that’s going to change.  You’re going to have some mobile devices that are specifically not going to be able to do that and those people in the end are going to find that no one is going to be trying to support you.

Gord: Thanks Danny.

Breaking “Auction Order” Explained

One of the things that raised eyebrows in my interview with Diane Tang and Nick Fox was the following section regarding how Google determines which ads rank first and climb into the all important top sponsored locations:

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.

This seemed a little odd to me in the interview and I made a note to ask further about that, but what can I say, I forgot and went on to other things. But when the article got posted on Searchengineland, Danny jumped on it at Sphinn

“Seriously? I mean, it’s not an auction. If it were an auction, highest amount would win. They break it all the time by factoring in clickrate, quality score, etc. Not saying that’s bad, but it’s not an auction.”

This reminded me to follow up with Nick and Diane. Diana Adair, on the Google PR team, responded with this clarification:

We wanted to follow up with you regarding your question below.  We wanted to clarify that we rank ads based on both quality score and by bid.  Auction order, therefore, is based on the combination of both of those factors.  So that means that it’s entirely possible that an ad with a lower bid could rank higher than an ad with a higher bid if the quality score for the less expensive ad is high enough.

So, it seems it’s the use of the word “auction” that’s throwing everyone off here. Google’s use of the term includes ad quality. The rest of the world thinks of an auction as somewhere where the highest bid (exclusively) determines the winner. Otherwise, like Danny said, “it’s not an auction”. So, with that interpretation, I then assume that Nick and Diane’s (which sounds vaguely like a title of a John Mellencamp song) comment means that Google won’t arbitrarily hijack these positions for other types of packages which may include presence on the SERP, as in the current Bourne Ultimatum promotion.

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.