Two Microsoft papers being presented at this week’s SIGIR’10 conference in Geneva, Switzerland explore the topics of Search Trails – The pages that a searcher travels through after performing a search for a query before reaching a final destination page.
The idea of delivering searchers to a final destination page, a page where previous searchers for a specific query often end up at before they either stop searching, or changed the focus of their search, is something that Microsoft has explored in the past.
I wrote about a patent filing from Microsoft a couple of years ago which explored how user behavior signals, such as how searchers browsed through pages to find information might be used to rerank search results. The post, Search Trails: Destinations, Interactive Hubs, and Way Stations, took a look at how search trails – the pages browsed between an initial query and a final page visited, might offer useful query suggestions to searchers as well.
That patent filing, and the 2007 SIGIR best paper, Studying the Use of Popular Destinations to Enhance Web Search Interaction (pdf) by Ryen W. White, Mikhail Bilenko, and Silviu Cucerzan, focused more upon the final destination pages found than the pages visited along the way. Ryen White is listed as a co-author in the earlier papers and patent filing on search trails, and he is one of the authors listed on the papers presented this week in Switzerland as well.
It looks as though those intermediary pages may have some value as well, and the idea of including those somehow within search results may be worth exploring.
What route do searchers follow to get to a final destination page, and how important are the pages along the way? Might other searchers with similar information needs and situational tasks to fulfill benefit from a search engine showing the search trails that other follow? How would those trails best be shown in search results?
The papers are:
- Assessing the Scenic Route: Measuring the Value of Search Trails in Web Logs (pdf) by Ryen W. White and Jeff Huang
- Studying Trailfinding Algorithms for Enhanced Web Search (pdf) by Ryen W. White, Jeff Huang and Adish Singla
The authors of the papers tell us that ranking documents for specific keywords may be an easier task than helping people who have more complex informational needs. Those needs can include learning about a topic that someone doesn’t know much about, and may want to become better informed.
Some of the search behaviors cited in the papers about how people might start with one search, and travel through a number of pages after their initial query describe patterns used to meet those needs under names such as information foraging, berry picking, and orienteering.
If you’re not familiar with those concepts, they are definitely worth exploring if you’re interested in learning some theories behind how people perform complex searches. The following are cited in the two recent Microsoft pages:
- The Design of Browsing and Berrypicking Techniques for the Online Search Interface
- Information Foraging
- Orienteering in an Information Landscape: How Information Seekers Get From Here to There
Conclusion
Many search related papers about searchers’ behaviors focus upon query sessions – where searchers may start out with one query term or phrase and possibly perform additional searches adding words to make their search more specific, or removing words to make it more general, correcting spelling mistakes, or even switching over to related terms.
Some query session refinements include people adding geographic terms to their queries, which may indicate to a search engine that specific queries have geographic intents behind them. If those refinements happen frequently enough, they may trigger maps showing up in search results. Other query session refinements may help power some of the “Did you mean” type suggestions that you sometimes see when you search.
Looking at Search Trails may provide a whole different range of searcher behavior type information. By studying the pages that people travel down, from their selected page amongst search results to a final destination page, there may be information that can be related to that initial query that just isn’t captured by looking at query sessions and refinements alone.
Will Microsoft start showing some search trails that are often followed by searchers for specific queries in their search results?
It’s a possibility. It might be an interesting addition to the search results we see today, and could benefit people who don’t know too much about a specific topic, but are interested in exploring it more fully than someone else who may just be looking for a quick and simple answer in their search results.
I believe that search trails are important in a sense that somehow every page that a user gets to provides a clue to reach the final destination. It would really be nice if search trails are included in the search results because it will the offer a wider perspective and a whole lot of possibilities.
Always a nice job covering the latest in search research, Bill.
I’m not sure if this counts as search trails, but I’ve noticed Internet Explorer 8 and Chrome are recording the URLs I go to by sending them to a Microsoft server. (you can monitor this using a network monitoring tool). Is there any SEO opportunity in using browsers made by search engine companies to navigate around my clients’ websites that I am trying to increase the ranking of? Is there any research about this?
PS: I think it’s generally good form to mention the authors when all the papers are written by the same people (ryen w. white from microsoft research etc,)
Hi Andrew,
I don’t know that it might be beneficial to show search trails for every query, but for some it could potentially be really helpful for other searchers. I’d like to see them – especially for topics that I might not know too much about.
Hi Casey,
Thanks.
I’m not sure that there’s an “SEO benefit” in that I imagine the more people who follow similar trails, the more confidence a search engine might have in presenting information about those trails, and in possibly using them when considering things like query refinements or query suggestions. Having one, or a handful of people following the same or similar trails might not have much of an impact. Having lots of people following the same trails, via something like Mechanical Turk, might present some odd statistics about many people performing almost identical browsing patterns over a short period of time, and would probably stand out. A system like this should be designed to be robust enough to avoid being manipulated in that kind of manner.
I’ll probably be digging into more research about query trails sometime in the future for a post – it seems like a good idea to try to write something about different methods of collecting user-behavior data such as mining query sessions and following search trails.
You’re right about including authors for the papers – I had intended to mention that Ryen White was one of the authors common to all of the papers and the patent – I usually include the names of inventors when I write posts about patents, and if I find related whitepapers which share an author, I try to mention that as well. I think it is helpful to mention that relationship. Thanks, again.
I found this article by searching “Assessing the Scenic Route: Measuring the Value of Search Trails in Web Logs best paper award”. I was just at the SIGIR banquet and this paper won Best Paper Award. (you heard it here first! check here in a few days if you don’t believe me). that’s 2 SIGIR best paper awards (including the one from 2007) about search trails by ryen white and his intern jeff huang. I guess this means the ACM SIGIR community also thinks search trails is an important topic.
Hi AJ,
Thank you very much for the update. This was a pretty interesting paper, and I’d love to see search trails integrated into search results. I haven’t had a chance to read through the other papers that were considered for Best Paper yet, but the Scenic Route paper was very well done.
I’m considering attending the HCIR Workshop in New Brunswick, NJ next month, which has Ryen White as one of the Chairs.
Perhaps not possible for Microsoft naysayers. However, recognizing that software is only as smart as the people writing the code, it certainly CAN be created and done better than it is today. BTW: Google does NOT do contextual search. As a matter of fact their core algorithm hasn’t changed much since inception.
Nice writeup Bill. I bet this could potentially be the biggest SERP shakeup since throwing bounce rate into the equation. I do wonder though. How will SE’s effectively determine what the final destination page is. Will they assume is was the last page a user visited before they changed the focus of their search as you mentioned? If you get any more information on this definitely post on it. In my opinion this will definitely reduce SPAM sites.
Excellent post Bill and many thanks for the historic perspective. We too often forget that the problem of finding information has been with us for quite some time. I am suspect that the search trails theory will lead us to better search results. All too often, searchers are flailing around and their trails will likely support this. Or, searchers are easily distracted by some other tidbit that takes them far afield of their original search. They may return later and that could be in a completely different session. Where I see the true benefit of search trails is the serendipitous discovery of new information. Reproducing another searcher’s path can bring me to information that I did not know I needed or was interested in. In a sense, it is “berry picking” by proxy.
Every time I read one of your articles about Microsoft, I can’t help but wonder about whether Bing will be successful in taking away market share from Google or not. Google, is the undisputed heavyweight champion in the area of search, but Microsoft seems to be doing some interesting research that may improve the accuracy of its search results. I don’t know if I’d personally make the switch or not .. Bing will have to come a long way.
Interesting article, as a web designer, looking at search trails and search engine submissions, being able to see where traffic is coming in from would be of an interest to me and would help SEO.
Excellent Article, keep up the good work.
Interesting article here.Keep up the good work..Thanks
Hi Alex,
Ask any search engineer, whether at Google or Yahoo or Microsoft, how far along they’ve gotten in perfecting search, and I’m sure that most would say that it is either in its infancy, or not too much further along than that.
Google’s algorithms, from crawling to indexing to displaying results, have changed significantly in the dozen or so years since they’ve launched. Google has published a number of patent filings on desktop search which show that contextual search is an area that they spent considerably amount of effort upon.
Hi Mark,
I’m not really sold on the idea that search engines consider bounce rate as part of ranking pages. There are many pages that fulfill informational needs of searchers without requiring that searchers visit other pages on a web site. Dwell time might be a little better signal, but even it has some limitations.
Looking at search trails could potentially be not only a great sources of information about query suggestions that could be presented to searchers, but those trails themselves could be very useful to people exploring a topic that they don’t know much about.
If you visit my links above to “Search Trails: Destinations, Interactive Hubs, and Way Stations,” and “Studying the Use of Popular Destinations to Enhance Web Search Interaction ,” there’s more information on that post, and in that Microsoft paper about how final destination pages might be selected.
Hi Marianne,
Thank you. I don’t see too many discussions about berrypicking, information foraging, and orienteering, so I thought those were worth sharing. If Microsoft decides to start sharing search trails for queries, they will be adding useful information that potentially increases the “discovery” element of search. Hopefully they will.
Hi Jonathan,
Papers like the recent ones on search trails show that Microsoft is will to take some chances, and offer something different in a way that makes it more likely that they can compete with Google. Hopefully they will take a bigger chance and start offering that kind of search trail information on search results.
Hi Jason and Kathy,
Thank you.
I would think that search engines already do this and have been collecting information on searches and phrases and the website destinations of users for years. haven’t they? I mean when you have access to so many searches world wide how can you not be collecting relevant data and user behavior? Reading Microsoft paper tomorrow…Ill be back
Hi GLADvertising,
You’re right. The search engines have been collecting information about searches and searchers’ searching and browsing activities for years. One of the challenges that they face is in deciding what information to look at, and how to use it to learn.
For instance, while they could look at individual queries from searchers, chances are that they could learn even more from looking at query sessions, where there are a number of related searches from the same searcher, and see how people might modify or refine their searches.
Or, as in this search trails approach, they might be able to learn from where people go on the web after conducting a search – which pages they browse after that search, and how long they spend on those pages, and if they seem to start a new query sessions after visiting certain pages.
While they might have been collecting that kind of information for a while, deciding that it might be helpful to show searchers some of the search trails that other searchers have followed is relatively new.