September 1st, 2006
Google’s Vertical Search Problem and the Law of Average Users
Poor Google — all these great vertical search options, from images to maps to books to shopping, and yet the vast major of their users still default to plain vanilla Google searching. In an apparent effort to change this, Google is testing a new search results page that pushes Google’s vertical search options (see left column of screen shot):
Google’s problem with vertical search adoption (as documented by Bill Tancer at Hitwise) is a lesson on how average people adopt and use applications. Most people use Google’s main search rather than the vertical options (which are plainly listed on Google’s main page) because they see “Googling” as a simple, elegant, one-stop-shop for any and every search need — this is how most people function. Using different search engines for different types of searches is several standard deviations of complexity away from average user behavior — the people who adopted Google for its simplicity are the same people who won’t stop to think what type of Google search they need to do. They just Google search. Just one, plain, simple application.
Life is complicated. People cling to simplicity and proven results. Behaviors are deeply ingrained and do not easily change.
The law of average users tells us that most people have significant limits to the amount of variation, complexity, and nuance they can tolerate in their daily application use. Adopting a new application for one purpose is a huge hurdle. Adopting multiple applications for variations on that purpose is just too much for most people.
This is another example of the Zen of 2.0. One search application is Zen. Many search applications is not Zen.
Powered by 
Design by 
Scott,
One of the things I find interesting is that most people have very little idea how search results are generated. I’m not saying they should, but the fact is that millions of people are placing near complete faith in a brand (Google) that will expose them — or NOT — to information that over time will have a huge direct and indirect impact on every dimension of that person’s life. You’d think that people would like to take greater control, and maximize vertical search applications. But our sophistication as humans is limited (or the interfaces and usability of the vertical search applications simply still suck).
Cheers,
Max
I think it might go deeper than what average people do.
People are mentally ill-equipped to accept probabilistic results. The fact that what a search engine does makes it probably right sits uncomfortably in the mind - we are set up to deal with greater levels of certainity than this. The great success of Google as a brand, I believe, is that it is has set itself up in the popular imagination up as authority above and beyond the practicailities of how its probabilistic search results are actually generated. People “ask Google” or “Google” for results (that disputed but commonly-accepted verb again!) as if there is some anthropomorphic entity called Google that can “know” things, because that’s easier for our minds to accomodate on a day-to-day basis than the reality that we are interrogating an algorithm that’s only probably - in the technical sense - accurate.
Stretch that brand any further and the mind rebels. No longer simply asking the all-knowing Google sage, one is suddenly asking Google Images or Google News for answers. It’s sufficiently not the same that the convenient congitive dissonance trick that we’ve played on ourselves fails.
Max — that’s precisely the issue — people can take greater control, but they don’t. Adopting Google search in the first place as a better way to find stuff on the web was a big enough leap for most people — and they significantly improved their results. And now its: But wait, improve your results EVEN MORE by using the right Google search for the job — it’s just more than most people have the cognitive energy to handle. We should all eat more fruits and vegetables, but mostly we just eat whatever is on hand.
Seamus — that’s fascinating. Google is in a sense a theistic phenomenon. And most Americans (and other Westerners) are wired for monotheism. Using Google’s vertical search applications is like believing in more than one God.
post
[…] People are continuing to think about vertical search from many different angles. Roger Ehrenberg, in his article on the possible manipulation of Digg results , provides a unique perspective by making a parallel between interest-based vertical communities and vertical search engines. This certainly brings up the question - are vertical search engine results more accurate and harder to game because of their focus and expertise in their specific domain? Susie Kang has written a fascinating article on using life stages behavioral targeting to improve conversion rates, a process that is particularly well-suited for vertical search engines. In this article on Google vs Yahoo, Donna Bogatin takes a different, content-oriented approach to vertical search : video, images, groups, news, maps, local and so on. Scott Karp explains Google’s Vertical Search problem in the Zen 2.0 blog. […]
[…] This post by Scott Karp over at Publishing 2.0 is tickling my design fancy. Ponder these quotes: […]
[…] life changing comes along. Scott Karp – September 10th, 2006 | Email | Print | Link | Del.icio.us Bookmark | Submit to Digg Categories: Technology, Attention, Email, AveragePeople […]
that is particularly well-suited for vertical search engines. In this article on Google vs Yahoo, Donna Bogatin takes a different, content-oriented approach to vertical search : video, images, groups, news, maps, local and so on. Scott Karp explains Google’s Vertical Search problem in the Zen 2.0 blog. Looking forward, the leading vertical search engines are continuing to innovate in a number of areas, to try to differentiate themselves from the big search engines. The most exciting areas are: integration with