Search Challenge: Evolve from "Result" to "Response"

I’ve been reading Sinclair Lewis’ novels lately. Lewis was the first American to win the Nobel Prize for literature. He possessed an extraordinarily insightful eye and fine-tuned wit. He was uniquely able to satirize human behavior in a way that makes one almost jealous of those who possess such interesting faults.

In one book, Lewis writes of an aspiring young man who just got a new automobile. Overwhelmingly pleased with the machine’s power and functionality, he boasts to friends he needs only turn the engine crank six times to get it started, that it only stalled twice on the way over, and that he cruised at the breakneck speed of 18 miles an hour. What a machine! Truly, he marvels, this is evidence of the powerful ingenuity of engineering minds.

I read this and thought only of search engines.

We at Jupiter Research just completed our annual consumer survey. The data are still in the rawest form, but I notice something pretty striking: users are overwhelmingly satisfied with the search experience, but only because they’ve gotten good at making it work. They only have to crank the engine a few times.

The Data behind the Details

Here’s a bit more detail. We asked:

In your experience over the past year, have search engines become:

• Significantly better at returning relevant results to my queries?
• Somewhat better at returning relevant results to my queries?
• No different than previous years at returning relevant results to my queries?
• Somewhat worse at returning relevant results to my queries?
• Significantly worse at returning relevant results to my queries?

The response was great news for the search engines. Forty-five percent thought the engines were “somewhat better;” 21 percent thought they were “significantly better.” Only four percent thought things were “somewhat worse,” and I had to round up to get to the one percent who thought things were “significantly worse.” So, overall satisfaction with search engines’ ability to return relevant results (remember everyone, that’s the main job of a search engine!) is pretty high.

The issue, though, appears when we ask what sorts of things people have done to get that relevant result. It put a bit of a damper on the satisfaction rating: 41 percent believe results aren’t relevant to their search. OK, so let’s call that a “disconnect” and dig deeper. The same percentage of respondents say they refine their searches with additional keywords. About a third feel too many results are returned, and too many of those results are sponsored links. A third also use more than one search engine to get the result that’s relevant to what they want.

For a large part of the Internet population, I’m beginning to see search is becoming a techie/geek activity. That’s a far cry from the vision of a clean and simple interface to the world’s knowledge that’s so often portrayed.

What’s Wrong with the Search Experience?
Negritude Ultramarine

Users are doing a fair amount of work to get the results they need from search engines, but give the credit for the results to the engines, not themselves. This is not the way the user-centric computer revolution was supposed to work.

The computer was supposed to transcend its status as a complex tool one must learn to use, and instead adapt to the needs of poor human users. With few notable exceptions (Ask Jeeves comes immediately to mind), there’s been too much of a focus in search engine development on “results.” People don’t want results. They want response.

A response is an answer. It’s a focused and clear return to a specific request. Queries get results; requests get responses. Results are raw; responses are complete.

I know there are technical issues behind this, but that’s not necessarily what I’m addressing. At issue is the mindset of the search community, which is far too focused on generating results. Take, for example, the recent SEO contest. The challenge was to generate the top result for the phrase “negritude ultramarine” on Google. The number of results for this nonsense phrase went from zero to thousands within a few short weeks. The top result (when I looked) was a forum page, jam-packed with the phrase.

OK. So you win. What does that accomplish? The phrase is completely empty. It’s meaningless (and a bit of an oxymoron at that). The contest organizers could just as easily have chosen “zxojhsswdf cjsdes.” But the contest was about generating results, not responses.

A far more interesting contest would have been a competition to generate the top result for something like “Reagan’s economic legacy for large cities.” That challenge would be more substantial, more meaningful for users, and ultimately more important for the development of search engines. With “negritude ultramarine,” the solution tends toward a brute force approach. With a more conceptual search, something requiring a response rather than a simple result, the solution would have to be more elegant, and more thoughtful.