Commentary

Metasearching For New Good Old Days

Remember metasearch? Dogpile wants to remind you what meta's about. Its research on the overlap (or lack thereof) of search engine results, coupled with rising consumer standards, make meta worth adding to your search radar.

The search landscape today is, of course, far different than it was in 1996, when Dogpile was born. Back then, there were a number of search engines, but search engine technology was embryonic, and there wasn't a vast amount of content to search. There are parallels to video search today. In 2010, you'll be reminiscing, "Remember when all you could find in video search were home movies and a few TV show transcripts?"

Today, search engines are pretty good. In 1996, an Internet user might have been looking for something that simply wasn't online. Today, the greater challenge is that search engines return so many results that the user seeks to find what's most specific to his or her needs.

I spoke with Brian Bowman, Vice President of Marketing and Product Development at InfoSpace.com, Dogpile's parent company, to find out how Dogpile is addressing this challenge.

advertisement

advertisement

Bowman, citing comScore data, remarked: "Fifty percent of all searches fail." On one level, it means that people aren't happy with the results returned. "Consumer behavior is showing they're dissatisfied with their search experience and are looking for alternatives," said Bowman.

I have a somewhat different interpretation. The failed searches may be a reflection of the "listen to the grumbles" phenomenon described by psychologist Abraham Maslow, where the more empowered people are, the higher-minded their complaints become. People don't just want to find a statistic; they want three sources confirming it. People don't just want the "Bull Durham" DVD; they want to see if there's a five-disc collector's edition with Kevin Costner saying his classic monologue dubbed in sixteen different languages. Consumers have such high expectations of search engines that they don't "feel lucky" when they find what they want; they attribute it to their searching skills, which they hone with every query and click.

Come to think of it, both interpretations arrive at the same conclusion. Dogpile avers that pooling results from multiple engines will satisfy the demanding search user.

Backing up Bowman is a study conducted by Dogpile in collaboration with researchers from the University of Pittsburgh and Pennsylvania State University, "Missing Pieces: A Study of First Page Web Search Engine Results Overlap." The study compiled a random sampling of 10,316 keyword searches across Google, Yahoo!, and Ask Jeeves, returning a total of 336,232 unique search results on page one. Only 3% of returned results (totaling 10,712 links) achieved first-page rankings across all three search engines, 12% of first-page results (39,959 links) were shared by two of the three engines, and 85% of results (285,561 links) appeared on just one of the engines' first pages.

If you're surprised at all, Dogpile welcomes you to try this yourself. Go to missingpieces.dogpile.com to test the Missing Pieces tool, and see how results overlap. In a cursory trial, "chekhov" returned one result on all three engines, "maytag dishwasher" and "watergate" returned two results common to all three engines' first pages, and "home improvement" returned three results shared by all three (including bobvila.com).

This might be a curiosity for you, and you'd be well-served to search for your company's or client's brands. Yet for Dogpile, this is key to its existence. The results common to multiple engines rank on Dogpile's first page of results. Users can compare the first various engines side by side by side. Another side will soon be added, as Dogpile will soon incorporate MSN's results.

Dogpile blends algorithmic and paid results, with the latter showing links preceded by "sponsored by," and all results noting which engine they came from. I wondered whether this "sponsored by" distinction was a little too subtle, although Bowman says a higher percentage of searches on Dogpile result in a click, and the internal tests with breaking out sponsored results separately didn't lead to any notable differences in user behavior.

The future for Dogpile brings a focus on local and mobile search, leveraging its parent company's strengths. The opportunity for marketers is hazier.

Dogpile doesn't deal with marketers and agencies directly. Yet metasearch may be due for a comeback. Look at other hot sectors of search: vertical search is essentially metasearch for vertical industries, local search usually involves aggregated listings from multiple providers, and video search's success will hinge on engines allowing searches of content from multiple entertainment and media properties. All are arguably some form of meta.

So will Dogpile bite into the leading search engines' market share? Bowman's take: "People will decide for themselves" whether metasearch delivers the best results. Ahh, search as the consumer democracy. That's an interpretation we can agree on.

Next story loading loading..