Search Engine Blekko Brings Human Evaluation to Search Results

Two years after he sold his second start-up, the online news aggregator Topix, to a trio of media companies, Blekko CEO Rich Skrenta was ready for a new challenge. During his 20-year career, the computer engineer had worked for AOL, Sun Microsystems, Unix Systems Labs, and Commodore Business Machines, and had also founded two companies of his own. But he had never worked on a search engine.

“Search is the hardest problem in Silicon Valley,” Skrenta says. “It’s the Holy Grail.”

As he was thinking about his next career move, he realized the quality of the available search engines was on the decline. The major search engines in the U.S. relied on social cues to rank the relevance of their results, but it had become too easy to game the system and use search engine optimization techniques to get a better ranking.

“There’s an entire industry to manipulate this giant corpus and put stuff on it,” Skrenta says. “If that’s the process, you can’t guarantee the quality of the search.”

So in 2007, Skrenta decided to leave Topix and take his core engineering team with him. The founders set out to create a new search engine that could guarantee the relevance and safety of its results. Instead of relying on social cues for ranking, Blekko would add a layer of curation, using human review to help determine the most useful links.

A Blekko search results page on heart disease

A Blekko search results page on heart disease

To do it, they created a function called slashtags, terms like “/health” or “/education” that users could include in search. Instead of just typing in terms like “free credit report”—a search that on another site might even pull up links associated with identity theft—users could add in “/finance” to see only results that had been reviewed by Blekko’s partners at The Motley Fool. Searching for “San Francisco Giants/sports” would only bring up results that had been ranked by partners at SB Nation. Slash tags like “/date” or “/blogs” would organize results chronologically, or search the content of blogs.

The Redwood Shores, CA-based startup also employs a Wikipedia-style model, where volunteers can apply to be curators in a given area. “They’re doing it because they’re subject experts,” Skrenta says. “When you start to pay people, it sort of corrupts the reason people do it. If you edit Wikipedia, it’s because you care about your reputation as a subject expert.” The main goal is to find knowledgeable curators who can give users high-quality results that they can trust.

The technological challenge of creating an algorithm and building a new search engine is massive, but adding in the additional element of human curation makes it even more complicated.

“We have a bigger investment in core technology than a basic startup,” Skrenta says. “It’s like building the 787 Boeing. It’s a fantastically complicated system. I’m an engineer, so I think it’s a lot of fun, but the scale of engineering challenge often keeps me up at night.”

It took three years, but the team launched Blekko in 2010.

The company is still working to refine the search and add even more levels of curation, but it has also created a second product, a social news site built on top of the core search technology. Called ROCKZi, the site launched earlier this year.

To Skrenta, it’s different than something like Google News, which uses standard search to pull up news stories and presents them with a different user interface. For him, simply creating a site called was “short-changing its potential.” Instead, the company created … Next Page »

Single PageCurrently on Page: 1 2

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

One response to “Search Engine Blekko Brings Human Evaluation to Search Results”

  1. Uatoo says:

    I want to suggest also UATOO,, a new human search engine, with new interesting and original features.