Can Blekko be a Disruptor in Search?

Blekko is a new search engine currently in private beta, and I have been playing with it for the past few days. Co-founder Rich Skrenta says upfront that Blekko is not a Google-killer, and I agree. However, for a few search enthusiasts to begin with, it is a very interesting Google alternative to come up in many years.

Blekko Search

If you are unfamiliar with Blekko, read this introductory article by Mike Arrington. For a detailed look, read this in-depth review by Danny Sullivan.

A SlashTag For Techmeme Leaderboard

I wanted to have a handy way to search all the websites that make up the Techmeme Leaderboard. It turned out to be simpler than I thought. A straight import of the OPML file helped create my “/TMTop” slashtag that I could use to get quality search results for anything related to technology.

For generic search terms like “credit card”, the difference between search results from Google and Blekko is obvious:

Google Credit Card

Blekko Credit Card for TMTop

Higher Relevance With Curated Search

When comparing approaches to filtering for relevance, I noted how Google search is built almost entirely on algorithms, with minimal human intervention directly on search results. Being a monopoly in the search business, Google has gone to great lengths to ensure that its search algorithm is fair and impartial with no human bias.

Blekko turns this principle upside-down, by giving end users the ability to curate their search. This mix of human + algorithmic filtering leads to potentially very high relevance of search results. Why potentially?

Keyword vs. Slashtag

Consider an example. Let’s say I’m searching to troubleshoot problems with iTunes on a Windows PC. The key question is: can Blekko’s “iTunes problems /windows” perform better than Google’s “iTunes problems windows”? The answer, at present, is no. Google’s first result is Apple’s official support site for iTunes on Windows, while Blekko doesn’t include as part of its “/windows” slashtag.

In fact, at present, even a plain search for “iTunes problems windows” without any slashtag on Blekko doesn’t return the Apple support site in the first few results.

These are difficult challenges for Blekko. Slashtags may not be as effective as you might think. This is because curation is an either-or affair – there is no ‘maybe’ as there can be deep inside an algorithm.

Combining Social Features with Search

Blekko has added social features by enabling you to “follow” other users’ slashtags. This means those who can aggregate a carefully curated set of websites within a slashtag stand a chance of being followed by several other users. This sounds appealing as anything social does these days.

But a reality check: who makes “following” popular on the web? Celebrities and Websites/Blogs whose primary objective is driving traffic to their own content. A slashtag may be a curator’s achievement, but it drives traffic to various sites by definition. Thus, I don’t see any popular brands, celebrities, or content creators to drive the social features of Blekko, hence I suspect it will remain restricted to the minority of search enthusiasts.

Impact on SEO: Slashtag Optimization (STO)?

Will Blekko’s human curation mean that algorithm-focused SEO will suffer? That largely depends on market share of Blekko’s adoption. Greg Sterling has a nice post discussing this issue.

Imagine being able to set default slashtags in your search preferences that filter content farms, adult websites, etc. Search will get a boost in effectiveness of several orders of magnitude. This, coupled with the transparency Blekko brings to the table about its internal SEO metrics, is one of the best things to happen in search, in my opinion.

Even if a minority of search enthusiasts adopt Blekko, I see two possibilities:

  • Google may tweak its algorithm to penalize content farms, as is being suspected
  • Google may offer tools to filter the web in its own searches

In my opinion, if either of these happen, Blekko has proved to be disruptive.

This entry was posted in Miscellaneous and tagged , , . Bookmark the permalink.
  • What I see as the big problem, as you hinted at, is content farms. They have gotten so sophisticated that it is hard to algorithmically filter them out.

    Google’s main underpinning of trust is the PageRank. Content farms have gotten very good at cross-linking and emulating the network of websites which lead to higher ranks, even if the actual content may not be accurate or from a knowledgeable person.

    At the very basis of every search is a person wanting to find out information on a topic. I think ultimately that trust is going to be sources from other trusted users, rating and sharing articles, to help score the ‘top’ sources. The social network is becoming more influential in this area.

    Although Blekko is a good idea at heart, the vast majority of the population won’t ‘get’ slashtags. Natural language search is really the only way a search can be done for the mass population, at least for now.

    As a niche search engine, it might work. But I don’t see it becoming a major disruptor.

  • Dregar,

    Agree with everything you said. That’s why I’m evaluating Blekko’s disruption by its effect on other search engines (primarily Google) rather than by the extent of its adoption by the mainstream.

    Nice thoughts from you, thank you!

  • Pingback: DataSift Curation Engine Aims for Relevance in Real-time by @ScepticGeek()

  • Pingback: The Filtering For Relevance Matrix (FORMAT) by @ScepticGeek()

  • Rmarshall406

    I tried Blekko, and I think it’s pretty cool. The slashtag thing is unique and I think it’s a great feature. It took a little bit of time to figure out. Not sure if it’s going to become a dominating force like Google. I don’t think I can stop using Google, which is why I really like another new search engine called Bweezy. It has Google results, but offers them in a more user-friendly way. There isn’t anything to “learn” and you can even open up search results in the same window.

  • Pingback: Mapping Startups & Services Filtering For Relevance In A Matrix – Mahendra Palsule - MediaNama()