Follow Us - | | | | or call   Brighton: 01273 907 920   London: 020 7183 8468 | Contact Us

The Don’t Be Evil Search Engine

January 26, 2012 Author: Tim Aldiss

I posted recently about how True Social Search is Coming and mentioned briefly about a backlash based on the fact that Google are now quite happily pointing people to their own properties, but felt compelled to put pen to paper because, what appears to be, a viable alternative to Google has appeared left field.

Briefly – my opinion on whether or not Google should or shouldn’t rank it’s own properties is this: if Google’s own search algorithm is able to return results from websites like Youtube, Picasa, Google News, Blogger, etc then what’s all the fuss? I mean it may well be the case that they are artificially inflating G+ right now, but surely that is set to change as more content appears on it and it gets more authority?

A list of Google properties

A list of Google properties

Anyway there’s a fly in the ointment, and it’s a big one.

Search Plus Your World

As a response to Google’s integration of Google+ and Google Search – Search Plus Your World (SPYW, or just Search+ as it’s being monickered) – and it’s seeming bias, with the help of ex-Gmail creator Paul Buchheit, and Blake Ross director of product at Facebook, a cartel of Twitter, Facebook and Myspace (of all people!) have created a cunning solution that could just be the start of the demise of Google.

This beast in the corner started life as just a simple browser add-on. However what it does is remarkable, and it’s not just remarkable because of the results it displays, but remarkable because of how it calculates the results. That’s right – it has it’s own algorithm. And it’s a clever one.

Don’t Be Evil

Let’s look at the results first. The bookmarklet is effectively a button you can press on your browser when you have already done your Google search (don’t forget that you have to search on Google.com if you are in the UK to see Search+ results).

The click of the button automatically updates the biased Search+ results to feature results from any and all other social properties. There are several good interactive examples of this over on the Focus On The  User website.

The clever thing about these news results is how they are compiled. Here’s the detail from a Danny Sullivan post on the subject:

—————————————————————————————–

We could debate which is the most best social account to show for each of the examples above in various ways, such as:

  • Should it be based on followers? If so, in many cases, accounts on the established social networks of Twitter and Facebook would easily trump Google+ accounts, though not always.
  • Should it be based on activity? Britney hasn’t posted to Google+ since January 18. Her last tweets are more recent, as of January 21. Does that count for more?
  • Should it be based on engagement? Even if Britney hasn’t been on Google+ recently, perhaps her engagement with fans is better there?

Rather than try to figure all this out — and potentially come under accusations they were selecting metrics to favor their own social networks — the developers of the Don’t Be Evil bookmarklet decided to let Google itself figure out which social account was most relevant.

For example, let’s go back to how the listing for Snoop Dogg was changed:

What happens is that the bookmarklet looks at which accounts are listed initially by Google in the People & Pages section. It then does a search for the people or companies listed to find which social profiles Google itself believes are most relevant, in the first 100 results.

In the case of Snoop Dogg, the bookmarklet sees that he’s listed as a People & Pages suggestion for a search on “music.” It then conducts a search for “snoop dogg” on Google, to see which social account Google lists first for him.

As you can see above, Twitter is the first social account shown, so the bookmarklet rewrites Snoop Dogg’s suggestion to link to his Twitter account. His MySpace account is found next, so it’s listed as the first alternative. Facebook is found after that, so it’s the second alternative. Google+ is found after that, so it becomes the third alternative.

According to the tool’s FAQ page, it checks for links from any of these social networks:

  • CrunchBase (weird, I wouldn’t consider this a social network)
  • Github
  • Google+
  • Facebook
  • Flickr
  • Foursquare
  • FriendFeed
  • LinkedIn
  • MySpace
  • Quora
  • Stack Overflow
  • Tumblr
  • Twitter

All the information the tool finds comes solely from Google itself, even the profile pictures, which are generated using Google’s own rich snippets testing tool, as the FAQ page explains.

Because only Google’s own information is being used, the tool can’t patch up a key weakness, which is that all the suggestions are initially based on people who are in Google Plus.

For example, you will never get Katy Perry as a music suggestion, despite her being one of the most followed musicians on Facebook or Twitter. That’s because she’s not on Google+, so she has no chance of appearing as a suggestion to music-related queries. If she doesn’t get suggested, then the tool can’t add her.

—————————————————————————————–

Brilliant! But what I think is so genius is that right now it only uses Google – Google’s index, and therefore Google’s own algorithm.

Google need to sit up and listen and make a decision. If they are starting to manipulate their own algorithm realising the limitations of it’s abilities then they need to find a way to be unbiased or this will drive people away.

What are your bets? A new cartel formed of Google, Facebook, Twitter & Myspace? A new search engine – MyFaceSpaceTwitGoo? Or several on-going class-action lawsuits?

Related Posts Plugin for WordPress, Blogger...

Leave a comment