Video Details

Search engines have traditionally operated on words and links however, this paradigm makes it difficult to account for what Ian Lurie in this Whiteboard Friday calls ‘random affinities’ – ideas that have connections to each other but without really any logical basis. This is the basis for the Ideagraph, which measures the correlation between different ideas.

Ian Lurie predicts Google is trying to use the Ideagraph to produce search rankings. To find the commonality between seemingly unrelated ideas, Ian think Google will use Google Plus as its source and seeing users as the previously missing ‘link’ between ideas. E.g. the idea that many clyclists may like beetroot would has no rational basis, however, if Google saw many Google Plus users posting content or engaging with content about cycling and also beetroot, it could draw from this that its likely that many people who are into cycling are also into beetroot. As such, it could better predict search intent produce results more pleasing to the user.

Finding Random Affinities

Ian suggests the following sources to find random affinities:

  1. Facebook Ad Planner – it will return results based on demographics and also the interests of users. If you enter a specific interest, it will show you related interests which are often based simply on many users liking both those things, even if the interests are albeit seemingly unrelated.
  2. Collaborative Filters – look at any sites that user collaborative filters to make recommendation. An example would be Amazon’s ‘people who bought this also bought…’.
  3. Followerwonk – look at the followers of a particular person and research into their bios and see if you can draw any interests in common amongst followers.

Using the Ideagraph to Impact Your Marketing

It’s always easier to market to people using the overlap of ideas important to them. Ian suggests implementing the following to best leverage and attain this data:

  1. Use Rel=Author and Rel=Publisher – this will help Google tie random ideas to a central person or company.
  2. Shema – Google recognises these markups and uses them to return more relevant results.
  3. Google Analytics Demographics – this will provide you with demographic data of visitors to your site. This data can be the basis for further research to find random affinities relevant to your target audience.