Where Einstein Meets Edison

How the Interest Graph will shape the future of the web

How the Interest Graph will shape the future of the web

Apr 1, 2011

What do Color, Quora, Hunch, Blippy, and StockTwits have in common? They are examples of companies that generate value for their users by leveraging the concept of the Interest Graph. The list also features some of the most promising startups right now, having raised close to $100 million in venture funding. Pure coincidence?

What’s the Interest Graph? The term surfaced around the launch of Gravity, which takes a data-mining approach to discovering people’s interests. The term then became more widely used in reference to Twitter. Essentially, the Interest Graph is an online representation of individuals’ interests, with people and interests being the nodes of the graph. Connections exist between people and interests as well as between interests and interests. The strength of the connection depends on the level of interest or relatedness between concepts and can change over time. Here’s an example: When living back in Europe, I used to be a big soccer fan, but since moving to the US, I have become a fan of the Boston Celtics. I also follow football games from time to time.  On my current interest graph, there is a strong connection to the Boston Celtics and a weaker connection to football and soccer. One can derive my general interest in sports, as these interests are connected to the sports node in the Interest Graph.

Online communities, from the early forums and discussion groups to the likes of StumbleUpon, Digg and Reddit, have always played a central role in grouping people around topics of interest. However, applying the Interest Graph concept to link online search and discovery to offline purchases catapults the realm of opportunities to a whole new level, as demonstrated for example by the latest Groupon valuation rumors.

The Interest Graph has been described as the “middle ground between Google and Facebook – between search, advertising, and the social graph”. Simply put, Google creates their version of the Interest Graph by mining my search queries and other data collected online, for example through Gmail or Google Maps.  It then offers advertisers a way to personalize their messages. One of the problems is the often high noise level in the data due to the lack of context (e.g. I might be looking up something for a friend rather than myself), which decreases relevancy. Recently, there has been a lot of buzz around social search as studies have shown that friend recommendations are much more powerful than traditional advertising in influencing consumer behavior and purchasing decisions.

While Facebook has been able to bring social trust to online search by enabling social endorsements and sharing via its “Like” button, some major flaws remain: We assume that 1,000 “Likes” are better than 100 and that “Likes” from within my social circle are better than “Likes” from strangers. However, I might have interests that I do not share with anyone in my social circle. Also, the algorithm behind the Facebook feed emphasizes recency over relevancy, burying useful bits of information in the digital nirvana. Ever tried to find a friend’s status update that you had seen in your feed a few days earlier? Good luck.

Like Facebook, Twitter has not been fully able to fix the problem of information overload. Some might even argue it has contributed to it. So what is the problem with the former poster child of the Interest Graph? The problem lies in the design: On Twitter, I am following both friends and other people who I believe have similar interests to mine and whose status updates I might find insightful. Individuals’ interests typically don’t align perfectly and I end up with a lot of irrelevant status updates. Hashtags mitigate the problem, by connecting people with particular topics. However, a look at social Q&A site Quora shows a better solution to the problem: I can follow specific topics (such as “Climate Change”) and questions (“What are some radical ideas to stop climate change”), in addition to people.

I strongly agree with Chris Dixon that we will continue to see a lot of innovation around specific verticals of the Interest Graph, further increasing relevancy for users. Location in particular is interesting. Meetup.com, a company that helps groups of people with shared interests plan meetings and offline gatherings, was started long before the term Interest Graph became fashionable and is a great example of how location and interest intersect. Color, a social app to share photos real-time with people around you, has surprised the tech world by raising $41m pre-revenue. By design, it has chosen location to be the most relevant factor to drive its content. It will be interesting to see how they can further increase relevancy within these highly local implied social networks created in real-time.

I mentioned Groupon earlier – now imagine the power of Groupon combined with a record of people’s current interests, nicely packaged for merchants, to stimulate demand and enable discovery that will ultimately lead to purchases. You could call it Personalized Micro-Groupons. It’s no coincidence then that investors have very high expectations for the companies mentioned above.

Companies that successfully design personalized filters for a given domain and that are able to produce and help discover relevant content and recommendations will shape the future of the web and reap the benefits of bridging the divide between online and offline interactions. However, there is no general recipe for finding the right mix of social, peer and/or expert opinion, and detailed personal profiles. Their success will depend on how well they are able to capture all the complexities of personal taste and preferences, which can change not only depending on location and time but also mood, a variable much harder to model.

About

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>