Why the interest graph is the future of social commerce

Truly social commerce, like truly social search, is one of the great unrealised hopes of the social web. It’s been frustratingly close for some time but remains tantalizingly out of reach.

The basic premise, of course, is straightforward: by combining social data (who I am, who my friends are) with commercial data (what I buy, what the retailer wants to promote) brands can deliver a more tailored, relevant shopping experience with benefits for both the user and the business.

Taken a step further, enlightened brands can enable users to sell direct to their social networks. This may be about identifying power users and enabling them to sell on the brand’s behalf (a kind of social affiliate model) or a more straightforward member get member approach, with baked in social context.

The possibilities are hugely exciting, both for the everyday user, for the micro-entrepreneur and for brands-who both open up a layer of rich, free data and potentially unlock a salesforce they could never afford using conventional distribution.

The reality has been, with rare exceptions, less compelling. Yes, there are some exciting outliers, such as Etsy’s gift recommendation engine, Giff Gaff’s social sales force or Converse’s new social sales platform enabling users to design and sell their own trainers but they remain in the minority.

Converse Create and Giff Gaff: rare examples of truly social commerce

There are a number of reasons for this:

  • Facebook’s  “like” button is undoubtedly the most ubiquitous social signal we have. There are 2.7 billion likes and comments a day. It captures a truly extraordinary amount of data at an amazingly granular level. If I’ve liked a pink toaster (I haven’t), it knows about it.
  • Integrating that data set with a retail database however is no small task. Cross matching the many and various things I’ve liked, or my friends have liked, and aligning them with product recommendations involves significant development time and effort.
  • Moreover, while the ubiquity of the “like” button has unlocked a new data stream for Facebook, its primary data set arguably remains the social graph, rather than the interest graph.  While the social graph is undoubtedly a hugely powerful data set, there are two flaws increasingly impacting its potency from a commercial point of view:
  1. There is a latency to the social graph: Perhaps it’s my age but my Facebook friends are primarily people I’ve known for a long while-school friends, college friends and former colleagues.
  2. While there is an implicit relationship between who I know and what I might like it is by no means a perfect relationship. Herd theory would suggest that the relationship between my network and my behaviour is profound, and Facebook have impressive data on the ripple effect of friends-of-fans but this doesn’t easily translate into the seamless, intuitive social shopping experience of our dreams.

The interest graph, by contrast, has both currency –what I’m reading, looking at and liking right now-and an explicit articulation of what I might like to buy.  Jeremiah Owyang recently observed that:

“Pinterest in the real world is like millions going to the shopping mall, filling shopping carts… then walking away.”

I’m not sure whether he meant it as a compliment or not, but it is this explicit signal of purchase intent that makes the interest graph so compelling-traffic referral from Pinterest would seem to bear this out. John Battelle’s seminal thinking on the “Database of Intentions” was evolved a few years ago to include the Commerce signal. Perhaps over time the “commerce intent” signal will become equally important.


The social graph versus the interest graph

Finally, although this is more a result of design and development choices than of an inherent difference in the data sets, services built on the interest graph currently have an open-ness and fluidity I’d argue those built purely on the social graph lack-they feel currently better placed to operate as a layer that sits across our entire online experience versus a destination.

It is the power of the interest graph that is fuelling the explosive growth of Pinterest, the buzz around Springpad 3.0 and (in part) the acquisition of Instagram. I found this quote from Mark Zuckerburg one of the most intriguing parts of the press coverage surrounding the deal:

“For years, we’ve focused on building the best experience for sharing photos with your friends and family. Now, we’ll be able to work even more closely with the Instagram team to also offer the best experiences for sharing beautiful mobile photos with people based on your interests.

We believe these are different experiences that complement each other”

The complementary point of course is an important one. Social data and interest data are by no means mutually exclusive.

In fact there’s a symbiotic relationship between the two. Social data often fuels early adoption of interest networks (it would be lonely on Pinterest if I couldn’t find some friends to get me started). Conversely, the interest graph rapidly creates new and dynamic social ties-so a fascinating data ecosystem starts to evolve with the social graph as a relatively static foundation and the interest graph as a dynamic, looser and rapidly evolving network that sits on top. A challenge for Google + (which I remain convinced is a much more compelling proposition than we give it credit for) is that it may currently lack the social foundation to build the interest graph upon-and unsurprisingly Facebook are unwilling to provide it.  The social graph still sits a kernel at the heart of many interest based services but I wonder if ultimately it will be less valuable commercially.

The social graph remains a kernel at the heart of the interest graph


So there are three key characteristics of the interest graph which I believe make it uniquely powerful:

  • Currency
  • Explicit signal of purchase intent
  • A layer, not a destination

But what does this mean for brands and businesses? I believe there are four interesting implications for any brand operating in the digital space:

The power of behavioural versus engineering solutions:

By triggering the right behaviour, services like Pinterest, Springpad et al have unlocked a data set that would otherwise have required a relatively complex (and ultimately blunt) engineering solution. How could changes in user behaviour resolve issues for your business you struggle to address with technology or infrastructure?

Providing a platform for self-expression:

These services have unlocked that behaviour by providing a platform for easy, frictionless self-expression. I don’t have to be a creator to show you what I like or find beautiful-and the majority of users aren’t-but I feel gratified when you appreciate my taste. How can you offer your users the same experience? What behaviour might it unlock?

Do one thing brilliantly:

What’s interesting about all these emerging services is that they do one thing brilliantly-photo sharing, pinning, springing, link sharing. By doing that they seem to create a different kind of user behaviour than more generalist networks-people seem to behave differently around a specialist service designed around a single passion, rewarding it with more quality input and a deeper emotional connection. Few brands though, have the confidence to do one thing well in the digital space-there’s invariably the temptation to add just a few more features. Among the rare examples of this kind of focus are Burberry’s Art of the Trench and Footlocker’s Sneakerpedia-and we love them for it.

Don’t forget to be interesting:

To survive in a world where consumers can signal at every turn what they find interesting, beautiful or entertaining, brands need to do interesting things. In a world where users can express themselves by assembling the images and articles that appeal at the touch of a button-and where those images and articles become a powerful purchase signal-our content needs to be something users are proud to share. The age of dull, uniform photography and corporate articles is over.  As John Battelle puts it:

“All brands are publishers. Learn to be a good one”.


Leave a Reply