If the first phase of e-commerce was the utilitarian hunt for staples, the next phase of e-commerce is about recreational shopping. So it needs to be a more fun and social experience which gives an edge to companies that encourage deeper social experiences.


A year and a half ago, I spent a few hours at the offices of Hunch, a New York-based startup, learning about their decision engine. By asking you seemingly random questions, the engine helped you make decisions. Hunch’s engine was a nice way to aggregate what you liked, then help you find information based on that assumption. For me, the real potential of this decision engine was commerce, and that’s why I thought perhaps Amazon should buy Hunch. It could use the decision engine to help customers sift through the ever-expanding array of offerings and make purchasing decisions. That little kernel of an idea still looms large in my thinking, especially as I wonder what the future of media and e-commerce looks like.

Social Spending

Last week, I was chatting with Lightspeed VC’s Jeremy Liew, who has invested in companies such as Bonobos, ShoeDazzle and LivingSocial. He pointed out that the first phase of e-commerce was about shopping for staples. It was utilitarian, and he pointed to the success of companies such as Diapers.com, Amazon and Zappos. The next phase of e-commerce is about recreational shopping, and as a result, it needs to be a more fun and social experience.

No wonder there seems to be a growing obsession with companies such as Groupon and LivingSocial, part of an amorphous category called “social commerce,” which means different things to different people. Elizabeth Yin, co-founder of the wedding apparel shopping service Shiny Orb, wrote in a guest column: “the social shopping space is comprised of e-commerce sites that facilitate interaction among customers as part of a shopping experience.”

If that is indeed the case, I have to say today’s social commerce companies need to build deeper social experiences. But how? And where does social commerce go from here?

Enter the “Interest Graph”

In July 2010, Chris Dixon — co-founder of Hunch — noted we would soon enter a phase where “one graph to rule them all” will give way to more-focused, social graphs built around concepts such as taste, location and trust. In other words, these concepts could become the underpinning of what is now generically known as the interest graph.

At its very core, the interest graph is a way to organize a social network based on people’s interests. For instance, if you’re a fan of Charlie Sheen and Lindsay Lohan, it’s clear self-destructive Hollywood stars and their lives are what you’re interested in. The interest graphs are built through various mechanisms: by following people whom you deem as experts, through your likes and shares, etc. In the middle part of the last decade, we tried to do this through tags.

These interest graphs are more like mini-Twitters. Just as you can follow someone — Will Ferrell, for example — without being his friend, you can have an asymmetrical relationship with someone who has similar musical interests or taste in watches. As a blogger for Asset Map, a San Francisco-based startup, noted:

Music, movies, books, articles — these are all things where people have tastes that aren’t always influenced by friends — or at least not a big group of your friends. It’s no surprise to me that the most successful music services so far are things like Last.fm and Pandora that are far more organized around your musical interest graph than your musical social graph (AssetMap Blog)

Interest Graph + Commerce = Transactions

Interest graph, for me, is the underpinning of a new kind of e-commerce experience. Think of it as a new kind of social commerce experience that goes beyond the notion of group shopping (Gilt Groupe, Groupon), shopping communities and recommendation engines. When Apple launched Ping, its music-oriented social network last year, to me it represented a template for social commerce.

Since Ping’s launch, I’ve downloaded songs based on the likes and recommendations of people who are not necessarily my friends, but who I follow because they have good taste in music. Sure, I have friends who are good at picking tracks, but Ping’s social layer has helped me discover new artists.

A few years back, I met Jeff Bezos and asked him why he was buying up content sites. I suspected the Amazon founder wanted to eliminate the “advertising” between commerce and content. If you remember, in 2007, Amazon bought DPreview, a digital camera community, and later acquired IMDB, a movie database.

As always, Bezos was a little ahead of the curve. In the post-Facebook, post-Groupon world, one can see a new kind of symbiotic relationship emerge between the interest graph and the “sellers.”

The concept is no different from enthusiast magazines of the past, such as Stereo Review, except there are “network effects” at play. Network effect, according to the Wikipedia definition is, “the effect that one user of a good or service has on the value of that product to other people.”

While enthusiast magazines were limited by the geographic boundaries and dollars publishers could spend on attracting new customers, in the Internet age, the network allows us to spread the word at a rapid clip, especially amongst people with similar interests. More importantly, since sellers can target the exact interest graph they want, they can skip advertising entirely. Instead, they can come up with an actual offer that leads to a transaction.

For entrepreneurs, I believe there are opportunities to create unique experiences around the concept of “interest graphs” that can be built off the backs of uber-networks such as Facebook and Twitter. These networks can help find the right kind of audience to build a viable channel for new commerce experience.

Uh oh!

Something is wrong with your Wufoo shortcode. If you copy and paste it from the Wufoo Code Manager, you should be golden.

You’re subscribed! If you like, you can update your settings

  1. Arnold Waldstein Tuesday, April 19, 2011

    This is a timely post. Thanks for writing it.

    I’ve been digging into the curation sites…and there are a host of them. They are all based on finding a way to cut through the social noise by focusing on curating the interest graph.

    My take on this from a marketing point of view:

    “Choosing context over friendship to filter the social web” @ http://bt.io/GwKV

  2. Agree. Interest graphs will be the ‘spoke’ data/social networks while Facebook and Twitter are the ‘hub’ social networks. I love the example of Stereo Review magazine because thats where someone who is really into stereos will go. However, now on an online version of stereo review, we can introduce that person to other stereo lovers and we can capture great data on which stereos they have owned (& want to own). And now advertisers can actually deliver a relevant offer to this person.

    Thats exactly our plan on http://meetoncruise.com where we are building an interest graph of cruising – past cruises, future cruises combined with some basic demo information can really add value to both cruise consumers and advertisers.

    Enjoyed the post – thanks.

    1. Arnold Waldstein Faraz Wednesday, April 20, 2011

      A commenter put it this way that said it well and worth sharing:

      “Nothing beats targeted relevancy. And social serendipity will not beat targeted discovery”@wmougayar http://disq.us/1q9j7b

  3. Of course I have a Facebook interest graph (my most boring), a Twitter interest graph and many others online and all different. BUT, my most important and valuable interest graph doesn’t live anywhere online. The ones online are superficial in comparison. So, I’m not sure how they pull this off as I suspect most are like me.

    1. I think that is where the opportunity lies. I think if you look at Instagram, they have created a photos-oriented interest graph but now need to find a business opportunity to go with it. Same goes for others, many of them still perhaps at the drawing board.

  4. Great post and one I certainly agree with in terms of the elimination of advertising, the network effects of social commerce and the rise of the shopping “experience”.

    We’re working on these things with http://sellsimp.ly. In fact social is baked right into the product as it’s a marketplace baked into Twitter, not Twitter baked into a marketplace. We’ll own a robust commerce interest graph.

    Looking forward to following through with some of the article’s claims.

  5. Dr. Phil Hendrix, immr Tuesday, April 19, 2011

    Unleashing Social Commerce, particularly for Mobile Users

    Thank you, Om, for a very timely post. I agree that “interest graphs” are a critical ingredient in social commerce – in mobile commerce, interest as well as other, more dynamic graphs will play an even more important role. As Fred Wilson pointed out, “sometimes you only want a social graph for a weekend, a day, an hour, or a minute” – this is especially true in mobile. In addition, signals can be inferred from mobile users’ digital traces, including location (“where I’ve been, where I am, where I’m going,” etc. – which may be viewed as “footstreams,” anaologous to clickstreams) as well as their spending, check-ins and other relevant behaviors. Mobile operators, credit card companies, the FourSquare’s of the world and Google, of course, see slices of mobile users’ behavior. By putting these pieces together one can (i) accurately infer mobile users’ interests at particular points in space and time, and (ii) present them with highly relevant, compelling “opportunities” that are actionable. The M-Commerce platform that delivers on this vision will create extraordinary value for consumers, merchants and advertisers alike (see http://j.mp/g5gxJk for discussion).

    I’m also moderating a panel on M-Commerce at the upcoming Social-Loco conference (May 5, SF – http://socialloco.net) where we will discuss these opportunities. Panelists include Michael Wu of Lithium (see Michael’s perspectives on “gamification” and the importance, care and nurturing of social networks at http://j.mp/hsjg1d) and Paul Bryan of Usography (see Paul’s critiques of leading retailers’ efforts to facilitate social commerce at http://j.mp/gKvzel). The panel also includes Pat Burns (Dash7 Alliance; see Pat’s presentation on “The Future of Check-ins” at http://j.mp/g9EojY), and Kim Finnell, CEO of deCarta, so we should have a great discussion.

    Dr. Phil Hendrix, immr and GigaOm Pro Analyst
    (author of Location – the Epicenter of Mobile Innovation, GigaOm Pro report, http://bit.ly/9ugm2M)

  6. interest graphs — forums and communities.
    Transactions are events in time – not something people want to keep up with unlike content.
    So, people need experts (who are in forums) to make their buying decisions. Actual e-commerce is price driven once the decision is made. I don’t see how interest graph etc. changes anything.

    1. If you see the interest graph in terms of forums, well you are probably not going to see the opportunity. I think the problem with forum is precisely the reason why newer forms of community need to emerge.

      1. Forums are today’s version of the interest graph. The issue is the data is not structured in a way where relevant offers can be delivered. So, we have to move beyond forums. Will be interesting to see which newer communities take off – and if they are similar across various interest graphs.

    2. Dr. Phil Hendrix, immr and GigaOm Pro analyst sharad verma Wednesday, April 20, 2011

      Sentiment analysis and network analysis (especially distinguishing strong connections and identifying influencers) do enable companies to make inferences from unstructured data and offer relevant, meaningful recommendations. A number of companies (see for example, GeoIQ, Lithium, and others) are making sense of and leveraging big, “messy” data for these very purposes.

  7. madhavshivpuri Wednesday, April 20, 2011

    I totally agree with interest graph as the next wave. But… it can’t be the data from FB likes because most people like other’s silly status updates, quotes from John Keats or a gaffe posted by an stranger on Youtube. It’s got be more real life.

    1. Yes…but Disney with Tickets Together on Facebook over a year ago launched what I found to be one of the first real uses of social comment.

      For big brands the interest graph surfaced early on Facebook:

      Social commerce on Facebook gets real with Disney and Diesel http://bt.io/GwWa

    2. I think FB is a good starting point and a nice way to jump-start your community but I agree, it cannot be the end-all if a service wants to build an engine that matters.

      1. True…

        Betting your future on FB is not wise. I see it more and more as a spoke and less as the hub for small companies and interest communities.

  8. Brad O’Neill Wednesday, April 20, 2011

    This makes sense, Om. The interesting aspect to the extension of this interest graph is going to be how the various explicit and implicit inputs are actually managed on behalf of a user. In your Sheen/Lohan example, you’re citing an implicitly derived insight… one that doesn’t need to be structured in a preset taxonomy or declared by the user. This could turn into just high-end targeted trash very quickly (which means it will)… So doing this well, and “in the wild”, is really difficult, but it opens the door to an entire range of new interactions… It’s getting to the state of literally predicting desires, which leads to an entire inversion of a commercial site’s value chain if followed through to the radical conclusion: “We know you are going to like something like this so we’re just going to make it just for you right now, show it to you, and see if you’ll actually buy it.”

    1. Agreed. I don’t think there is any denying to the fact that unless properly managed this can get ugly pretty fast. I do believe that what separates interest graph from previous efforts of structured data mining is the element of time.

      What I mean by that – the graph needs to have a “time” input built into it.

  9. this is exactly what we have built – http://www.getabl.com

    Interest graph – tighter concentric circles of friends sharing local buying experiences.

  10. Obviously, you are heading away from social in getting at hyper-targeted commerce. My belief is that it is very contextual and will probably be an opt in process, where “like” means more than affinity.

    In certain spaces people will tell you exactly what they think. In others, they will definitely not. I tend to believe the latter is the social space.

Comments have been disabled for this post