Posted in: amazon, Business, facebook, marketing, social networking, web20

Social Shopping and Expertise Identification – The Missing Piece in the Puzzle?

I have been noodling on this whole idea of social shopping, which I think is still a poorly-defined concept. I have to warn you in advance that this is a fairly academic-style post.

What does it mean to “shop socially”?

To me, social shopping is about leveraging the collective wisdom of others, be they friends or others, to make better purchase decisions.

I think it’s just that simple. It’s not just about what my friends do and not just about what everyone on Amazon does either. There are very strong use cases for each instance. Part of why I’ve struggled with this whole concept is that I haven’t seen a good breakdown of what people really mean by “social shopping” – so I’m going to break it down in a way that make sense to me.

First I think there are 3 different types of people whose input I value as a shopper. The three broad classes can be found below:

  • Friends – People I actually know or to whom I have some connection.
  • Smart Strangers – People who are not my friends, but are likely to have some particular expertise on a particular topic. Think about a power reviewer on Yelp – someone who knows a lot about your local environment but is not exactly a friend of yours. You’re relying on their expertise as an individual, not the weight of collective opinion.
  • The “Collective” wisdom of others – What did other people, facing the same decision as you, actually choose to do? The best example of this is Amazon’s feature that tells you what other folks who looked at the same product actually chose to buy.

Now that we have the classes of people out of the way, what else do you need besides people to make social shopping work? I think there are two other key pieces you need:

  • A Broad and Deep Collection of Ratings and Reviews – You need to make sure that you have good breadth of reviews and that each object reviewed has some depth behind it to reduce the effect of noise.
  • Aggregation of Actual Purchase Behavior – At the end of the day, you want to be able to combine reviews and actual purchase behavior. This is not just so that you can do the Amazon-like piece around telling people what the collective did when presented with the same choice. You also want to be able to connect reviewers with actual purchase behavior. In many cases, you get this “for free” – if I only review restaurants that I’ve actually visited or cameras that I actually own, this is moot. However, it reduces (but does not eliminate) the possibility of gaming by posting fake reviews on products you don’t actually own.

With all of that background in place, why don’t we have more social shopping applications? It’s my theory that when most people hear “social shopping” they think of the use case where people only want to know what their friends purchased or use their friends as a key source of input when making decisions. This is only one of the few use cases imaginable, but it’s really interesting.

If this is the use case that seems to have captivated people (I’d argue that Beacon was a clumsy attempt at enabling social shopping), why isn’t it working. I think the missing piece in the puzzle is the ability to identify experts among your friends and contacts.

Expertise Identification– Of all of the things that I think are holding back social shopping for the “friend” use case, I think expertise identification is one of the biggest ones. What is expertise identification? Simply put, it’s the ability to determine who in your social network is an “expert” on a given topic. This is where I actually think the real opportunity lies and where a properly-implemented system like Beacon or some other system that silently keeps track of what your friends are buying / reviewing on the web might be made useful. I have no interest in knowing about each and every purchase made by my friends and associates on Facebook, MySpace, LinkedIn, or any other social networking site. Right now, it seems like there are two major ways that I see people figure out who in their social network might be able to answer a question for them:

  • Post a blast announcement to your friends (update your IM status, Facebook status, send an email, etc)
  • Rely on coarse “clues” to guess who might know something about the product (your own memory, clues from information in their profiles, etc.)

To be fair, a friend of mine was working on a concept called friendput in this space that relied on very explicit input and contribution by users to build up expertise. I think the only way you’re going to be able to actually surface experts within a network (most likely) is through passive profiling of what people buy, read, and write. Relying on active user contribution might work, but I’m not sure why I’d want to tag myself as an expert or identify myself as an expert manually. You get into all of the complicated questions of how to map across individual tag taxonomies and what not. Also, “expertise” is a relative term – I might not consider myself an expert in some subject where my friends might think I know quite a bit. Relying on self-identification is both prone to gaming and time consuming for the user.

The reason (I think) we haven’t seen systems like this emerge yet is largely privacy driven. It’s one thing to look at questionable material on the web. Actually buying questionable products is a much stronger signal of one’s intents and interests. I don’t know that most people today would want to have their purchase histories a) indexed by a 3rd party service or b) aggregated across their activities and identities on the web and c) made visible to their social networks. The ability to exclude certain purchases or websites would be an obvious must-have feature. But I can imagine a number of folks who’d rather not use a system than have to invest a lot of energy in manually managing the information in the system.

If someone can find a way to get this right, I do think it’s a big idea. And thanks for reading if you made it this far – feel free to share your thoughts in the comments section.

Comments (3) on "Social Shopping and Expertise Identification – The Missing Piece in the Puzzle?"

  1. Hey, Charles, a friend pointed me to your post. I think you’re exactly right: the key to a truly useful social shopping app is connecting users with trusted ‘experts’ who can provide meaningful shopping guidance. I’m the founder of a recently-launched social shopping startup, Lootist, that is directly addressing that function.

    We opted for the “active user contribution” route (i.e., our specialists self-declare themselves and their qualifications), and I’ll explain why. The thinking is that our model will appeal to the “one-percenters”, the group of people who actively contribute the majority of content on sites with socially-driven content. Whether it’s on a personal blog or in Wikipedia, these folks are motivated to share their specialized knowledge in large part by growth of their online reputations. We’ve designed Lootist to provide the recognition they deserve.

    Regarding another of your points… we’ve also implemented a user-endorsement system that helps to decrease the subjectivity in judging the relative expertise among our specialists. That is, the specialists who are the most knowledgeable and helpful get higher visibility.

    Anyway, that’s the nutshell of my approach to the challenge you’ve posed; I hope you’ll check it out. I’d love to have a more in-depth conversation about it.



  2. You shop like a guy – value experts and reviews too much. InStyle, Lucky etc are like recreational shopping; going to the mall just to see what is new. A lot of the social shoppings sites are similar, just crowd sourced (e.g. Stylehive – a Lightspeed Company, ThisNext etc)

  3. Thanks for the friendput plug Charles :). The site’s goal was exactly how you define social shopping above: better purchase decisions through collective wisdom, and you’re spot on that explicit content creation is a challenge, even when folks are appropriately incentivized (ie helping out a friend as through friendput). I think you’re also correct when you say “actually buying…products is a much stronger signal of one’s intents and interests”; looking at purchase behavior data across users, together with user intent as expressed through context and past behavior, might indeed be the way to get to your definition of “social shopping”. Enter richrelevance 🙂

Leave a Reply

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

Back to Top