6 Comments

Summary:

Information from Facebook on user “interests” is nothing compared to the affinity data that could be mined from Pinterest. The potential to aggregate, analyze, and deploy the site’s data is huge. Baynote CTO Scott Brave examines the types of affinity data that Pinterest could monetize.

thumb tack

Pinterest, and sites like it, are valuable for driving traffic to retail sites. No doubt about it. However, the big value proposition that retailers and Pinterest may initially overlook is the underlying user data. Information from Facebook on user “interests” is nice, but it’s nothing compared to the kind of interest and affinity data that could be mined from Pinterest. And although mining this data may require some investment on the part of Pinterest and perhaps retailers, the potential to aggregate, analyze, and then deploy it is huge.

In my experience as chief technology officer of Baynote, a provider of personalized customer experience solutions, I’ve seen a big shift among retailers, who are now using analytics to study the customer experience. In today’s era of big data, retailers are looking at massive data sets of seemingly incongruous information about customers, products, and purchases to find patterns that can be used to deliver more relevant content and experiences that increase revenues. GigaOM’s own Derrick Harris recently pointed out new insights that retailers can gain from big data analysis in his article on what Orbitz has learned about consumer preferences. The curated content revolution, driven by Pinterest and sites like it, should excite retailers because of the rich “affinity” data they have made possible. Affinity data reveals valuable relationships between consumer behaviors, products and content that can then be used to create more targeted marketing.

The concept of mining affinity data — or “affinity modeling” — isn’t new. Online businesses have been using this technique for years to power personalized content and product recommendations on their sites. As highlighted in the recent article in the New York Times Magazine, “How Companies Learn Your Secrets,” retailers can use affinity modeling to find connections that are counter intuitive yet highly profitable.

This thinking is not so different from how retailers should approach Pinterest. Consumers are expressing affinity for products by pinning, repinning and creating their own boards by topic. Surfacing these affinities is one way that Pinterest could monetize their data for retailers and marketers.

Let’s examine the kinds of affinity data that Pinterest could monetize.

Product to theme

Perhaps the most obvious affinity is the connection that curated content naturally makes between products and themes. Retailers organize products by categories such as running shoes and hiking shoes. When consumers pin the same items, however, they may choose new themes to group them in, such as “killer hikes” or “my wedding outfit.” This type of affinity data provides retailers with new insights into how consumers view their products, and gives retailers new ideas for ways to market their products.

Cluster to theme

Individual pinning choices are interesting, but there is an even greater opportunity to analyze segments of people who express an affinity for a product or category in aggregate. The connections between groups of people based on defined dimensions, such as gender and location, is very useful to marketers using segment based techniques. Drawing segmentation connections from curated data allows marketers to create new segments or map existing segments to categories, brands or products that they could use for more accurate and targeting. Similar to Facebook, Pinterest could tally the number of users who fit a targeting criteria. For example, Pinterest could determine based on clustered affinity data that 30,000 women in San Francisco are more likely to prefer Seven Jeans to other brands. And, more importantly, Pinterest would know who they are.

Product to product

Another way Pinterest’s affinity data is useful is that consumers also group different brands together under the same themes. For example, products from J.Crew and Urban Outfitters grouped in the same theme might help retailers decide on which complementary products to develop, potential partnerships to form, inventory to stock, and opportunities for up-sell or cross-sell. Retailers may not have connected these items on their own because they are in different product lines, or suit different purposes. With this affinity data retailers gain new information about how their products are related to other products from the consumer perspective.

Similar to the open graph in Facebook, Pinterest has an opportunity to monetize affinity data. This data could greatly help advertisers and retailers figure out how to target ads, content, and products to consumers. It could also prove to be the missing link between what we know today from Facebook and what we wished we knew about what people are truly interested in.

Scott Brave is a founder and chief technology officer at Baynote, a provider of personalization solutions for online retailers. Previously, he served as lab manager for the CHIMe (Communication between Humans and Interactive Media) Lab at Stanford University. 

Image courtesy of Flickr user wayne’s eye view.

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

  1. Nir Zion Pengas Saturday, March 24, 2012

    insightful commentary scott – thanks.

  2. David Fishman Sunday, March 25, 2012

    Good luck with the “big data” boondoggle. I agree there’s gold in them there hills, but until retailers can learn to harness this information and organizations can leverage across depts. it’s just another network! Polyvore has been at it longer than Pinterest …. and I’m not clear how retailers benefit. Not to mention that Polyvore is not profitable and Pinterest probably will never be either.

    The richest retail data is that which can be recycled back to the shoppers to assist in the decision making and buying process. Until tools are made available to the retailers that enable them to provide these services to the consumer …. Pinterest and Polyvore will just be self-serving networks …. and of little value to the retail audience they “claim” to serve.

    The problem begins at the retail level and hence must be solved

  3. @David Agreed there are structural issues with retailers that need to get “fixed” before Big Data from Pinterest and Polyvore can be useful.

    Fundamentally I believe that big data can help inform marketing and merchandising decisions, but my experience has been that few retailers are actually capable of acting on the data in a timely fashion. It’s not that they don’t have the tools or the data, it’s that the merchandising cycles are still on a very long fixed timeline, and realtime data simply can’t impact what products get purchase, manufactured, marketed, etc. It’s too big an operational dependency. H&M is something of an anomaly in its short cycles, as are other specialty retailers who do reactive purchasing/mfrg, but I’d venture to say that even *they* couldn’t make maximum use of internet and mobile data right now. The general retail industry is still pretty focused on trying to get more intelligence from stores and partners, which is problematic enough. Scrubbing data for returned items alone can screw up assumptions made earlier about “hot” products or segmentation.

    HOWEVER: rant about perfection aside, if merch teams could simply get in the habit of looking at what gets shared and what types of editorial styling appeal to buyers, it would be a step in the right direction. My sense now is that it’s not in their daily “to do” list. Sometimes the most powerful insights are qualitative, not quantitative.

  4. Scott, great insight into the value of affinity data for retailers. A similar use case was done by a marketing agency who used HPCC Systems to analyze usage patterns and behaviors on Pinterest data. I can see more companies taking advantage of open source platforms such as this to help with analytics and obtain value from their data.

  5. While it’s true that Big Data can’t solve all the world’s problems, this is a bit more tangible than that. As consumers curate content in mass they are creating a new data stream that retailers should care about. In a C2B (consumer to business) driven world, retailers need to find ways to listen to their customers in new ways. Content curation is one new way that consumers are communicating with each other and a new way that businesses can learn from to build more personalized and valuable customer experiences.

  6. Marion Shaw Kumar Sunday, April 1, 2012

    Perhaps a bit ahead of the curve for some readers, but foretells the promise of retail differentiation and profitability in big data analytics. Thank you Scott for sparking this discussion!

Comments have been disabled for this post