In March, Washington Post economics columnist Catherine Rampell had some less-than-nice things to say about the people who munge users’ social data to target advertising. She notes wryly that, within the space of a few weeks, she saw advertising on Facebook for a Mormon dating Web site turn into advertising for engagement rings and then, a few days later, advertising for baby products. “Dang, I thought, Facebook thinks I move fast,” she quips.
Like Ms. Rampell, I too often find myself baffled by the scattershot advertising and recommendations I see across the Web. Why, for example, do apparel companies send me emails touting the latest discounts on silk blouses? Between my social login at some sites and stores, and my purchase history at all of them, there’s ample data to show I will likely never have an interest in floral print skirts, let alone some time in the next week.
Other recommendations systems often fail to pick up on these patterns or take a very long time to do so because they don’t connect as deeply with people as they could. Namely, they don’t use the rich stores of information people have developed on social networks. These additional data are a win-win. Shoppers don’t have to wade through the potentially dizzying and overwhelming variety of products it’s possible to offer online — making their experience more pleasant — which has an obvious benefit for retailers.
Indeed, in my admittedly anecdotal experience, shopping and interacting with some of our clients, very few companies have really embraced social networks and social media. And it’s at their cost. J. Crew’s spring financials show, for instance, that people who connected with the retailer through social media spent nearly twice as much as the average shopper. This is not to mention the potential of doing more than connecting on a Facebook page.
Assimilating social network data into users’ recommendations allows us to see the whole person, not just their activity as it relates to the products on a specific site. A person would almost certainly make a different recommendation to someone with a passion for high-performance cars and no children, and her friend with a similar interest and a large family to drive around as well. Moreover, armed with this knowledge, we can instantly come up with relevant suggestions, rather than slowly emerging from a fog of confusion as they begin browsing online and we slowly discover their tastes and preferences.
This is really nothing more than an evolution of the same sort of relationships we build in “real life.” Just as the sales staff at my local housewares store know me and can then offer suggestions tailored to my whole person — not just on the basis of what I’ve bought in the past — adding social network information can inform our calculus when it comes to building people suggestions online. The servers running our code may be somewhat faceless and imposing, but it’s we humans who are telling them what to do. We’re bringing the know-how and helpfulness of a great shopkeeper online and into the twenty-first century.
The benefits of using social data are clear. Everyone wins when it’s easier for people to find what they’re looking for. And while, as my fellow data nerd Ivan pointed out in a previous note on this blog, this is a complicated problem that demands a complicated solution under the hood, our hard work makes using the technology easy. Our focus has allowed us to refine not only what we can do, but how we do it, especially at the point of integration. Get in touch with us at firstname.lastname@example.org to learn more.