Not long after my Facebook initiation, I tried to access it from a hotel room far from home. Facebook felt it needed to verify my identity by showing me photos of three “friends” and asking me to identify them. The “friends” were all classmates from schooldays decades in the past: I couldn’t even match their names to their high school faces, let alone their faces of today.
This episode comes to mind every time I think of using social data in commerce. Do I want to dine where my old (in every sense) classmates dine?
I imagine the social commerce tools of the future will be able to identify not-really-social people like me, and make reliable decisions about how to engage us. For now, a social commerce tool that can gather and use social data is a big first step. Big, because the world of commerce hasn’t really figured out how best to apply social data, and in which contexts or circumstances. Heck, I’d say the world of commerce is still struggling with much older and more familiar problems, such as how, where and when to place ads.
And big, because the quantity and quality of social data will change how we go about personalizing customer interactions.
Graphdive, a new company in the social and personalization arenas, provides social data for ecommerce. It does more than gather and analyze social data: it creates user profiles, interest segments, and recommendations. The user profiles contain demographic data, some of it inferred by analyzing the user’s network. A person might be identified as married based on characteristics of their friends, posts and likes.
Users’ key interests will be inferred from likes and checkins and posts. Graphdive applies semantic analysis to equate concepts such as trombone, euphonium, and brass band competition, enabling it to determine that these three likes represent a strong interest in brass music. These interests are stored in the user profile, and enable marketers to search for people with a specific interest, creating interest-based segments. Marketers can also use the Graphdive service to identify other users who are similar to their best customers. Graphdive uses the information to select product recommendations for users, matching user interests with the product catalog a retailer provides.
Graphdive has a handful of clients and pretty good press attention dating from September 2012.