Caution: Your data may mislead

Ever wonder what you have in common with yourself? I didn’t really, either, but an app I was using for social analytics showed me my own account and presented me with a view of what I had in common with @kateo.

According to this metadata, here are some of the things I share an interest with myself about:

• Big Data, Data Visualization And Infographics, Dataviz and Infographics. Well, OK, those were gimmes.
• Parenting. I’m publicly on record (in TIME magazine, among other outlets) as being child-free by choice. So that’s actually an understandable semantic link; it’s just a misleading one.
• Both Country Music and Classical Music. I live in Nashville, a.k.a. Music City, and yes, I have ties to country music and the industry, but this one serves more as proof that computer-led analysis can be imbued with the jumpy biases of its programmers, since “Nashville” = “country music” to many people who don’t know anything else about the city. And Classical Music, while I respect it, has less significance in my digital life than, say, bacon does, and that’s saying something since as you can probably infer from the Vegan, Vegetarian, and Raw Food tags above.
• Pay Per Click Marketing, Ecommerce, Testing & Optimization Software, Advertising & Marketing, Email Marketing. Sort of, I guess. They’re all, like, fractional pieces. But I get that “digital behavioral strategy” is a pretty esoteric conceptual space. And I’ve certainly expressed interest in topics relating to each of these areas online. So those are forgivable oversimplifications.
• Sports. Ha!
• Horror. I can’t even. Maybe we should interpret that as part of a set with QR Codes. Or US Politics.

If you were trying to use this metadata across a user base to build targeted messaging and experiences, based on how my own authentic interests align and misalign with this data, I can tell you you’d miss more often than you’d hit. Which would maybe be OK if you’d built learning cycles into your process, so you could continually refine your understanding of your audience and what resonated with them.

Data is just dots. Analysis is trying to draw lines between or around those dots, but there’s no guarantee you’ll produce anything truly meaningful. It usually takes some understanding of context to make any sense, or meaning, out of data, and that’s more true the more abstract and open-ended the data is, such as social metadata.

A sound business data strategy involves both framing up data collection so that what you collect is most useful, and looking at the data collected in the context of business realities.

Now if you’ll excuse me, I have a virtual reality nature hike to plan gamification strategies for.