We’ve been talking about big data for quite some time. In conversations with sales and marketing professionals, I have spent more than a little effort discussing how “big data” is much more nuanced than just predicting sales. (Although it’s really good for that.)
I read a piece on Wired today about the Netflix algorithm. It’s a discussion about how Netflix uses data from your viewing history to predict what you will want to watch next. In the article, Carlos Gomez-Uribe, VP of product innovation for Netflix says, “A lot of people tell us they often watch foreign movies or documentaries. But in practice, that doesn’t happen very much.”
So, the sad news is, we can’t fool Netflix. They know that we are really watching Pineapple Express and Hot Tub Time Machine, even though we say that we like foreign films with a historical basis.
I know enough about it to know that the same thing is happening at Safeway. When the “customer offer” prints out with my receipt, Safeway is applying my history. Their algorithm probably involves a blend of my purchases, possibly weighted toward more recent purchase history, added to a database of Safeway products, possibly weighted by profitability. The offer is either: 1) a coupon for something I recently purchased, or 2) a near-adjacency, like a more expensive brand of Greek yogurt, if I usually buy the cheap stuff.
As they say, history doesn’t lie. But, is that what we really want from marketing? Giving us more and more and more of the same stuff we bought last week or month or year?
This is where the REAL marketing use of big data will start to come into play. Personal history does not tell the whole story about where customers will go next. Change-enthusiasts don’t want to keep buying the same old stuff. If you only push their previous purchases, they may leave you in favor of a new marketplace where their preferences are not cast in stone. On the other hand, the change-averse will leave if you do the opposite: propose shiny new things when they want the same old stuff.
Netflix might be missing the boat if they don’t factor in my stated preferences for independent British comedies or documentaries about economics. Safeway might be missing out if they don’t suggest saffron-dusted grilled tilapia, with a coupon for a few cents off the saffron. It might sound good to me, and 5% of the time, I might buy the pricey saffron, even if I normally only buy salt-and-pepper.
Most of us WANT to live a little better, even if the actions captured by our personal histories show that we eat Cheetos in front of reruns of My Name is Earl. We want our lives to be more interesting, glamorous, trendy, than they are right now. Every once in a while – not every time, but on occasion – we want Netflix, or Safeway, or Amazon, or BevMo to assume that we might enjoy something a cut above the normal fare. Maybe I’ll never buy any of the suggested products, but I’ll think better of these marketers because they thought better of me. And their good opinion of me will make me a “stickier” customer, likely to do more business with them.
I’ve seen a lot of algorithms for customer analysis; I’ve written more than a few. So far, I haven’t seen many that propose to raise the bar, to help customers reach for a higher, better life. Let’s get to work on stretch goals for big data marketing, thinking about what customers WANT and not just what they HAVE DONE. Take the remarkable step of catering to the highest common denominator in the data, instead of the lowest.