American Apparel using data to drive decisions
WGSN’s Sarah Owen & Holley Klein report on how American Apparel chief digital officer uses data to promote engagement.
With his previous post on the insights team at Fox Broadcasting Company, Thoryn Stephens brings quite an analytic eye to his role at Los Angeles-based retailer, American Apparel.
At the recent ClickZ Live conference in New York, the chief digital officer divulged some of the secret sauce behind analysing raw data from his time as a research scientist at Cytokinetics.
“You might be asking how I went from synthesizing molecules to optimising booty shorts,” he noted before offering up some gems on sifting through mountains of data to detect customer behaviour clues.
Here’s what we learnt from Stephens:
On “fake metrics”: “Often times when we get caught up with digital analytics, we get caught up in what I call “fake metrics” such as Facebook likes or registered users. It doesn’t really focus on the value of the customers which is ultimately retention in actually driving engagement.” Speaking from previous experience, Stephens, “had tens of millions of registered users but when we looked at the current consumer value we actually found that we had thousands of users that were negative value; they were serial exchange and returners due to our exchange policy. Without that metric we never would have known.”
On measuring data at American Apparel: “Everything we launch we make sure we can measure it. That’s why my CEO loves me so much, because I can tie everything back to a return. Whether it is a return on investment or a return on ad spend; every penny that I spend I say that can get 5x return on that spend and 10x, 30x depending on the specific channel that we are optimizing against.”
On the data maturation curve: “What I have seen first hand at a number of billion dollar companies, starting at the basis of the curve, ultimately are we collecting the right amount of information on a variable level, for a given page, on a given website. It’s not just how much you are collecting, it’s what is the value of everyone of those variables. Number one is the actual collection of the data. Two is do we have the actual recording of the data, then you get into the more meaningful business impact such as dashboarding, and then the really fun stuff such as predictive analytics.”