Thursday, June 7, 2012

Less is More (Data)

In theory, more marketing data mean more insights.  Essentially, a sufficiently smart approach using modern (i.e. relatively inexpensive and powerful) hardware can boil down vast quantities of data to find the few bits that really matter.  Simple math, really.

In theory, a pint of Haagen-Dazs serves four.


One serving.  Two if my wife uses a VERY small spoon,

Here in the real world where it rains and sometimes dogs bark for no reason at all, less data work better than more data.



Here’s a favorite example.  Moons ago, I helped my employer--an email marketing agency--pitch a major credit card issuer.  As usual, their request involved some version of “take our email to the next level” and accordingly, they sent us examples of some recent campaigns.

As the strategist, I looked at the emails to divine just what kind of data they might have used to drive offers or creative.  Based on my professional judgment and perhaps a pint of Guinness, I suspected that they had not used ANY data for those purposes.  

A travel promotion, for instance, had overly broad offers, such as “travel this summer” rather than “fly this summer” or “visit Dubuque this summer.”  Surely, the issuer would know if its customers would prefer to fly or drive, to travel within the US or abroad and so on.  As a card issuer, they could easily build models off of past transactions and fine-tune the offers and content.  Say it with me: IN THEORY.

In this pitch process, we had no opportunity to ask questions before the presentation, so we had to wing it.  I reasoned that the prospect must have had some good reason not to have access to customer data so I developed a strategy that used implicit preference data to drive offers and content, rather than depend on their data.  The implicit preferences would come from what email links they had clicked.

My hunch turned out to be correct.  The card issuer told us that indeed, they had customer transaction data, several PETABYTES worth of them.  A petabyte equals one thousand terabytes, which in turn equal one thousand gigabytes.  At the time, frequent analyses of that much data lay beyond the reach of that marketer.

Over the following years, I came across the “we could tell you, but then we’d have to kill you” issue with data.  Marketers HAVE the data, but some combination of technology, process, regulation, legislation or just plain orneriness prevents them from using those data to power more effective communications.  As a result, marketers hoping to use data need to use their own orneriness to find solutions that make the most of smaller, more available data.

The Chinese have a saying: it is better to light a single candle than to curse the darkness.  (Aside: I once used that line in a presentation in Beijing.  They told me that thought it was an American expression.)  Marketers should seek out those single candles when darkness looms.  It beats bumping into things.

No comments:

Post a Comment