The Dark Side of Automation

When I was a kid, I programmed my Commodore 64 to pick lottery numbers for my Dad.  Sounds easy, but I wrote a flexible program that could pick numbers for the daily numbers or Lotto (in its 40, 48 and 54 iterations) and it picked numbers without repeats.  It took me a few hours to get it right.

During one of my programming sessions, my Dad ventured into my programming dungeon in the basement, ordered me away from the computer, and typed the following:


Of course, the computer shot back with ?SYNTAX  ERROR

“Stupid computer,” he said.  “If computers are so smart, why can’t it understand what I want?”

I laughed it off and continued writing my program.

I couldn’t have foreseen a world of self-driving cars, smart homes, school robotics clubs, or the ubiquity of computers.  In that world, Dad’s question is a lot more relevant.

There’s an old joke about a computer programmer’s wife, who sends her husband to the store with a simple command.  “Get a loaf of bread.  And if they have eggs, get a dozen.”  The husband comes back home with 12 loaves of bread.

Computers are very good at performing tasks exactly as they’re described.  Nuance and insight are not their strong suits, which is why we need humans to tell the computers what to do.

As the computers are put to use automating marketing tasks, though, that blindness to nuance often causes more problems than it solves.  Some of us longtime marketers even file examples of dumb automation away when we see the havoc they wreak.  Motivity Marketing CEO Kevin Ryan showed us a bunch of them at his latest iMedia Breakthrough Summit keynote – acquisition ads trying to convert longtime and loyal customers, jaw-droppingly inappropriate instances of ad targeting, ridiculous canned replies to sincere pleas for help, and a lot more.

Tech and marketing are irreversibly intertwined, but advances in tech have left us lazy and less curious about what drives success.  The folly is that the tech universe believes that no one should care why an algorithm determined that, all other things being equal, an ad banner performs a smidgen better with a blue background than with a green one – just that making that adjustment from green to blue helps drive success.

But in order to replicate that future success, we need to understand why that small change drives better response.  (It satisfies our curiosity, too.)  There’s a reason behind it, and perhaps thinking about why will allow us to predict performance better than we otherwise might.  It certainly beats testing infinite combinations of variables to arrive at the answer.

With all of the emphasis on data’s role in marketing communications, we seem to have forgotten a key tenet of our roles and responsibilities.  Computers are good at iteration, humans are good at insight.