When Trusting Data Can Backfire

I just visited Re/code, one of my favorite tech-driven media sites. While reading a story about yet another HP executive departure, Re/code's ad server displayed a Dunkin' Donuts ad to me. My reaction? Context-driven advertising is getting pretty darn impressive -- but it's also destroying businesses.

Through some sort of tracking software and a third-party ad network (I suspect), the online world knows I'm a daily Dunkin' Donuts coffee drinker (iced latte in the summer; hot latte in the winter). At first glance, it's wise for Re/code to serve me up that coffee ad. But take a longer-term view of me -- the customer -- and you'll see how analytics-driven business decisions (made by humans or machines) can sometimes backfire.

What's Higher Value?

Think of it this way: I'm a coffee drinker and a business co-founder. I consume:

  • coffee
  • cloud services
  • latte
  • hardware
  • cappuccino
  • software
  • caffeine
  • and high-end technology

Look at that list closely. Some online algorithm somewhere in the cloud decided to promote a $3 cup coffee to me -- when I'm already a once-daily customer and can't possibly spend more than the $100 or so I already give to Dunkin' Donuts each month.

Somehow, the data and the algorithm also overlooked the fact that I'm a business builder -- and I'll need to spend more and more on IT services, hardware and software over the next few years. The potential loser in all this is Re/code ... which somehow isn't monetizing my hunger for innovative IT platforms.

Good Data, Bad Conclusion

And therein resides the danger with data. Just like statistics, you can spin the data to say what you want. You can misread the data -- spending all your time on clear trends ("Joe is a coffee addict; sell him more coffee") and overlooking the long-tail trends ("Joe turns on a new cloud service every 90 days or so. Sell him more cloud services"). 

Further complicating matters, data analysis often relies too heavily on numbers and too little on business relationships.

If Amy Katz and I were building a company together (ahem...), we'd somehow blend industry conversations with data analysis to pinpoint the next market trend -- and unique ways to capitalize on that trend. Heck, perhaps Amy and I would even be willing to enter shrinking or distressed markets -- if we knew we could add value to those markets or somehow redefine them or jumpstart them.