Confessions of a Market Researcher: The Day I Beat My Data to Death

The day started out innocently enough.

I was looking forward to one of my very favorite tasks, data analysis.

  I had indulged myself with a printout of 8 banners of cross-tabs plus several variations on some multivariate output. My big, padded recliner and a fresh cup of coffee awaited. Ah, a day spent uncovering interesting themes, discovering unexpected patterns. If I got lucky, I might even find proof points to contradict widely held assumptions. What fun!

So with a pad of post-its and 2 highlighters (green for particularly useful results, red for items requiring deeper digging), I began.

About 40 minutes into it, I found myself feeling disappointed. My pages had very few green highlights. And not many red, either. Just…nothing. I referred back to my original hypotheses for the project, and started looking at the variables most relevant to them. The results? Mud. Boring, boring mud.

Just face it like a challenge, I told myself. So back to my computer.  I’ll rethink the ranges here, modify my rules for cutting outliers there, create some new variables over here, consolidate some variables over there. That should do it.

Huh. Still nothing.

Okay. Let’s go back to the original records. Maybe I have a lot of bad records. Unfortunate, but it can happen.  Hmmmm…no. The records had been cleaned sufficiently before it seems. That’s not it.

Okay, back to data analysis. Maybe I could modify my analysis plan. If I can’t get the story one way, I’ll get it another. Thank god for SPSS. CHAID, cluster, factor, regression…more banners….more more more. I wouldn’t let anyone else near the data. I was going to make this work,  @*#%!

And that’s when it happened. I lost it. I spent the next 4 days beating my data to death.  I poked it, pulled it, pummeled it. I sliced, slashed and skewered it. I tore it up into a thousand pieces and glued it back together into a Frankenstein mess. I abandoned it. I restarted from scratch. Seven times. I created 43 new variables, each of them a total waste. It wasn’t analysis any more; it was an act of violence. I was angry at the data, and I was exacting my revenge.

After the 5th cup of coffee on the 4th day, I took a walk. Crisp, New England autumn air always seems to put things into perspective. And the truth of the situation became evident.

I had a data set with no story to tell. It was time to let it collect dust in peace.  I wasn’t happy, and my client wasn’t either. Sure, we confirmed some stuff we expected. But we should have been able to do more. Much, much more.

[This happened more than 10 years ago. But I still think of it from time to time. I know a lot of people think of market researchers as calm, analytic types.

  But, alas, I, know…some of us have a darker side. How about you? Add a comment and let me know!]