Did babies named Ava cause the U.S. housing bubble?
If you follow the trend of babies born named Ava, it matches the U.S. housing bubble perfectly. Though that Ava in the chart is adorable, I want to curse her for the horror she has wrought on the economy!
But wait: this is just an example of the danger of confusing two correlated facts with the concept causation.
(Thanks to BusinessWeek’s Vali Chandrasekaran for this great example. Click through for some more.)
I see this all the time in presentations we make to grocery buyers. And it’s hard to fault anyone for going this route.
Going on a sales call is far from an exercise in objectivity. You have a case to make when you’re on a sales call, and if there’s some correlation that makes your case, any salesperson worth their salt is going to use it.
Sales teams never have the funding they need to bring in the analytic horsepower they need to run real experiments on the right way to price an item, shelve a product, run a promotion, or make an assortment decision.
I have a hard time avoiding this tactic when senior executives who are looking to make a particular point want you to do it, the facts be damned! But it’s bad for business in the long run — both for the retailer if they accept such specious reasoning and for any manufacturer who succumbs to it.
It’s easy to see why it’s not in a retailer’s interest to accept bad data in making decisions: they’ll end up with the wrong item, the wrong price, the wrong whatever for their shoppers. For a manufacturer, they often feel like if they can at least sneak a decision through, it’s in their interest. But it’s not. It’s just a ticking time bomb. If it’s a bad decision, any buyer or category manager will notice it with enough time. You’ll lose trust and — maybe more importantly — the decision you won won’t last.
If enough analytic resources are in place, you can present good decisions based on good data that favor your interests in the long run. And if the right results aren’t there, it might be better to just be quiet and present nothing at all.