Years ago, a grumpy old professor of mine provocatively said, “The best check of a man’s intelligence is to see whether he generalizes from a single data point.”

What am I talking about? I’m talking about making a decision based on one experience, or statement, or person’s opinion.

It happens all the time. But it’s something to shun, especially in marketing.

I see it in content review cycles.  All too often, marketers radically rewrite marketing pieces thanks to one person being critical. Maybe that person is a salesperson, or a client, or a marketing director, or a spouse — but it doesn’t matter. Opinion is opinion.  It’s based on hunch, instinct, habit, rhythm. It can be skewed by a poor night’s sleep, by a desire to win political points, or out of an overabundance of caution.

I also see it in campaign creation. Often a salesperson or an engineering VP comes over to marketing as says, “Oh, we HAVE to run a campaign about XYZ, because if we don’t something horrible will happen.” Often, when you dig into this opinion, it stems from a single experience. The salesperson talked to a customer who wanted purple widgets, and suddenly the salesperson concludes that all people want purple widgets.

What these opinionated people don’t understand is that their opinions are just single data points. The problem with single data points is that they often distribute on a bell curve, and with one data point, you have no idea where the opinion is on the curve. It might be the most pessimistic view, or the optimistic view. It might be the view held be geniuses, or by idiots. With just one data point, there’s no way to tell.

What trumps opinion? Data. Data trumps opinion. That’s why the direct response guys have it right. Market testing, delivering thousands of relevant data points, allows you to see the distribution of opinions. It helps you to guide choices toward the most common view.

But what trumps data? In most marketing organizations, often relationships trump data. If you’re an up-and-coming marketing manager, you don’t tell your marketing director to buzz off — even if his opinion is a load of codswallop. You don’t tell your closest channel partner that he’s all wrong about your campaign. You don’t tell your spouse that she’s a programmer, not a marketer, and doesn’t know anything about headline creation.

In corporate marketing, this is often a real problem. Again and again, I’ve seen a senior marketer derail good work for no good reason, based on a hunch that proved to be completely wrong.

So, if you’re in an organization that values hunches over data, how do you get past this problem? I’m asking every marketing manager I meet, and would love your comments. How do you get past the single data point? When should you let a single opinion change your marketing?