When you have a great idea—a true light-bulb moment—you don’t want to be bothered with pesky marketing research that tells you it won’t work. “People who really, really want to believe, will believe,” says Steve Cuno in an article at MarketingProfs, “regardless of where the evidence points.” But to avoid potentially costly mistakes, he has this advice:

Conduct a valid predictive test. “The trick is to quit asking people to tell you their behavior,” says Kuno, “and discreetly watch it instead.” Send out a pair of identical offers that differ only in their headline, for instance, to see which works best; then retest to validate your findings. “If you get the same results, the evidence is that you’re on solid ground.”

Make unemotional decisions. “Your objective is to sell widgets, not to bolster your ego,” he notes. “As you design a valid test, resolve in advance to accept the results, even if they fail to support your hunches.”

Maintain control groups as a matter of policy. Without a representative selection of customers who haven’t been exposed to a particular campaign, there’s no way to know whether a spike in sales was driven by the offer or coincidence.

Finally, resist the urge to jump to conclusions. Even after a valid test, you might want to read the data in a way that favors the result you wanted. Don’t guide the evidence; let it guide you.

“The more you train yourself to eschew unwarranted leaps and instead approach marketing from a sober, rational standpoint, the more you will find yourself creating and refining campaigns that are demonstrably and measurably successful.”

Source: MarketingProfs.