Let us shed some light onto a common misnomer in today’s language and in database marketing — the word, correlated. The term has been used in marketing for years, either behind the scenes or when marketing a product to an audience. Look at cholesterol medication, for example. Marketers have cleverly implied that these products help prevent heart problems because cholesterol and heart disease are correlated. Something isn’t quite right here though.

Marketers, particularly those focused on database marketing, need to be aware that a correlation between two sets of data does not indicate that that one dataset affects the other. In other words, a correlation does not provide statistical evidence of a cause-and-effect relationship. We’re not saying that it wasn’t good marketing to point out the correlation between cholesterol and heart disease, but be careful basing your marketing decisions purely on correlations in your database.

Did you know that the occurrence of diaper rash and road construction is highly correlated? Does this mean that one causes the other? No. The missing link is warm weather. Diaper rash and road construction both happen to occur during the warm months of the year, but neither are directly related to each other (at least to our knowledge).

Unfortunately, marketers have had a hand in causing some confusion about this subject, so consider this our attempt to help set the record straight at least in the marketing community. Correlations can help guide your marketing decisions, but don’t always completely rely on them. When you analyze your target audience and how they respond to your marketing, take a moment to consider one level beneath what the correlations are telling you. Is it really only because they’re men that they’re responding better? Is the true trigger the fact that they’re between the ages of 18 and 24? Digging deeper using advanced database marketing or taking a closer look at your audiences’ attributes will reveal a better picture of what really causes them to act.

If you have the time and capability, run a statistical regression analysis to help uncover the variables that truly cause higher propensities to respond. With enough data and time, a skilled database marketing team can predict the actions of an audience with extreme accuracy, allowing you to segment groups that flat out respond.

And remember, next time you see that cholesterol commercial don’t be so quick to put down your delicious sausage grinder. You can eat with a little less guilt, knowing that ill fate is only correlated.

Read more: Maslow’s Hierarchy of Needs and Database Marketing