You've probably relied on market research at times in your business. Maybe you've informally asked people about concepts, had someone on staff do a more extensive set of interviews, hired an agency to perform a study, or used information from existing surveys.

No matter what type you've done, chances are strong that you've paid attention to a particular type of question--a critical one that should unlock important parts of your strategy. You ask people what they like, or what they'll do in the next six months, or how much they'll pay. And many of the answers you're getting back are uninformed garbage.

Not that people are pulling your leg, although some percentage probably is. But this is perhaps the least reliable type of question you can ask. A perfect example is offered by the recent Pew Internet Project research on public perceptions of privacy and security after the Snowden revelations.

For example, 91 percent of adults feel that they've lost control over how their personal information is collected and used by companies. Eighty percent of people using social networks voice concern over third parties and advertisers accessing data they share, while 70 percent have at least some concern about the government accessing the information. But many think the government can or should regulate advertisers' use of that personal information.

The public feels the least secure using mobile phones and social networks. But those are the mechanisms they often use when communicating. More than half of people were willing to share information about themselves with companies for free access to online services.

We are, all of us, a bundle of contradictions. One thing has our attention at one minute, and then something else the next. Look at the mass distracted attention we exhibit when it comes to national issues. We're sure we understand ourselves, and yet how many of us go off to therapy because we don't understand ourselves? We go into a store to buy one thing and then are manipulated to add on something else, even though we had no intention at the start to do so.

All of this is a collection of examples of why you cannot trust what consumers predict about their own behavior, or even what they report they regularly do. (As an experiment, estimate how much time you spend online or watching TV in a day and then measure it for a week and take the average. You will be surprised.)

That doesn't mean you should never ask people about what they think they would do. But bring that salt shaker and be ready to test each of the claims. It's the best way to get some certainty at a far lower price than taking people at face value.