It's a common goal of marketers to start projects with research. (Well, most people want to, but they seldom do it. That's a topic for another time). The idea is that, armed with some measurements of the audience's attitudes, needs, and so on, we will have a better chance at structuring a successful solution. It's assumed that the hard work will be in interpreting and using this data, and almost no one thinks about the data itself. You need to be sure to ask is the data is even accurate. Those who do wonder about the data usually do so from the perspective of statistics. If we didn't get enough data points in the sample, then it might not be valid for the whole audience. Similar worries can arise from other sources: the sample is not segmented properly, there is error in the results, or perhaps the questions weren't formulated properly. All these concerns are valid. But, I want to bring up something else. Even if these statistical potential anomalies are dealt with, you still might have invalid data. This hidden source of error lies in the Heisenberg Uncertainty Principle. Heisenberg (a German physicist) taught us that, at least in the realm of subnuclear physics, you can know the position or the motion of a particle, but not both at the same time. This is true because every attempt to know these features alters the particle somehow. In other words, the measurement tools themselves introduce changes in the subject. We can apply the same "observer effect" to our understanding of market research. Imagine a focus group. You've convened this focus group to learn if one of your new offerings will meet a need. So, you ask the group something like "Do you have a need for a waterproof flashlight?". This seems harmless enough. But, the problem is that just that question, just the thought of that question, introduces new thoughts to the members of the target audience. They had never thought of a flashlight that's waterproof. So, now they are thinking about it. And whether they say "yes" or "no" to the question, there is no doubt that the question itself caused a change in the audience. You might think that the problem here lies in the formulation of the question. Perhaps it would solve the problem to ask something more like "What needs do you have?" However, even this seemingly unprejudicial question causes changes. For one, people don't think of having "needs". They live their lives in an active way; they don't contemplate their existence and thematize their mental states. To even to ask them to do this causes them to disengage from their active "use" of their needs and think about them objectively. And this conceptualization is a different view, thus their perception is altered by the question. Does this make market research impossible? Couldn't we do some contextual research instead of asking questions? If the subject knows she is under observation, we will have the same problem, just at the level of action instead of verbally. What if she doesn't know she's under observation? This might be the best approach, but I still think we have a fundamental problem. Even the observational act places the subject into a certain artificial context. That is, to even observe another human being is to alter their existence to "fit" into your frame of reference. In this sense, it is perhaps impossible to know what "market needs" anyone really has. Perhaps it's just as impossible as ever really knowing anyone. Human beings are just not objective beings (like trees or rocks). So, should you give up on market research? Probably not. Just be careful. Be aware of how your observations and study can alter the object of your study. And because of this, don't be a slave to the research data (or researchers). That is, be prepared to downplay or even eliminate research that doesn't seem to fit. Sometimes, there is no substitute for good old human judgment.