Consumer research is complex enough on its own. But if, in addition, we rely on memory as the main source… we are building on shaky ground.

Why? Because memory does not work as we think it does. It is not an ordered archive of memories that we can access to retrieve what happened, as we saw or felt it. In reality, memory stores only fragments of what happened, and each time we remember, the brain reconstructs the scene… filling in the gaps with what it thinks should have happened. And that new, modified version is the one we store again.

The more we remember something, the more we change it. And yet we are convinced that it is true.

A mistake of the brain? Not at all. It’s a brilliant efficiency strategy. As Leonard Mlodinow explains in Subliminal, we receive about 11 million bits per second of sensory information… but our conscious mind can only process about 50. So, the brain summarizes and reinterprets what we experience so we don’t collapse.

The result: we remember well the general, but poorly the details. And on top of that, we do it with total security.

This has huge implications for marketing. Because if you ask someone how many yogurts they ate last month, they won’t lie… but they won’t get it right either. They simply don’t know.

Luckily, today we have other ways to find out: shopping panels, real data, automatic measurements. No need to ask if you can already observe.

And there’s something else: the brain is even worse at predicting the future. There is a huge gap between what we say we will do and what we actually do. Neuroscience proves it over and over again. So what’s the point of continuing to ask “would you buy this product?” as if it were a reliable prediction?

The key is to choose the right technique for each type of question. It’s not about stopping asking, it’s about knowing when to observe, when to measure, and when to investigate differently.

This is not new. But it is more urgent than ever.