What Does Consumer Research Actually Tell You?

Suppose two brand assets are each correctly attributed to the intended brand by 70% of respondents.

At the top line, they look the same.

But imagine that for the first asset, the incorrect responses are scattered across many different brands. For the second, most of them go to one particular competitor.

The percentage is the same. The problem it describes is not.

Be clear about what each measure actually tells you. Look at additional measures or patterns when they could resolve an ambiguity that matters to the decision, rather than assuming one result captures the whole consumer response.

Every research measure answers a particular kind of question.

Preference tells you which option someone says they prefer. Choice records which option they select. Agreement tells you whether someone endorses a statement. Brand attribution tells you which brand someone connects with an asset.

Sometimes that is all you need.

If the decision really does come down to which of two packages consumers prefer, a preference measure may give you the evidence required. Adding more measures does not automatically make the study better.

Problems arise when we expect a result to answer questions it was never designed to answer.

A preferred package may not communicate the intended positioning particularly well. A well-liked asset may be strongly associated with another brand. Two claims can receive similar agreement scores while creating different associations.

The headline result can be correct and still leave something important unanswered.

It helps to distinguish between what a measure directly records and what we conclude from it.

If consumers prefer Package A to Package B, the research provides evidence about preference. It does not automatically establish that Package A will attract more attention on shelf, communicate the brand more clearly, increase purchase, or perform better in market.

Those may be reasonable questions to investigate, but they are different questions.

The distinction matters because research findings can become broader as they move from the data into a decision. “More people preferred this option” can easily become “this is the stronger option,” even when strength depends on several things the preference question did not measure.

A useful discipline is to ask: what did we actually measure, and how far does that evidence allow us to go?

Aggregate results can hide meaningful differences.

A 70% attribution score tells you how many people selected the intended brand. It does not tell you where the other 30% went.

If those incorrect responses are widely scattered, the asset may simply have a weak connection to any particular brand. If they consistently point to one competitor, you have learned something quite different.

The same principle applies elsewhere.

Two packages might have similar preference scores but communicate important attributes differently. Two messages might receive similar levels of agreement but produce different patterns across the audience.

Looking beneath the total does not mean searching through the data until something interesting appears. The useful patterns are the ones that could change how you understand the result or what you decide to do next.

Sometimes there is useful information in how readily a response is made.

In a timed task, two people might give the same answer while one responds almost immediately and the other takes considerably longer. Across a set of comparable responses, those differences can add information about how readily an association or judgement is made.

That evidence needs to be interpreted carefully.

A fast response is not automatically a better response. A slow response does not mean someone secretly disagrees with the answer they gave. Timing can also be affected by the task, the stimulus, reading difficulty, attention and other factors.

Response time is not a window into what someone “really” thinks. It is another piece of evidence about the response.

Whether it is useful depends on what you are trying to understand.

The same caution applies in the other direction.

Asking someone why they chose a package or reacted to a message can reveal things another measure cannot. They might point to a word, feature, colour or association that helps explain how they understood what they saw.

Those explanations are useful evidence. They are not necessarily a complete account of everything that produced the response.

That distinction matters because consumer research is sometimes framed as a choice between what people say and what they really think. Human responses are not that simple.

Stated answers, choices, associations, attribution patterns, response time and open-ended explanations can all tell us something useful. They tell us different things.

There is no need to keep adding measures until every possible part of the consumer response has been captured.

Doing that can make a study longer and the results harder to interpret without making the decision any easier.

An additional measure is useful when it helps resolve something that matters.

If two packages have similar preference scores, intended associations may help you understand an important difference between them. If two assets have similar attribution, the pattern of incorrect responses may reveal a competitive problem. If stated responses are similar, response time may tell you something about how readily those responses are made.

If that additional information would not change how you interpret the result or what you do next, you may not need it.

Different measures will not always point in the same direction.

Imagine that Package A is preferred more often, while Package B is more strongly associated with the premium positioning the brand is trying to communicate.

That does not necessarily mean something has gone wrong with the research, and it does not require deciding that one measure is more truthful than the other. The measures are telling you different things.

The disagreement may expose a trade-off that was not obvious from either result alone. If communicating a premium position is central to the decision, the association result matters. If preference is the priority, the preference result may matter more.

The research cannot resolve that strategic question on its own. What it can do is make the trade-off visible so the team can make the decision with better information.

There is rarely one measure that reveals what consumers “really” think.

What people say matters. What they choose matters. The associations they make may matter. How readily they respond may matter. Their explanations may matter.

When different measures point in the same direction, that can strengthen the overall evidence. When they do not, the disagreement can itself be informative.

It is not automatically a reason to decide that one measure is telling the truth and another is not.

Start with the result you need to understand. Be clear about what your measure can support. Then decide whether anything important is still unresolved.