When the Numbers Lie: How Poor Research Can Lead to Bad Marketing Decisions

Data can make a decision look intelligent. That does not necessarily make the decision right.

There is something powerful about numbers in business. Put a percentage on a presentation, add a chart, produce a few graphs and suddenly an opinion can look like evidence.

But numbers do not automatically create truth.

Good research does.

This distinction matters because businesses increasingly describe themselves as data-driven. We hear about consumer insights, market intelligence, analytics, surveys, dashboards and customer data. Yet behind many strategic decisions are research processes that are poorly designed, poorly sampled or poorly interpreted.

And sometimes, the most dangerous research is not research that produces no numbers.

It is research that produces very convincing numbers that answer the wrong question.

Consider a company planning to launch a new product. It conducts a survey and discovers that 78% of respondents say they would consider buying it. The management team becomes excited. The presentation is prepared. The projections are developed. The marketing budget is approved.

But there is a problem.

Who were the 100 people surveyed?

Where were they recruited?

Were they existing customers?

Were they from the same geographic area?

Were they actually part of the target market?

Did the questionnaire influence their responses?

And perhaps most importantly, did “I would consider buying” actually mean “I will pay for this product”?

Those are very different questions.

This is where research becomes more than collecting responses. Research is about designing a process that gives decision-makers a credible understanding of reality.

A poorly designed questionnaire can produce beautifully organised nonsense.

A biased sample can produce statistically impressive findings that do not represent the wider market. A leading question can push respondents towards an answer the researcher already expects. A small sample can be treated as if it represents an entire population. Correlation can be interpreted as causation. And a research report can be presented with such confidence that nobody stops to question the assumptions behind it.

The problem is not the numbers.

The problem is what happened before the numbers appeared.

Marketing decisions are particularly vulnerable to this because consumers are complex. People do not always behave according to what they say they will do.

A consumer may tell you that price is the most important factor when choosing a product, but repeatedly purchase a more expensive alternative. Someone may say they prefer a local brand but continue buying an international competitor. A customer may claim that they want more features, yet choose the simpler product because it is easier to use.

This does not mean consumers are dishonest.

It means human behaviour is complicated.

People answer questions based on what they remember, what they believe about themselves, what they think sounds reasonable and what they expect they might do in the future. Actual behaviour takes place under different circumstances.

That is why marketing research should not stop at asking consumers what they think.

It should attempt to understand what they do, why they do it, and under what circumstances their behaviour changes.

This is particularly important in markets where businesses rely heavily on assumptions.

In many African markets, including Ghana, marketing decisions can still be influenced significantly by personal experience, management intuition, anecdotal feedback or what competitors appear to be doing. Experience has value, but experience is not the same thing as evidence.

A business owner might say, “My customers don't like this.”

The researcher should ask:

Which customers? How many? Compared with what? How do we know?

That is not challenging the manager for the sake of challenging them. It is protecting the decision.

Because research is ultimately about reducing uncertainty.

It cannot eliminate uncertainty. No research can perfectly predict what consumers will do next month, next year or after a competitor changes its strategy. But good research can make uncertainty more manageable.

Poor research does the opposite.

It creates false confidence.

And false confidence can be more expensive than admitting that you do not know.

Imagine spending millions launching a product because research suggested strong demand, only to discover that the research captured interest rather than purchasing intention. Imagine repositioning a brand because a small focus group disliked its identity, without understanding that the participants were not representative of the brand's core customers. Imagine changing your pricing strategy because a survey indicated that customers wanted lower prices, only to discover that the real problem was poor perceived value.

These are not simply research mistakes.

They are business mistakes created upstream by research mistakes.

This is why the quality of the research methodology matters.

Sampling matters.

Questionnaire design matters.

Measurement matters.

Data cleaning matters.

Analysis matters.

Interpretation matters.

Even the question itself matters.

Sometimes organisations rush into collecting data before deciding what they actually need to know. They ask dozens of questions because technology makes it easy to create a survey, then search through the responses hoping something useful will emerge.

That is backwards.

The business problem should determine the research question. The research question should determine the methodology. The methodology should determine the data required.

Not the other way around.

There is also a growing temptation to confuse more data with better research.

A company may have thousands of social media comments, website visits, customer transactions and survey responses. That sounds impressive. But volume does not automatically equal insight.

A million observations can still produce a poor decision if the underlying question is wrong.

The real competitive advantage is not simply having more data than everyone else.

It is being better at asking the right questions, collecting the right evidence and interpreting it intelligently.

This is where marketers and researchers need each other.

The researcher brings methodological discipline. The marketer brings commercial context. When those two perspectives work together, research becomes more than an academic exercise. It becomes a decision-making tool.

The ultimate purpose of marketing research is not to produce a 40-page report that sits on a manager's desk.

It is to help a business make a better decision.

Sometimes the research should tell you to proceed.

Sometimes it should tell you to change the strategy.

Sometimes it should tell you to abandon the idea entirely.

And sometimes, the most valuable conclusion is simply:

“We do not have enough evidence yet.”

That answer requires intellectual honesty.

In an environment where everyone wants quick answers, admitting that you need more evidence can feel uncomfortable. But responsible decision-making is not about appearing certain. It is about knowing why you believe what you believe.

This is why I am increasingly convinced that one of the most important questions a marketer can ask is not, “What do the numbers say?”

It is:

“Can I trust the process that produced these numbers?”

Because numbers can inform strategy, but they cannot rescue poor research design.

A beautiful dashboard cannot correct a biased sample.

A sophisticated statistical analysis cannot fix a badly constructed questionnaire.

Artificial intelligence cannot magically transform weak data into strong evidence.

And a confident presentation cannot turn an assumption into a fact.

The numbers do not lie by themselves. We make them lie when we ask the wrong questions, collect the wrong evidence or interpret the evidence beyond what it can actually tell us.

For businesses, the lesson is straightforward.

Do not become obsessed with having data.

Become obsessed with having credible evidence.

Because in marketing, a wrong decision supported by numbers is still a wrong decision.

And sometimes, the most dangerous number in the room is the one everyone believes without asking where it came from.