The Data Is Clear. So Why Don't People Change Their Minds?

Imagine spending hours analysing a problem.

You check the data.

You test different explanations.

You find a pattern.

The evidence is strong enough to support a recommendation.

You present it.

And the response is:

"I'm still not convinced."

At first, that might seem frustrating.

If the evidence is clear, why wouldn't someone change their mind?

But perhaps that's the wrong question.

Because people don't stop being human just because there is a dashboard in front of them.

Data doesn't arrive in an empty room

This is something I find particularly interesting because my background isn't only in data. I also studied Management Psychology and spent years working with people in banking, education and my own business.

And one thing those experiences have taught me is that decisions are rarely made in a vacuum.

By the time an analyst presents a finding, the person receiving it may already have an opinion about the problem.

They may have spent months supporting a particular strategy.

They may have designed the process being analysed.

They may have defended a previous decision to senior management.

They may even have their own performance measured against the outcome.

So when new data challenges that position, they're not simply receiving information.

Sometimes they're being asked to reconsider something they are already invested in.

And that's a very different situation.

Imagine a simple example

A customer support manager believes slow response times are caused by understaffing.

It makes sense from their perspective.

The team feels busy. Employees complain about workload. There always seems to be another ticket waiting.

So the proposed solution is obvious:

We need more people.

Then the data is analysed.

And it shows something different.

Ticket volumes haven't increased significantly.

Instead, certain categories of tickets are repeatedly being transferred between agents. Some tickets are being reopened. Workload distribution is uneven, and one particular part of the process is creating delays.

Suddenly, the evidence suggests that recruitment might not be the first problem to solve.

The process might be.

From an analytical perspective, that's useful.

From the manager's perspective, it might be much more complicated.

They may have spent months arguing for additional headcount.

Perhaps they've already told their team that staffing is the problem.

Maybe they've raised it repeatedly with senior management.

Now someone arrives with a dashboard and effectively says:

"Actually, that might not be the main issue."

The numbers haven't changed.

But the context in which those numbers are received changes everything.

We naturally look for evidence that fits what we already believe

Psychology has a name for one part of this: confirmation bias.

We tend to notice, interpret and remember information in ways that support our existing beliefs.

And analysts aren't magically immune to this either.

If I begin an analysis convinced that discounts are damaging profitability, for example, I need to be careful that I'm not simply looking harder for evidence that confirms my theory.

The same applies to the person receiving the analysis.

If someone has believed for months that understaffing is causing a problem, evidence supporting that explanation may feel immediately convincing.

Evidence challenging it may be examined much more critically.

That's not necessarily dishonesty.

It's human behaviour.

Then there is the problem of investment

Now imagine that the business has already invested £100,000 in a new process or system.

Six months later, the analysis suggests that it isn't delivering the expected results.

Technically, the historical investment shouldn't determine whether continuing with the same approach is the best decision.

But psychologically?

Walking away is difficult.

Time has been invested.

Money has been invested.

People may have put their reputation behind the decision.

The analysis isn't simply asking:

"Does this work?"

It may also feel like it's asking:

"Was the previous decision wrong?"

Those are not the same question, even if a spreadsheet treats them as though they are.

Being right isn't always enough

I think this is where analytical work becomes particularly interesting.

You can be technically correct and still fail to influence a decision.

A perfectly accurate analysis presented without understanding the people receiving it may go nowhere.

That doesn't mean changing the numbers to make people comfortable.

And it definitely doesn't mean telling stakeholders what they want to hear.

It means recognising that presenting evidence and communicating evidence are not exactly the same thing.

Saying:

"Your assumption was wrong."

creates a very different conversation from:

"The data suggests there may be another factor contributing to the problem. Can we look at this part of the process?"

The underlying finding hasn't changed.

The possibility of having a productive conversation probably has.

Analysts have biases too

There's another side to this that I think is easy to forget.

It's tempting to position the analyst as the rational person in the room, surrounded by emotional stakeholders who refuse to listen to the evidence.

I don't think that's realistic.

Analysts bring assumptions too.

We choose which questions to investigate.

We decide which metrics matter.

We make decisions about exclusions, categories and definitions.

We form hypotheses.

And sometimes we become attached to our own interpretation because we've spent hours building it.

The fact that an opinion came from an analysis doesn't automatically make the person presenting it completely objective.

Sometimes a stakeholder challenging an analysis isn't resisting the data.

Sometimes they've noticed context the analyst missed.

That's why questions matter in both directions.

So what should an analyst do?

For me, the answer isn't to become an amateur psychologist in every meeting.

It's much simpler.

Be curious about the reaction as well as the result.

If someone disagrees with a finding, I don't think the first assumption should be that they "don't understand the data".

Maybe they know something I don't.

Maybe the metric doesn't capture an important part of the process.

Maybe the analysis challenges something they've believed for a long time.

Or maybe the evidence really is strong and they simply need time to reconsider their position.

Understanding which of those situations you're dealing with matters.

Because the goal isn't to win an argument with a dashboard.

The goal is to help the business make a better decision.

Final thought

We often talk about data as though presenting enough evidence should automatically lead to the rational decision.

Real organisations don't work like that.

They are made up of people with experience, assumptions, incentives, previous decisions and opinions of their own.

Data can challenge those things.

But it doesn't make them disappear.

And perhaps that's why good analysis requires more than finding the right answer.

It also requires understanding the human being who has to do something with it.