Why More Data Does Not Always Improve Leadership Decisions

Leadership Judgement Series – Part Four

Businesses have never had more information available to support important leadership decisions.

A significant investment proposal might arrive at a leadership meeting supported by financial forecasts, historic performance, CRM data, market analysis, customer research and competitor intelligence. Increasingly, AI can add another layer of analysis before anyone enters the room.

In principle, better information should mean better leadership decisions, yet it doesn’t always work that way.

McKinsey’s research into organisational decision-making found that 57% of respondents believed their organisations consistently made high-quality decisions, while only 37% said decisions combined both high quality and appropriate speed. The same research found that 61% believed most of the time their organisations spent making decisions was used ineffectively. Even among C-suite respondents, the figure was 57%.

The problem is not that organisations have too much data, nor is it that leaders should rely more heavily on intuition.

The problem is that having more information does not tell a leadership team which information deserves to influence the decision. That requires judgement.

Data Is Not the Same as Evidence

Data, information and evidence are often treated as though they mean broadly the same thing.

For leadership decisions, there is an important distinction.

A business considering further investment in an established product, for example, may know exactly how much revenue it generated last year, its current gross margin, the size of the sales pipeline and its customer satisfaction score.

All of those figures may be completely accurate.

However, imagine that the real strategic question is whether customers will continue valuing that product in the same way over the next three years.

Historic revenue tells leadership what customers bought. Pipeline tells them what the sales organisation currently expects to sell. Satisfaction measures how existing customers responded to the questions the business chose to ask.

None necessarily tells leadership whether customer buying criteria are changing, whether an emerging competitor is reframing the market or whether something previously regarded as a differentiator is becoming standard.

The data isn’t wrong.

The danger lies in assuming that because something can be measured accurately, it must be the evidence that matters most to the decision.

More Information Can Increase Confidence Faster Than Accuracy

This is where having more information becomes particularly interesting.

Research by Claire Tsai, Joshua Klayman and Reid Hastie examined what happened when people were given increasing amounts of information relevant to a judgement. Across three studies, confidence increased more than accuracy as additional information was provided, creating a growing gap between how certain participants felt and how accurate their judgements actually were.

That research was not conducted in executive boardrooms, so it would be wrong to claim that it proves leadership teams make worse decisions whenever another dashboard appears.

It does, however, expose an important human tendency.

Feeling better informed and being better able to judge something are not necessarily the same thing.

Consider an investment case arriving at an executive meeting supported by twenty pages of analysis. Revenue has grown. Pipeline has increased. Opportunity numbers are rising. Sales activity is ahead of last year and the forecast remains positive.

Five different indicators appear to support the recommendation.

Look more closely, however, and they may not represent five independent pieces of evidence at all. Pipeline value, opportunity count, forecast revenue and sales activity may all ultimately depend on the same underlying belief: that customer demand will continue developing broadly as it has in the recent past.

If that assumption is wrong, several apparently independent metrics can become wrong together.

The volume of supporting information creates reassurance because the evidence appears to converge.

What may actually be converging is the same assumption expressed through several different measures.

That is one reason assumptions are so difficult to expose. They do not always sit obviously at the bottom of a spreadsheet waiting to be challenged. They can become embedded in the measures that leadership subsequently uses as evidence that the original assumption was correct.

More information can therefore make an interpretation feel increasingly objective without making it increasingly accurate.

When Noise Starts Looking Like Evidence

As organisations grow, almost everything becomes more measurable. There are more systems, more reports and more specialist functions producing analysis. Sales has its measures, marketing has another set, finance has its forecasts and operations has performance data covering increasingly detailed aspects of delivery.

Most of that information exists for good reason.

The leadership challenge is deciding what deserves attention.

Visibility and importance are not the same thing.

Imagine a leadership team reviewing sales performance. Pipeline coverage remains comfortably above target, but conversion is taking longer than expected. Because the pipeline is visible and measurable, attention naturally turns towards sales discipline, deal progression and forecasting accuracy.

Meanwhile, experienced salespeople are repeatedly noticing something more difficult to quantify. Buying committees are getting larger. Finance and procurement are entering conversations earlier. Customers who previously made relatively straightforward decisions now need greater internal consensus before committing.

The dashboard says the business has enough pipeline. The conversations suggest the way customers buy may be changing, but which deserves more weight?

There may not yet be enough qualitative evidence to prove that buying behaviour has changed across the market. Equally, treating the pipeline number as the more important evidence simply because it is easier to quantify risks missing a change that could eventually affect the entire go-to-market model.

This is why noise is not simply information we don’t need.

In leadership decisions, noise can be information that is useful, accurate and professionally presented but less important than something quieter competing for attention.

Accurate Data Can Still Describe Yesterday’s Reality

There is another problem with good data in that it can remain accurate long after it has stopped being a reliable guide to the future.

Leadership teams understandably place significant weight on experience. They know which channels have generated growth, which customer segments have been profitable, where pricing has held and which capabilities have historically differentiated the business.

That knowledge has value, but leadership judgement depends on more than knowing the numbers. It depends on knowing whether the numbers still describe the world in which the decision will operate.

A business may have compelling evidence that a particular sales channel has generated its highest returns for five years. That does not automatically mean another three years of investment should follow if customers are beginning to buy differently.

A proposition may have maintained strong margins because customers historically valued specialist expertise. That advantage becomes less useful as evidence for future pricing decisions if competitors have caught up or buyers now place greater weight on integration, implementation or ease of doing business.

An established customer segment may have exceptionally strong retention. That does not necessarily prove loyalty if customers are staying because switching remains difficult and new entrants are gradually reducing that friction.

The historical evidence can be completely correct in every case.

The key question is whether it remains current enough to justify the next leadership decisions.

That distinction between knowledge and knowing matters because businesses rarely move from “correct” to “wrong” overnight. Markets shift progressively. Customer expectations evolve. Competitors improve and technology changes the economics of previously successful approaches.

An organisation can therefore continue accumulating increasingly accurate evidence about a reality that is gradually disappearing.

The Commercial Cost of Asking for More Data Before Making Leadership Decisions

There are, of course, plenty of situations where the right leadership decision is to ask for more information. A major acquisition, capital investment or strategic change should not be approved simply because everyone is tired of analysing it.

There is a different problem, however, when the request for more information stops reducing meaningful uncertainty and starts postponing the moment when judgement has to be exercised.

“We need more data” is one of the easiest statements to agree with around an executive table.

Another forecast can be prepared. Additional customer research commissioned. More scenarios modelled. Another month’s performance reviewed before the decision returns to the agenda.

Each step appears diligent. Collectively, they can become expensive.

McKinsey’s decision-making research estimated that respondents spent an average of 37% of their working time making decisions, with more than half of that time perceived as ineffective. The researchers estimated that, for managers in an average Fortune 500 company, the wasted time could theoretically equate to more than 530,000 working days annually and around $250 million in labour costs.

For most businesses, the more important cost will not appear as a line marked “ineffective decision-making”.

It appears when an investment remains undecided while a competitor moves. When leadership continues funding a familiar priority because the emerging alternative has less historical evidence behind it. When another round of modelling creates greater precision around assumptions that nobody has properly challenged. Or when senior management attention becomes consumed by analysis that no longer materially changes the decision.

There is also an opportunity cost

Some of the earliest signals of change will never arrive neatly quantified. They may appear as recurring customer comments, unusual objections during sales conversations, frontline employees noticing new patterns or competitors behaving in ways the existing reporting model was never designed to capture.

Waiting until every weak signal becomes a statistically robust trend may produce better evidence. It may also mean learning about the change after the commercial opportunity to respond has narrowed considerably. At some point, therefore, the leadership question has to move from:

What else can we know?

to:

Do we know enough to exercise judgement?

Separate Fact, Interpretation and Assumption in Making Leadership Decisions

One practical way to improve a major decision is not to add another source of information, but to become more disciplined about the information already available.

Before approving an investment, changing strategy or making another consequential decision, separate three things that frequently become blurred together.

The distinction looks obvious when written down. Around a leadership table, it frequently isn’t.

An interpretation repeated often enough starts sounding like a fact. An assumption embedded in a forecast begins to inherit the authority of the spreadsheet around it.

Separating them forces a different conversation.

Does retention genuinely demonstrate loyalty, or could switching costs be contributing to it? What current customer evidence supports the interpretation? What would we expect to observe if the assumption were becoming less true?

Before commissioning another piece of analysis, five questions are particularly useful:

• What decision are we actually trying to make?
• Which evidence would genuinely change our view?
• What are we measuring because it matters rather than because it is available?
• Which part of our understanding is based on current evidence rather than inherited knowledge?
• What important evidence might not appear in the data at all?

Those questions do not reduce the need for information. They improve the discipline with which information is used.

Better Leadership Decisions Do Not Mean Less Data

None of this is an argument for leaders ignoring data, rejecting analytics or returning to instinct-led management.

Reliable evidence is fundamental to good judgement precisely because it helps leaders test assumptions against reality.

The distinction is that evidence informs a decision; it does not make the decision.

A useful contemporary example comes from a different leadership domain. Gartner reported in 2025 that increased efforts to give business leaders greater access to people data and improve data literacy had not, by themselves, led to better talent-related decisions. Its recommendation was to improve the relevance and salience of the insight presented to leaders.

The principle travels well beyond HR.

The objective should not be to give leaders the maximum amount of information available. It should be to ensure that the evidence reaching them is reliable, relevant, current and capable of improving their understanding.

Because data can be accurate, plentiful and professionally presented — and still be the wrong evidence on which to base the decision.

Leadership judgement begins when leaders recognise that distinction. So, before asking for another report, dashboard or piece of analysis, there is a harder question worth asking:

What evidence would genuinely cause us to make a different decision?

If the honest answer is “nothing”, the organisation may no longer have an information problem but it may have a judgement problem.

Next in the Leadership Judgement Series

Why Leadership Teams Need Constructive Friction to Make Better Decisions

Strong evidence only improves judgement when leadership teams are prepared to let it challenge what they already believe.

When disagreement disappears too quickly, consensus can become comfortable long before the quality of the decision has been properly tested.