Successful leadership has never been a choice between instinct and evidence. Experience, judgment, pattern recognition, customer knowledge, and data all contribute different information to a decision.
The challenge is knowing how much weight to give each one.
As businesses collect more information, it is tempting to assume that the answer to uncertainty is simply more data. Dashboards expand, teams track additional metrics, and leadership meetings become filled with numbers. Yet an organization can be extremely data-rich and still make poor decisions if it is measuring the wrong things, interpreting results without context, or mistaking correlation for explanation.
For women leading growing businesses, the strategic advantage therefore does not come from becoming “data-first” at the expense of judgment. It comes from becoming more deliberate about the relationship between evidence and decision-making.
Used well, data can challenge assumptions, expose problems earlier, sharpen resource allocation, and help leaders test whether what they believe is happening matches what is actually being observed. Used poorly, it can give weak decisions an appearance of certainty they do not deserve.
Move Beyond the Intuition Versus Data Debate
Leadership is often framed as a choice between trusting your instincts and following the numbers.
That is the wrong distinction.
Experience can help leaders recognize patterns quickly, understand nuances that are difficult to quantify, and notice when something does not feel consistent with what they know about a customer, team, or market. Those instincts can be valuable, particularly when decisions need to be made before complete information is available.
But experience can also reinforce assumptions. We naturally notice evidence that fits what we already believe, remember dramatic examples more readily than ordinary ones, and sometimes become overly confident because a similar decision worked in the past.
That is where data-driven decision-making becomes valuable. Data provides another source of evidence against which assumptions can be tested.
Suppose a leader believes a marketing campaign is performing exceptionally well because engagement is high and customers frequently mention it. Conversion data might reveal that the campaign is generating attention but very few sales. That does not mean the leader’s observation was wrong. The campaign may genuinely be engaging. It means engagement and commercial effectiveness are different questions.
Data sharpens judgment when it helps leaders distinguish between them.
Data Is Evidence, Not an Answer
One of the most important leadership shifts is to stop treating data as though it speaks for itself.
Every dataset reflects decisions about what was collected, how it was defined, who or what was included, when it was measured, and what was ignored.
If customer satisfaction is measured only among people who completed a survey, the results tell you about respondents, not necessarily every customer. If marketing performance is judged by last-click conversions, other interactions that influenced the purchase may disappear from the analysis. If employee engagement is measured immediately after a major organizational announcement, timing may influence the result.
Data can therefore be accurate while still being incomplete.
The leadership question is not simply, “What does the data say?” It is, “What exactly does this data allow us to conclude?”
That difference prevents numbers from becoming a substitute for thinking.
Start With the Decision, Not the Dashboard
Many businesses collect information because their software makes it available.
That is not the same as having useful data.
A dashboard may contain dozens of metrics while still failing to help a leader answer the question she is actually facing. Before deciding what to measure, define what you need to understand.
Are you trying to determine whether customer acquisition is becoming too expensive? Whether a new product is generating repeat purchases? Whether an operational bottleneck is limiting capacity? Whether one customer segment is more profitable than another?
Once the decision is clear, the information required becomes easier to identify.
This approach also reduces a common problem in data-heavy organizations: people searching through metrics until they find something interesting enough to discuss.
Interesting data is not necessarily decision-relevant data.
A strong measurement system begins with the strategic question and works backward.
Choose KPIs That Reflect the Outcome You Actually Want
Key Performance Indicators are valuable because they create focus. They can also become dangerous when the metric begins replacing the objective it was originally designed to represent.
A sales team measured solely by the number of deals closed may prioritize volume even when those customers have poor retention or low margins. A customer-service team measured only by call duration may shorten interactions at the expense of solving the problem properly. A marketing team rewarded for traffic may produce more visitors without increasing meaningful commercial activity.
The KPI can improve while the business outcome deteriorates.
Good leaders therefore ask what behavior a metric is likely to encourage.
Customer lifetime value may sometimes reveal more than the number of new customers acquired. Contribution margin may matter more than revenue growth. Repeat purchase behavior may tell a different story from initial conversion.
The right metric depends on the decision.
Rather than creating an enormous scorecard, focus attention on a manageable set of measures that reflect the outcomes the business genuinely cares about while keeping enough supporting information available to understand what is driving them.
Use Leading and Lagging Indicators Together
Many business metrics describe something that has already happened.
Revenue, profit, customer churn, employee turnover, and completed sales are important, but by the time they move significantly, the underlying cause may have been developing for months.
Leaders can strengthen their understanding by combining these lagging outcomes with earlier signals.
A decline in repeat website visits may precede customer churn. An increase in support requests may provide early evidence of product friction. A growing sales pipeline may indicate future revenue, although only if historical conversion patterns remain reasonably stable.
Leading indicators are not predictions in the strict sense. They are signals that may help identify change earlier.
The discipline is learning which signals actually have a meaningful relationship with the later outcome and continuing to test whether that relationship still holds.
Do Not Let an Average Hide the Story
Aggregated numbers can make a business appear healthier or weaker than individual segments really are.
Overall customer retention may look stable while retention among the company’s most profitable customers is falling. Average order value may rise because a small group of high-value customers is purchasing more while most customers spend less. Company-wide employee engagement may appear unchanged even though one team is struggling significantly.
Whenever an important metric moves, leaders should consider what lies beneath the average.
Segmenting data by customer type, product, geography, acquisition channel, team, tenure, or other relevant characteristics can reveal patterns that a single headline number conceals.
The purpose is not to slice data endlessly until something appears interesting. It is to understand whether an average is representing a reasonably consistent reality or combining groups experiencing something very different.
Use External Data to Challenge Internal Assumptions
Internal data tells you what is happening inside your organization. It does not automatically explain what is happening in the wider environment.
Competitor behavior, demographic changes, economic conditions, regulation, technology, customer expectations, and shifts in distribution can all affect business performance.
This is where external data can become strategically useful.
Imagine an e-commerce business considering its first physical location. Internal sales information might reveal where existing customers live and which products perform particularly well. A specialized commercial real estate data platform may provide additional information about properties, ownership, locations, or markets.
But none of those sources alone can identify a “perfect” location.
Foot traffic may be high but poorly matched to the target customer. Strong local demographics may coexist with expensive rents or intense competition. Historical property information may tell you little about how consumer behavior will change after the store opens.
The advantage comes from combining different forms of evidence and asking how they fit together.
Look for Disconfirming Evidence
Data is often used to prove an idea that leadership already prefers.
A stronger approach is to ask what information would make you reconsider the decision.
Suppose a company plans to enter a new market because customer research suggests strong demand. Instead of collecting additional evidence confirming the opportunity, leadership might examine competitor economics, customer acquisition costs, regulatory differences, local retention patterns, or evidence from businesses that attempted similar expansions unsuccessfully.
This does not mean trying to kill every idea.
It means deliberately looking for evidence that challenges the preferred interpretation before significant resources are committed.
A confident decision is stronger when it has survived serious questioning.
Understand Where the Data Came From
Data quality is a leadership issue, not merely a technical one.
If employees use different definitions for the same metric, reporting can become misleading. If customer records contain large gaps, segmentation may be unreliable. If one department counts revenue when a contract is signed and another counts it when payment arrives, leadership may believe two reports describe the same thing when they do not.
Before acting on an important figure, understand how it was produced.
Who owns the data? How often is it updated? Are definitions consistent? Is the sample representative? Are there missing values? Has the measurement method recently changed?
These questions may appear operational, but strategic decisions often depend on their answers.
A beautifully designed dashboard cannot rescue unreliable underlying information.
Numbers Do Not Explain Everything That Matters
Some information relevant to leadership is difficult to quantify.
Customers may stop buying because they no longer identify with the brand. Employees may leave because they do not trust their manager. A sales process may be underperforming because prospects repeatedly misunderstand the offer.
A metric can show that something changed. It may not explain why.
This is where qualitative evidence becomes valuable.
Customer interviews, sales conversations, employee feedback, complaint records, support transcripts, observational research, and direct experience can add context to numerical patterns.
If conversion drops, analytics may tell you where in the process customers leave. Speaking with customers may help explain why.
Strong leadership combines both forms of evidence rather than assuming that information becomes more valid simply because it can be expressed as a number.
Be Careful With Prediction
The original article presents predictive models as a way for leaders to understand what will happen next. Predictive analytics can certainly improve forecasting, but probability should not be confused with certainty.
Forecasts are built from historical relationships. Those relationships can weaken when conditions change.
Customer churn models may identify behaviors associated with customers who previously left. Demand forecasts may help businesses estimate inventory requirements based on historical sales, seasonality, promotions, and other variables. Neither system knows the future.
Unexpected competitors, economic changes, supply problems, unusual weather, new regulation, changing customer preferences, and product developments can all alter the pattern.
Leaders should therefore use forecasts as decision inputs rather than promises.
A forecast is particularly valuable when it allows you to prepare for several plausible outcomes rather than convincing yourself that one outcome has become inevitable.
Turn Forecasting Into Scenario Thinking
One useful way to avoid false precision is to consider ranges and scenarios rather than relying entirely on a single forecast.
Instead of asking whether demand next quarter will be exactly 20 percent higher, ask what the business would need to do if demand increases by 10 percent, 20 percent, or 30 percent.
What would happen to inventory?
Would staffing become constrained?
Would cash flow support the growth?
What if demand remains flat?
Scenario thinking turns prediction into preparedness.
This is a more practical strategic approach to data because it accepts uncertainty rather than attempting to eliminate it.
The strongest use of data is often not accurately predicting one future. It is helping the organization understand how different futures would affect its choices.
Create a Culture Where the Numbers Can Be Questioned
Data-driven organizations can unintentionally create a new form of hierarchy in which whoever owns the dashboard appears to own the truth.
That is risky.
Leadership teams should be able to question how a metric was constructed, challenge whether a result has been interpreted correctly, and raise information that does not fit the dominant narrative.
This does not mean dismissing evidence whenever it becomes inconvenient.
It means creating enough intellectual openness that both the data and the interpretation of it can be examined.
If employees believe questioning a senior leader’s preferred interpretation is career-limiting, an organization can have excellent analytics and still make poor decisions.
Data quality matters. So does the quality of the conversation around it.
Watch for Metrics That Change Behavior
Measurement does not merely observe organizations. It can change them.
Once employees know a number is being monitored, rewarded, or discussed constantly, they naturally pay more attention to it.
That can be useful when the metric aligns closely with the desired result. It can also produce unintended behavior.
If speed is rewarded, quality may suffer. If teams are pushed aggressively toward quarterly revenue, they may discount too heavily or pursue poor-fit customers. If managers are evaluated on employee-retention numbers without context, they may become reluctant to address persistent underperformance.
Leaders therefore need to look beyond whether a metric improved.
Ask what people did to make it improve.
Sometimes the behavior behind the number is more informative than the number itself.
Measure Decisions, Not Just Results
Business results are affected by factors outside leadership control.
A strong decision can produce a disappointing outcome because conditions changed. A weak decision can occasionally produce a good result because circumstances were favorable.
If leaders judge decision quality entirely by the eventual outcome, they may learn the wrong lesson.
After significant decisions, return to the reasoning.
What did we believe at the time?
Which assumptions were most important?
What evidence supported them?
What did we underestimate?
Which signals should we have taken more seriously?
This creates a record of decision quality rather than simply business performance.
Over time, it can help leadership teams improve how they interpret evidence, not merely how much they collect.
Know When There Is Enough Information to Act
Data can reduce uncertainty, but it rarely eliminates it.
There is always another report that could be commissioned, another customer segment that could be analyzed, another scenario that could be modeled.
At some point, the cost of obtaining additional certainty exceeds its practical value.
Leadership requires deciding when the evidence is sufficient.
The threshold should depend partly on the size and reversibility of the decision. A small marketing experiment can often be launched with less certainty because the cost of being wrong is limited. Acquiring another business, entering a heavily regulated market, or committing substantial capital deserves much more rigorous analysis.
The goal is not perfect information.
It is enough information for the level of risk you are taking.
Use Data to Improve Judgment, Not Avoid It
There is a reassuring quality to numbers. They can make a difficult decision feel more objective.
But leadership responsibility does not disappear because a spreadsheet supports the choice.
The data may reveal that one product line is less profitable. Leadership still needs to decide whether to close it, improve it, reposition it, or accept lower margins because it serves another strategic purpose.
Analytics may identify employees statistically more likely to leave. Leadership still needs to decide what interventions are appropriate, ethical, and useful.
Evidence informs judgment. It does not remove the need for judgment.
That is particularly important in strategic decisions, where the most consequential questions often involve values, trade-offs, uncertainty, and futures for which historical data does not yet exist.
The Strategic Advantage Is Better Questions
Data can help leaders see patterns earlier, challenge assumptions, allocate resources more carefully, and detect opportunities that would otherwise remain hidden.
But its greatest value may be more fundamental.
It forces better questions.
What are we actually trying to improve? What would success look like? Which evidence would change our mind? Are we measuring the outcome or merely a convenient proxy? Does this pattern hold across customer groups? What information is missing? Which assumptions does the forecast depend on?
Those questions create the real strategic advantage.
The most effective leaders do not replace intuition with data or instinct with algorithms. They become better at moving between evidence, experience, context, and judgment, knowing that each sees a different part of the problem.
Strong data leadership is therefore not about allowing numbers to make the decision.
It is about becoming harder to mislead, including by your own assumptions.
