top of page
Search

The Most Dangerous Number in Your Dashboard Is the Precise One

Writer: Nakel Nikiema
Nakel Nikiema
22 hours ago
5 min read

Picture two forecasts on the same slide. One says revenue will grow 16.3% next quarter. The other says growth will land somewhere between 15 and 20%. Which one would your leadership team trust?

Most rooms pick the first one. It sounds measured. It sounds like someone did the work. Yet the second forecast may be the more honest of the two, because it admits what nobody in the room actually knows.

That instinct, trusting the number that looks most exact, is one of the quietest ways data leads smart leaders in the wrong direction. I wrote The Dark Side of BI about this problem and the others like it. This post walks through three of the most common traps and the questions that get you out of them.

Why precision feels like proof

From the time we're children, numbers are presented to us as settled facts. The multiplication table never lies. A geometry problem solved correctly stays solved. So when a number shows up with decimals attached, we read the decimals as evidence of rigor: data was collected, models were run, someone checked.

Sometimes that's true. But precision is a format, not a guarantee. A figure can be calculated to two decimal places from data that is incomplete, out of date, poorly defined or simply measuring the wrong thing. The decimals don't tell you any of that. They make the number harder to question.

There's a social side, too. In many organizations, challenging a well-formatted dashboard in a meeting feels awkward, even career-limiting. Nobody wants to look unprepared or "anti-data." So people nod, the number moves into the plan, and the assumptions underneath it never get a hearing.

Trap one: the average that describes nobody

Ask for a quick summary of almost anything (revenue, satisfaction, website traffic) and you'll usually get an average. Averages are useful. They're also one of the easiest ways to hide what matters.

Take a small agency with five clients paying about $5,000 per project and one large client paying $100,000. The average payment works out to more than $20,000. Mathematically correct, and completely misleading about what a "typical" client looks like. Plan hiring or pricing around that number and you'll be planning for a business you don't have.

The same thing happens with daily traffic that swings between quiet and busy days, with satisfaction scores that blend delighted customers and angry ones, and with performance numbers that combine very different teams. The average is accurate. The picture it paints isn't.

The fix: whenever someone gives you an average, ask to see the spread. What does the range look like? Is one large value pulling the number up or down? Would the median tell a different story?

Trap two: the metric that stopped measuring the goal

Every metric is built. A customer satisfaction score comes from a survey, a set of questions, a sampling method and a rule for adding it all up. A conversion rate depends on what counts in the numerator, what counts in the denominator, and how long the window is. At each of those steps, someone makes a choice.

Most of those choices are reasonable. But under the pressure of a quarterly review or a board meeting, the choices can start bending toward the story people want to tell. A time window gets shifted. A definition gets quietly reworded. A comparison is anchored to whichever baseline looks best. Two companies can add the same number of customers in a quarter. One reports explosive month-on-month growth, the other a modest year-on-year improvement. Both statements are true, and they leave very different impressions.

Over time, a metric that was meant to track a goal can replace the goal. Teams start managing the number instead of the outcome it was supposed to represent.

The fix: keep a written definition next to every key measure, and make any change to that definition visible. When a number improves, ask whether the underlying reality improved, or only the way it's counted.

Trap three: the dashboard that decides for you

Dashboards are meant to make information easier to act on. They can also distort it. A chart's axis can start at a convenient point and make a small change look dramatic. A time range can be picked to show the best stretch. Colors can signal "good" or "bad" before anyone reads the numbers. Uncertainty, the honest range around a forecast, is usually the first thing left off the screen.

None of this needs bad intent. Often it's a design habit. But the effect is the same: the dashboard frames the decision before the leader does.

The fix: ask what the dashboard isn't showing you. Where does the axis start? Why this time window? Where's the range of uncertainty?

Five questions to ask before any number shapes a decision

You don't need to become a data scientist to protect your decisions. You need a habit of asking a few plain questions, every time:

  1. How did this number come about? Where did the data come from, and who decided how to calculate it?

  2. What is it being compared against? Last month, last year, a target, a competitor? Would a different comparison change the story?

  3. What does the spread look like? Is the average hiding very different groups or one big outlier?

  4. What's the honest range? If this is a forecast, what are the reasonable high and low ends?

  5. What would change my mind? If you can't name any evidence that would change the decision, the data isn't driving it. Something else is.

These questions aren't an argument against analytics. They're the case for better analytics. Organizations that ask them get more out of their data, not less, because their numbers have to earn the trust they're given.

Go deeper

The Dark Side of BI: When Data Misleads Decision-Makers covers these traps and more: garbage in, garbage out; correlation mistaken for causation; overfitting and false patterns in predictive models; cognitive bias in analysis; the ethics of data communication; and the systems organizations can build to reduce the risk. It's written for leaders who rely on numbers and want to rely on them wisely.

If your team is making high-stakes calls from dashboards nobody fully trusts, that's the kind of fog I help leaders cut through. Learn about my business consulting at https://www.nakelnikiema.com/business-consulting, or write to me at consulting@nakelnikiema.com.

Nakel Nikiema is the Founder & President of Nakel & Associates and brings 30+ years of international leadership and consulting experience. He is the author of The Dark Side of BI and host of The Intelligence Edge podcast.

 
 
 

Comments

bottom of page