KPIs That Actually Matter: How to Pick Metrics That Do Not Lie

Every dashboard in the world is full of numbers. Very few of them tell you anything you did not already know, and some of them actively mislead you. The difference between a KPI that helps and a KPI that lies is not the number itself. It is the question the number answers.

A key performance indicator is only useful if it changes what you do. If a metric goes red and nobody changes their behaviour, it is decoration, not management. Here is how to choose metrics that earn their place.

1. Start With the Decision, Not the Number

The most common mistake is picking metrics because they are easy to measure or because everyone else tracks them. Reverse the logic: start with the decisions you actually make, then find the numbers that would make those decisions easier.

If you are deciding whether to hire another person in customer service, you need to know how many tickets arrive per week and how long they take to resolve. If you are deciding whether to raise prices, you need to know how many customers churn after a price change, not how many visits the website got. Every useful KPI is attached to a decision it informs.

2. Goodhart's Law Is Real

"When a measure becomes a target, it ceases to be a good measure." This is not a philosophical curiosity. It is the most reliable law in management.

The moment you start rewarding a number, people will optimise the number instead of the thing the number was supposed to represent. Sales teams hit call quotas by making calls nobody picks up. Support teams hit resolution targets by closing tickets without solving the problem. Delivery teams hit deadline targets by cutting quality.

The defence is not to stop measuring. It is to pair every metric with a guardrail metric that catches gaming. If you measure calls, also measure outcomes. If you measure speed, also measure quality. The second metric keeps the first one honest.

3. Leading Indicators Beat Lagging Indicators

Lagging indicators tell you what already happened. Revenue, profit, churn: all true, all in the past, all impossible to change. Leading indicators tell you what is about to happen. They are earlier, noisier, and infinitely more useful.

For a subscription business, churn is lagging; the number of customers who did not open your product this week is leading. For a shop, monthly revenue is lagging; footfall and basket size are leading. For a service business, client satisfaction surveys are lagging; the number of times you had to redo a job is leading.

The best KPI sets have a few of each. The lagging ones confirm you are winning. The leading ones tell you early enough to do something about it.

4. Fewer, Better Metrics

A dashboard with forty metrics is a dashboard with zero metrics. Attention is the scarcest resource in any business, and every number you add dilutes the ones that matter.

A good rule of thumb: each person should have no more than three to five KPIs, and they should fit on one page. If you cannot explain why a metric is on the page, remove it. If removing it would not change any decision, it was never a KPI, it was trivia.

5. Ratios Are Smarter Than Raw Counts

Raw counts are almost always misleading because they ignore scale. Revenue is up is a meaningless sentence until you know how many customers it took to get there. Ratios strip out the noise of size and show you the underlying behaviour.

The most useful ratios in a small business are simple: revenue per customer, cost per order, tickets per customer, hours per project, repeat purchase rate. Each one answers a question a raw number cannot: are we getting better at this, or just bigger?

6. Make the Numbers Visible and Weekly

A KPI that lives in a spreadsheet nobody opens is not a KPI, it is an archive. The metrics that change behaviour are the ones people see regularly and discuss as a team.

Weekly is the right cadence for most small businesses. Monthly is too slow to react, daily is too noisy for anything except the most operational numbers. A fifteen minute weekly review of five numbers, with one question per number (why is this better, why is this worse), beats a quarterly report nobody reads.

7. Kill Metrics That Outlive Their Purpose

Metrics have a lifecycle. A KPI that was useful when the business had ten customers may be actively misleading when it has a thousand. Review your KPI set quarterly and kill anything that no longer drives a decision.

The sign that a metric has died: it has not changed any decision for two quarters. Give it a dignified burial and replace it with something that will. A dashboard is a living tool, not a monument.

8. The Test of a Good KPI

Before you put any number on a dashboard, run it through three questions. Does it attach to a decision I actually make? Would I change my behaviour if it moved? And does it measure the real thing, or just a proxy that can be gamed?

If the answer to any of them is no, the number stays in the spreadsheet. The point of a KPI is not to look data-driven. It is to be better at deciding, and a handful of honest numbers will always beat a wall of impressive ones.

Tags

#business #operations #management