What it means
Not every metric is a key performance indicator. Metrics are the wide pool of things a business could measure, while KPIs are the small handful deliberately chosen because they track progress against a specific goal.
The most useful distinction is between lagging and leading metrics. Lagging metrics such as quarterly revenue tell you what has already happened, while leading metrics such as qualified pipeline or trial sign-ups indicate what is likely to happen next and give you time to react.
Metrics need a written definition or they drift. Two teams reporting active users can produce completely different numbers because one counts anyone who logged in and the other counts anyone who completed an action, and the argument that follows wastes more time than the metric saves.
Ratios usually beat raw counts. Revenue tells you little without knowing the cost of getting it, which is why acquisition cost, payback period, gross margin and retention rate carry more decision-making weight than headline totals.
Every metric can be gamed, so the choice of what to measure changes behaviour. Rewarding call volume produces short calls, rewarding closed tickets produces prematurely closed tickets, and sensible dashboards pair each headline number with a quality counterweight.
Good discipline is to keep the reported set small and stable. Five to ten metrics that everyone understands beat a forty-line dashboard nobody reads, and changing definitions mid-year destroys the comparability that made the numbers useful in the first place.
In practice
Real-world examples
Example 1
A logistics firm tracks on-time delivery percentage alongside cost per drop. When the operations team pushes cost per drop down by consolidating routes, on-time performance falls, and the pairing of the two metrics reveals the trade-off before customers complain.
Example 2
A recruitment agency replaces its headline placements-per-consultant metric with gross profit per consultant. Consultants stop chasing low-fee roles simply to hit a count and start working the vacancies that genuinely fund the business.
Example 3
A hospital ward reports average waiting time and immediately looks good, until someone adds the 90th percentile. The average was hiding a small group waiting far longer, and the ward redesigns triage around that tail rather than the mean.
Think of it
Metrics are like the dashboard dials in a car. Speedometers, fuel gauges, and warning lights tell you exactly how the vehicle is performing so you can drive safely.
Formula
Calculation
Most business metrics are simple ratios of the form outcome / input. Two of the most widely used are customer acquisition cost, which is sales and marketing spend divided by new customers won, and the payback period, which is acquisition cost divided by monthly gross profit per customer.
A subscription software company attracts 40,000 website visitors in a month. Of those, 2,000 start a free trial, a visitor-to-trial rate of 2,000 / 40,000 = 5%, and 300 convert to paying plans, a trial-to-paid rate of 300 / 2,000 = 15%. The overall conversion rate is therefore 300 / 40,000 = 0.75%.
Sales and marketing cost $180,000 that month, so customer acquisition cost = $180,000 / 300 = $600. Each customer pays $80 a month at a gross margin of 75%, giving monthly gross profit of $80 x 0.75 = $60 per customer.
Payback period = $600 / $60 = 10 months. If the trial-to-paid rate improved from 15% to 20%, the same 2,000 trials would produce 2,000 x 20% = 400 customers on identical spend, cutting acquisition cost to $180,000 / 400 = $450 and payback to $450 / $60 = 7.5 months.Case study
Seen in the real world
Pellworth Analytics is an invented software business used here for illustration. Its board pack ran to 38 metrics, and the monthly meeting spent most of its time reconciling numbers that different teams had calculated in different ways.
The finance lead cut the pack to seven metrics, each with a one-line written definition and a single named owner: new monthly recurring revenue, net revenue retention, gross margin, acquisition cost payback, active accounts, cash runway and support response time. Everything else moved into a reference appendix that nobody was obliged to read.
In this fictional case the meeting shortened from three hours to 75 minutes and, far more importantly, the team spotted a rise in acquisition cost payback from 10 months to 14 months two quarters earlier than it would have done before. The number had always existed; it had simply been buried among 37 others.
Watch out
Common mistakes
- Measuring everything that happens to be easy to count. A large dashboard of convenient numbers usually crowds out the few that would actually change a decision.
- Leaving metrics undefined. Without a written rule for what counts, two teams will report different figures for the same thing and trust in the whole pack collapses.
- Judging performance on a single number. Almost every metric can be improved at the expense of something else, so headline figures need a counterweight.
Questions
People also ask
What is the difference between a metric and a KPI?
Every KPI is a metric, but a KPI is one of the few metrics deliberately chosen to track progress against a stated objective.
How many metrics should a leadership team review?
Most work well with five to ten headline numbers reviewed monthly, with deeper operational detail held at team level.
Should metrics change over time?
Definitions should stay stable so trends remain comparable, but the chosen set is worth revisiting annually as business priorities shift.