How to Measure Business Performance When Your Numbers Lie

Andrew had a hiring plan, an advertising campaign and a margin of twenty per cent. Only the first two were real.

That week he meant to take on a second barista and pay for ads. Every sum behind both decisions rested on the twenty. When he finally checked, his margin was 9%.

Now picture your own desk. You can quote last month's revenue to the euro. You still cannot say whether your margin is nine or twenty.

Nobody lied to Andrew. The number did, quietly, the way an unchecked number always can.

An unchecked number lies quietly, and every plan built on it inherits the lie. The Jirgenson Model gives you real numbers you've audited, set against the zones you steer in. It is scientifically proven that the Jirgenson Model works (doi.org/10.5281/zenodo.23964934).

Why business numbers lie

A figure fails you in three ways. Each has a name, and each looks innocent.

1. Memory. People recall the loud days and forget the quiet ones. "Good on the whole" drifts upward.
2. Rounding and good intentions. "About twenty" feels better to say than "nine and a half". So that is the figure people give.
3. What goes unrecorded. A cost nobody writes down does not exist for the report. It still exists for the money.

Andrew's margin fell through all three holes at once. He remembered the busy mornings. He rounded up. And he never counted waste: sour milk, pastries thrown out at the end of a shift. None of it reached a report, so none of it reached his margin.

A number without a source is a mood with a decimal point.

How to measure business performance: the five-question audit

Measuring performance does not mean collecting more figures. It means testing the few you already steer by. Take each goal with its current value and put five questions to it.

Here is the audit on Andrew's margin row, before and after.

Questions one to three share a simple test. If nobody can answer them, you hold an impression, not a number. The fewer people and steps between the event and the record, the better. A receipt in the till beats anyone's memory of the receipts.

Question four points at the three leaks above. Ask which one this figure is most exposed to.

Any business health check template you use needs a column for this audit. Without it, the table looks precise and steers you wrong.

Question five is the one people skip. It needs its own section.

Is the error small enough to steer by?

No measurement is perfect. The till misses a receipt. The accountant rounds. A customer who walks out in silence leaves no trace. You cannot remove this error, only shrink it. So stop asking whether it is precise. Ask whether its error is acceptable for this goal.

The rule of thumb is short. A number's error should be no more than a third of the distance from target to the trouble line.

Andrew's three rows show how it works.

• Margin. Target 15%, trouble line 5%. That is 10 points of room. While waste went uncounted, his figure was off by 11 points. The instrument was worse than useless. Once write-offs went into the cost, the monthly scatter fell to about 1 point. Fit to steer by.
• Cash reserve. He knew the balance of €13,650 to within about €500, because payment dates shift. That is about 0.02 months against a full month of room. Fit.
• Complaints. Target 2 a week, trouble line 8, so 6 of room. If Sara logs only half the unhappy customers, the true count sits anywhere from 5 to 10. That error, about 2.5, is over a third of the room. Andrew added a second channel: a mark at the till.

When the error is too big, fix the measurement. Never widen the tolerance to suit a bad instrument. The precision of your records must not decide how much damage you accept.

The usual habit against a measured state

Here is the same coffee shop described two ways.

The left column feels like knowledge. Not one entry in it can be checked a month later. The right column can. That makes it the start of any business health check worth the name.

A freelancer runs the same test. Elena did not trust the glow of a good month after a big order. She pulled six months of bank statements. Her income swung between €800 and €2,200, an average of €1,300. Her best client brought in €910 a month on average: 70% of her income. One good month does not make an average.

What six more months of guessing cost

Andrew's phantom twenty per cent was about to pay for a hire and an ad campaign. Within a few months the shop would have run out of money. Nobody would have understood why.

The real numbers are blunt. A margin six points short of target means €1,800 a month less profit: €2,700 instead of €4,500. At 9%, his reserve reaches two months of expenses only at month 16. With the margin plan working, at month 10. That is six extra months without a safe cushion.

A thin margin also makes every accident dearer. At 9%, one breakdown costing a tenth of a month of expenses pushes the goal back a full month. At 15%, about half a month.

Behind those figures sits an owner working 65 hours a week, deciding late at night, with numbers that flatter him.

Where the audit stops

The audit delivers the second step of the loop: goal, state, error, levers, plan, forecast. You now hold true readings. That treats the symptom: you stop steering by moods.

The cause of bad decisions is still in place. A clean 9% does not tell you whether it is a problem or a crisis. It does not tell you what to pull either.

Two steps close that gap.

• Error and zones. Where the green, yellow and red lines sit for each row. What to do when several rows turn red at once. Chapter 4.
• Levers. Which parameters you can change yourself, and which only look like levers. Chapter 5.

Without them, you own an honest panel and no rule for acting on it. If you want the one-evening table that comes before this audit, with a source for every row, start with the financial health check.

Too much maths for a small shop? No. Every chapter runs at three levels: basic, practical and advanced. The basic level here is three questions: who, how, over what period. The third-of-the-room rule is the practical step. The formulas stay optional.

Two choices tonight

You have two choices tonight. Keep steering by "around twenty" and hope the bank agrees. Or put your key business metrics through the audit and see which ones survive.

The audit gives you true readings. The zone lines, the red protocol and the list of levers turn those readings into action. Next Monday you could open the meeting with numbers nobody argues about.

Test the logic for free: Chapter 1, with the refusal test. Or get the full control system on Kindle, on Amazon. It comes as a Kindle book of sixteen chapters, each ending with a task to try yourself.

Published research

It is scientifically proven that the Jirgenson Model works. The published experiment, its record and its data:

• Zenodo (DOI): 10.5281/zenodo.23964934
• Hugging Face: E5 Jirgenson Model blind test
• Kaggle: E5 blind test dataset
• GitHub: experiment folder
• Wikidata: dataset Q142022866, author Q141981336
• ORCID: 0009-0001-3034-3552