Operational KPIs for a Small Business: Seven Numbers That Drive Decisions

A good metric is a number somebody does something with. Everything else is a report. A small business does not need twenty metrics but seven - and each of them needs a ready answer to one question: what will we do if it moves.

The seven metrics

Rows three and seven are the ones most businesses do not measure, and both have a direct effect on money.

Why a metric with no action is worthless

Because the purpose of measuring is not to know but to decide. A number nobody knows what to do with becomes, within a month, a dial people glance at and nod to, and after two months nobody opens it.

So write one sentence beside each metric: if it passes X, we do Y. For example: if response time exceeds two hours, somebody else joins the queue. If debt over 30 days passes a certain amount, supply to those customers stops. That sentence is what turns measurement into management.

Trend, not number

A single number is nearly always meaningless. A 30% close rate is good or bad only relative to last month and to the type of enquiries. So every metric should appear next to its previous value, and ideally next to a three-month average.

More important is not reacting to every fluctuation. In a small business numbers jump week to week because of one large deal or a holiday. Reacting to every jump produces unnecessary changes; what justifies action is a trend across three consecutive points.

In Israel it is also worth marking in advance the weeks that are not comparable - the autumn holidays, Passover, and periods when a share of customers are simply unavailable. Comparing a holiday week against an ordinary one produces wrong conclusions in both directions, and the simple fix is comparing them against the same period last year rather than against last week.

Where the data comes from

In most small businesses four sources cover all seven metrics: the invoicing system, the CRM or the spreadsheet standing in for it, the order system, and a short manual record. What matters is that each metric has exactly one source - two sources for one metric means people will argue about the number instead of discussing the conclusion.

So record three things beside each metric: where it comes from, who owns it, and how often it updates. That is also what makes a standing report possible instead of reassembling it each time, as covered in a one-page weekly management report.

What not to measure in a small business

• Metrics that take an hour to collect - they will not be collected.
• Metrics with no owner - nobody will notice when they move.
• Averages over few cases - in a business with five deals a month, an average is noise.
• Metrics about people rather than processes - they change behaviour rather than outcomes.
• Anything measured only because it can be measured.

The fourth matters most culturally: the moment a metric is perceived as an employee assessment, the data starts being managed rather than reported - and that is very hard to correct afterwards.

How to start

1. Pick three metrics only for the first month, not seven.
2. Write the action sentence for each - if X then Y.
3. Measure for four weeks and see whether the data is even available.
4. Add two more only once the first three collect themselves.
5. Drop any metric that led to no decision in a quarter.

The last step is almost always skipped, and it is what keeps the list short. A metric that produced no decision in a whole quarter is not a metric - it is a habit.

How do you know a metric is working?

Three simple checks after a quarter. Is it collected without anyone chasing? A metric needing a weekly reminder will vanish in the first busy month. Does anyone notice when it moves? If you can change the number in the spreadsheet and nobody comments, it is not being read. And did it change anything? At least one decision per quarter - a change in practice, a hire, stopping supply to a customer, a price change.

When a metric passes all three, it earns its time. When it fails one, drop it and add another - a list of five metrics that work is worth more than twelve well-formatted ones nobody looks at.

What changes as the business grows

The metrics themselves change little, but who reads them changes a lot. In a business of two or three, the owner sees everything anyway, so measurement mainly surfaces trends that are hard to see day to day. In a business of seven to ten, metrics become the only way to know what is happening in areas you no longer touch.

That is also the point where segmentation earns its place: not just average response time but by lead source or by rep; not just error rate but by product type. Segmentation produces the information you can act on, and its logic is the same as in where did this lead come from.

Sources

• How to improve business efficiency
• Israel Tax Authority - information for businesses