What a service business should look at every week

Six numbers, looked at every week, will tell an owner more about their business than any monthly report. The test for whether something belongs on a weekly list is not importance, it is whether the number can meaningfully change in seven days and whether you would do something differently if it moved. Almost everything that gets reported weekly fails both.

The two questions that build the list

Most reporting is assembled by asking what data is available, which produces a page of things that are true and useless. The better construction is two filters applied in order, and they eliminate most candidates immediately.

First: can this move in a week. Revenue in a service business with a two-month sales cycle cannot, so watching it weekly produces noise that people then explain. Second: would a bad reading change what you do on Monday. If the honest answer is that you would note it and carry on, it is not a weekly number, whatever else it is.

Those two filters also explain why the useful weekly numbers are almost all leading rather than lagging. Lagging figures, revenue, profit, margin, are the ones owners care most about and the ones that move too slowly and too noisily to be read on a seven-day cycle. They belong on the monthly list, where they are excellent.

The six

1. Enquiries that reached a human

   Every route, counted, including the calls that were missed and then returned. This is the top of everything and it is the number most businesses cannot produce at all, because missed calls leave no record anywhere except the carrier's log.

2. Quotes issued, and the hours from enquiry to issue

   Two figures from one measurement. The count tells you whether enquiries are being converted into a priced proposal at all, and the elapsed time is the single most actionable operational number in the business.

3. Quote acceptance rate, on a rolling basis

   Rolling over the last eight or twelve weeks rather than weekly, because a single week's sample in a small business is too small to mean anything. Rolling smooths that without losing responsiveness.

4. Open quotes with no contact in seven days

   Not a performance metric, a work list. Every item on it is priced work sitting untouched, and reviewing it weekly is the review. This one is unusual in that reading the number and doing the job are the same activity.

5. Booked capacity for the next two weeks

   The percentage of available technician hours already committed. It is the earliest possible warning of a quiet month, arriving while there is still time to release reactivation work rather than after the month has happened.

6. Jobs completed but not yet invoiced

   Work finished, cash not requested. In a business where job sheets come back on paper this figure is routinely larger than the owner assumes, and every day it sits there is a day the company is financing its customers for free.

Notice what is not on the list. No revenue, no profit, no margin, no headcount, no marketing spend, no website traffic. All of those matter and none of them meets both tests, so they belong in the monthly review where they can be read against a sample large enough to mean something.

Five of the six are things that have not happened yet. The one that already happened, uninvoiced work, is the one you can fix this afternoon.

The fifth is the one owners most often argue with, on the grounds that they already know how busy they are. They usually do, for this week. Two weeks out is different, and the gap between an intuition about capacity and a percentage is where the quiet month gets caught early enough to do something about.

How to actually run it

The mechanics decide whether this survives past the third week, and most attempts die because the collection is too expensive relative to the value.

• Fifteen minutes, same time every week, same person, whether or not anything looks interesting
• One page or one screen, six numbers, this week and the previous four, so a trend is visible without anyone building a chart
• The person who collects it is the person who can act on it, because handing it to somebody else adds a translation step and a delay
• Write one sentence a week on what changed and why, and keep the file, since after three months that log is more valuable than the numbers
• Do not add a seventh number without removing one, because the list grows to fill the meeting and then the meeting gets skipped
• Compare only against your own previous weeks, never against a benchmark from somebody else's business

The weekly written sentence is what gets dropped, and it is the highest-value part. Numbers without a written explanation of what changed are unreadable six months later, and the log of explanations is what turns a reporting habit into an actual understanding of how the business behaves. It also catches the self-deception, because writing down last week's reason makes it obvious when the same reason has now appeared four times.

The trap of the number that always looks fine

One failure mode is worth naming because it is invisible while it happens. A metric that has never once prompted an action is not evidence that everything is well. It is usually evidence that the metric is measuring something the business cannot influence, or that the threshold for concern was never defined.

The fix is to set the trigger before you start. For each of the six, write down the reading that would make you do something and what that something is. Without those thresholds the weekly review becomes an act of observation rather than a decision point, and observation is a comfortable habit that changes nothing.

Which means the sheet has four columns rather than one, and the last two are the ones that make it a review instead of a report. The thresholds below are written as shapes rather than as figures, because the numbers have to be yours: a benchmark from somebody else's business is the one thing on this page that will actively mislead you.

Filling in the third column is the hard part and it is the part worth doing, because a threshold you set in advance is a decision and a threshold you set while looking at a bad number is a rationalisation. Doing it before the first review also surfaces something uncomfortable and useful: for one or two of the six you will find you have no idea what a bad reading would even be, which means you have never had a view on that part of the business at all.

The related discipline is a review that is allowed to be boring. Most weeks nothing will have moved and the correct outcome is a fifteen-minute meeting in which nobody does anything. That is the system working. The temptation to find significance in ordinary variation is what makes people abandon weekly reviews, usually after chasing three phantom trends in a row.

Where this fits

These six are not a general management dashboard, they are the instrumentation for a specific set of problems: enquiries dying before they become work, quotes going out slowly, capacity going unnoticed, and cash sitting in a van. That is deliberate, and it is because those are the failures that show up in almost every unimproved service business.

They also make every later change legible. Any improvement you install will be defended or attacked with anecdote unless there is a before, and eight weeks of these six numbers is the cheapest before you will ever get. That is the same argument for measuring first that applies when taking over a business at all, which is in the first ninety days after you own it.

Which layer each number belongs to, and why fixing them out of order wastes money, is in the four-layer growth engine. The first number on the list is almost always the one that is broken, and that is the unanswered phone. If you would rather see the whole diagnostic laid out at once, it is on the system page.