How to measure backlog in weeks, not dollars

"We've got about $600,000 on the books." It is the standard answer, and it is not usable information.

Six hundred thousand dollars is fourteen weeks of work for one crew and four weeks for three. Without capacity in the sentence, the number means nothing.

The short answer: divide signed work not yet started by what a crew genuinely produces in a week. Under four weeks and you start selling from desperation.

The calculation

Backlog in weeks = signed work not yet started ÷ weekly production capacity

Signed work not yet started. Contract value of work you have won and not begun. Exclude anything in progress, exclude verbal agreements, exclude quotes outstanding. Only signed, only not started.

Weekly production capacity. What you actually put in place in a normal week, in revenue terms. Not what you could do in a perfect week. Take the last three months of completed revenue and divide by the number of weeks.

That second number is where most people flatter themselves. Use the real one.

Six weeks is a specific, actionable fact. $600,000 is trivia.

What the thresholds mean

Like cash, the bands matter because each one changes behaviour.

Under four weeks is the number that costs you money, and it costs you through pricing rather than idleness.

When the schedule is nearly empty, a contractor takes a job at 14% because an empty crew costs money and 14% is better than nothing. That logic is correct in isolation and ruinous as a pattern. Thin backlog does not just mean less work — it means worse-priced work, which produces less margin, which makes the next month tighter.

This is why backlog is a leading indicator. It tells you about your pricing power four to six weeks before your margin report tells you about your prices.

The other end of the problem

Very long backlog looks like success and carries its own costs.

Past about sixteen weeks, your lead time is long enough that customers start going elsewhere — particularly for anything urgent, which is usually your best-margin work. You are also locking in today's prices against costs that will be several months older when the work is done, which in a period of material inflation is a real exposure.

If backlog is running long, the answer is usually not more crew. It is higher prices. Scarce capacity that is priced as though it were abundant is the most common unforced error in the trades.

Measuring capacity honestly

This is the harder half, and it is where the calculation goes wrong.

Do not use your best month. Use a rolling three-month average of completed revenue. It absorbs weather, holidays, and the week everything went wrong.

Adjust for crew changes. If you have added a crew since that average was set, scale it. If someone left, scale it down.

Adjust for job mix if it is genuinely different. A quarter of large simple jobs produces more revenue per crew-week than a quarter of small complex ones. If your mix has shifted meaningfully, your capacity number has too.

For most trades businesses a crew-week of production lands somewhere fairly stable, and once you know yours the conversion from dollars to weeks takes ten seconds.

Reading it alongside the other numbers

Backlog on its own is one dimension. Read with margin, it tells you considerably more.

Backlog falling, margin holding. Demand is softening but you are keeping discipline. Sales problem. Fix the top of the funnel.

Backlog rising, margin falling. You are buying work. The most dangerous combination, because it looks like growth on every report and is quietly destroying the year.

Backlog rising, margin rising. You have found something that works. Do more of it, and consider whether you are still under-priced.

Backlog falling, margin falling. Both things wrong at once. This is the point where the business needs a decision rather than more effort.

That second row is the one to watch for. It is exactly what happens when a contractor responds to a soft month by dropping prices, wins more work, and reports a record quarter — while making less money than the quarter before.

Using it to make decisions

Hiring. Do not hire off one busy month. Hire when backlog has been above ten to twelve weeks consistently for a quarter, because a new crew member costs money immediately and produces later.

Pricing. Backlog above twelve weeks is permission to raise prices on the next bid. Backlog under four is a signal to increase sales activity, not to drop prices — though dropping prices is what almost everyone does.

Marketing spend. The time to spend on lead generation is when backlog is healthy and falling, not when it is already empty. Lead generation has a lag of weeks to months. Turning it on when the schedule is bare means the leads arrive after the crisis.

Turning work down. With twelve weeks of backlog, saying no to a bad job costs you nothing. With three weeks, saying no feels impossible. Which is why the work you accept at three weeks is worse than the work you accept at twelve — and why backlog management is really margin management.

Track it weekly

One line, once a week: signed work not started, divided by weekly capacity.

The trend matters more than the level. A backlog that has gone from eleven weeks to seven over a month is telling you something urgent, even though seven weeks still sounds comfortable. By the time it reaches four, the pricing damage has already started.

It is the earliest warning the business gives you — earlier than cash, earlier than margin, earlier than the P&L. Which makes it worth ten seconds on a Monday morning.

Measure it by crew, not in total

A single company-wide backlog number hides the problem it exists to reveal.

A business with two crews — one doing full replacements, one doing repairs and service — might report ten weeks of backlog and feel comfortable. Split it and the picture changes:

The combined figure of six weeks says "watch it." The real situation is that one crew is comfortable and the other runs out of work inside a fortnight — and next week somebody will take a service job at 12% margin to keep them busy, because the alternative is sending people home.

Backlog only means something at the level where the scheduling decision gets made. If you schedule crews separately, measure them separately.

The same applies to anything that constrains you independently: a single certified installer, one piece of equipment, one estimator. If work cannot start without that resource, it has its own backlog and its own thresholds.

Two things follow.

Sales effort should be aimed, not general. Knowing the service crew is at two weeks tells you exactly what to sell this month. A combined number tells you nothing actionable.

Work that can move between crews is worth more than it looks. A job either crew can run is not just revenue — it is flexibility that keeps both above the discounting threshold. That is worth deliberately building capability for.

Adding one column to the tracking — which crew — costs nothing, and turns a vague sense of comfort into a specific instruction.

Common questions

Should quotes outstanding count as backlog?

No. Backlog is signed work only. Outstanding quotes are a separate pipeline number and they are worth tracking, but mixing them makes backlog useless as a capacity measure. A quote is a maybe; backlog is a commitment you have to staff and schedule.

What if my jobs vary hugely in size?

Weekly production capacity in revenue terms already smooths this out, because it is an average of what you actually complete. If your mix has shifted structurally — say from mostly small jobs to mostly large ones — recalculate your weekly capacity, because revenue per crew-week changes with job size.

Is a long backlog always good?

No. Past roughly sixteen weeks your lead time starts costing you work, particularly urgent jobs, which are usually your best margin. You are also locking in today's prices against costs that will be months older when you deliver. Long backlog is usually a signal to raise prices rather than to hire.