Pick one number to decide the next project, and keep it in one place
Whether to take the next project is decided by the headroom left against promises already made, not by revenue.
Key points
- Watch headroom on promises, not revenue
- Count the hours left after commitments
- One place, read just before you quote
- Decide in advance: decline or move it
Contents
The number to watch is headroom, not revenue
A&A perspective
The number meant here is a single figure you look at once, before deciding whether to take the next project.
A&A perspective
Automate & Augment (A&A), who write this article, design and build systems that hand fixed-procedure work to AI and leave the judgement with a person.
A&A perspective
From that position, the figure to watch is not revenue. It is the headroom left against the response already promised to your current customers.
A&A perspective
Watch revenue and the number improves every time you accept. A reason to decline never appears in it. Watch headroom and the number drops the moment you accept. Because the drop is visible on the spot, the reason to decline appears in the same place.
From the sources
Paul Graham is a writer who has advised the founders of new companies. In an essay published in July 2013 he wrote about what turns risky once you are paid by the hour.
From the sources
A product company that is merely being extra attentive to a customer, he writes, is thanked for the attentiveness. But once that attentiveness itself carries a fee, once they start paying you by the hour, Graham writes that they expect you to do everything.
A&A perspective
A&A reads this from the contractor's side. When work is paid by the hour, there is no ceiling on what the other side expects. So the total hours you have sold will not tell you how much more you can take on. What it does not give, the time left after delivering what you already promised does.
A&A perspective
Headroom has a second advantage. Revenue only moves after you accept, while headroom moves before you accept. The only number usable when deciding is one that moves before the decision.
Hypothetical example
Suppose you have 40 hours left this month. If 30 of them are committed to three projects already under way, 10 remain. Now an enquiry arrives that you estimate at 20 hours. Revenue tells you only that accepting would make it larger. Headroom tells you that you are 10 hours short.
Count headroom in the hours that are left
A&A perspective
One subtraction is enough. From the hours you can still work this month, subtract the hours already committed to existing projects. What is left is the headroom you can offer the next one.
A&A perspective
What you subtract covers more than delivery work. Replies and meetings for current customers belong there too. Count without them and headroom reads larger than it is. A headroom figure that reads large becomes the excuse for taking on too much.
A&A perspective
Take committed work from either the quote or the time log, never from both. Adding both counts the same work twice. The quote gives the amount you promised; the time log gives the amount it actually took. Decide which one you use before you start, and do not change it afterwards.
Hypothetical example
Suppose the hours you can work this month are 40. Two existing projects have 25 hours committed, and replies and meetings for enquiries take 5. That leaves 10 hours, and those 10 are the headroom available to the next project.
A&A perspective
Match the counting period to the length of your projects. If one project finishes in a few weeks, count by the week; if most run for months, count by the month.
A&A perspective
You can also count projects instead of hours. You set a ceiling on how many you hold at once, then count how many you hold now. Counting projects is easier, but it treats projects of very different sizes as one each. For a company whose projects are close in size, counting them is the method more likely to stick.
A&A perspective
In a month when headroom goes negative, keep the number as it is. Floor it at zero and you lose how far over you are.
Hypothetical example
Suppose the headroom line reads minus 15 hours. From that one number you can read not only the answer to decline, but also which delivery date you would have to move to make it fit.
| Aspect | Watching revenue | Watching headroom |
|---|---|---|
| When it moves | After you accept | Before you accept |
| On accepting | It goes up | It goes down |
| Reason to decline | Never appears | Appears on the spot |
| What is counted | Money sold | Hours left on promises |
Keep it in one place, read just before you quote
A&A perspective
Scattered numbers are not peculiar to contract businesses. There is an account from a different industry of numbers that were scattered the same way and were brought into one place.
From the sources
On 15 September 2026 Anthropic published an article announcing new workflows and integrations for small businesses. One of the owner stories in it concerns a company called Mothership Coffee Roasters, which runs six cafés in Las Vegas and also does wholesale and catering.
From the sources
At the company of its founder, Juanny Romero, the numbers sat in four places: the till, the accounting software, the team chat and email. There was nowhere to see them together once a week.
From the sources
Romero connected those four numbers to Claude and began doing the forecasting and the weekly planning there. She now runs all six locations from one consolidated place.
From the sources
In-store margins rose to 22%, which she credits in part to tighter scheduling, ordering and inventory reporting.
A&A perspective
That account comes from food and retail, where the cost structure and the shape of a busy period differ from contract work. The source is Anthropic's own announcement, and the 22% is the company's own report.
A&A perspective
What A&A takes from it is not the margin. It is the point before the margin: numbers sitting in four places, with nowhere to see them together once a week. That state, and only that, carries over to a contract business unchanged.
A&A perspective
Choose the place from among the things you already open every day. A time log, a spreadsheet or a note to yourself will all do. Add a new tool and you add the effort of going to look, and that effort becomes the reason not to look.
A&A perspective
The position to put it is just before the quote goes out. At the moment the enquiry arrives it is still too early, and once the quote has been sent it is too late. If the number catches your eye immediately before you write the price and the date, the decision changes on the spot.
A&A perspective
Updating once a week is enough. Fixing it daily stops your hands, and updating monthly leaves it stale by the time you decide. Collecting it automatically from the time log removes the weekly typing. Write the update date beside the number and you will not read a stale figure without noticing.
Decide in advance what you do in a short month
A&A perspective
Putting the number somewhere is not enough. Without an answer decided for the month it comes up short, you will accept anyway. Looking alone does not stop your hand. So write what you do in a short month on one line, in the same place as the number.
A&A perspective
There are two answers: decline, or move the delivery date back. Both need wording you send to the other party, so writing them in advance means you do not have to think on the day the enquiry lands.
Hypothetical example
A line for declining could read: this month is full with existing projects, so we cannot take it on; we could start from the second week of next month. For moving the date, put the timing first, as in: start the second week of next month, delivery four weeks after that.
A&A perspective
Declining is not a decision to lose the other party. It is a decision to protect the response owed to the customers you already have. When a delay appears after you accept, what you lose is not the new party but the trust of the one already paying you.
A&A perspective
Write the condition for an exception on that same line. You might decide, for instance, that additions from existing customers are taken even above headroom. Allow one exception at most. The more exceptions there are, the less you look at the number at all.
A&A perspective
You might think of outsourcing as a way to add headroom, but what it adds is working hours only. The time spent checking what comes back stays with you. When you count headroom, keep the checking time for work you plan to outsource on your own side of the subtraction.
A&A perspective
At the end of the month, put the headroom you had at quoting time beside the hours the work actually took. If the gap falls the same way every month, your estimate of available hours is generous. More than the figure itself, the direction of the gap is what shows you where to correct.
Summary
A&A perspective
Whether you take the next project is settled less by how good the project is than by the number you look at before you accept.
A&A perspective
The number to watch is not revenue. It is the headroom left against the response you have promised to your current customers.
A&A perspective
Count headroom as the hours you can work this month minus the work already committed.
A&A perspective
Keep it in one place and read it just before the quote goes out. For the month it comes up short, write in advance whether you decline or move the delivery date.
A&A perspective
There are situations where none of this is needed. If you only ever hold one project at a time, you know your headroom without looking.
Whether you take the next project is settled less by how good the project is than by the number you look at before you accept. The number to watch is not revenue but the headroom left against the response already promised to your current customers, because revenue only moves after you accept while headroom moves before. Count headroom as the hours you can work this month minus committed work and minus the time owed to current customers, keep it in one place, and read it just before the quote goes out. For the month it comes up short, write in advance whether you decline or move the delivery date. There are situations where none of this is needed: if you only ever hold one project at a time, you know your headroom without looking.
FAQ
What if headroom is short but I really want the project?
You can still accept. If you do, decide at the same time which existing delivery date you will move. Accepting above headroom means delaying a promise somewhere. Accept without choosing where, and who gets delayed is settled afterwards, by accident.
Does narrowing it to one number make the decision crude?
It does make it cruder. Narrowing to one is not for accuracy; it is so that you actually look before you accept. An accurate figure split across four places does not get read before the decision. When you want more precision, correct how headroom is counted rather than adding more numbers.
Can I hand the updating of this number to AI?
The updating can be handed over. What cannot is deciding which record you use, and deciding what you do in a short month. Automate the updating without settling those two and a figure arrives every week that nobody can say is right.
Sources & editorial note
Primary pages read for this article. Publication dates below belong to the sources; access dates record our research.
- Claude for Small Business launches new workflows, integrations, and training programs
Anthropic · September 15, 2026
Accessed 2026-10-21 - Do Things that Don't Scale
Paul Graham · July 2013
Accessed 2026-10-21