Five questions that predict whether a project will be profitable

The usual way to qualify a project is to check the fee and the fit. Both are answerable in five minutes, which is most of their appeal. Neither one predicts what the project earns you, because the fee is revenue, and revenue is not margin. Margin is the fee divided by every hour the work eventually consumes, and nearly all the variance in that denominator is set by things you could have asked about before saying yes.

Getting the fee right does not protect you. In IPSE's late payment research, 35% of self-employed people said a client had not paid them on time in the previous twelve months, and the average amount owed was £5,230. Every one of those was a project someone had agreed a price for. The price was never the problem.

So treat qualification as forecasting. Five questions, each of which moves the denominator in a direction you can estimate. Ask them in the scoping call, before the proposal, when asking them still reads as diligence rather than suspicion.

Throughout, I'll use one illustration and keep substituting into it: a fixed-price project at €9,000, scoped at 100 hours, so a quoted rate of €90 an hour. The 100 hours already include two rounds of consolidated client feedback, because a scope that assumes zero revision is not a scope, it is a wish. These numbers are invented for the arithmetic. Your own figures go in the same slots.

Every additional approver is a revision round you did not quote

Ask who signs off, and then ask who else sees it before that person signs off. The second question is the one that produces the real answer, because the marketing manager who briefed you rarely has final say and rarely volunteers that.

The mechanism is not that more people means more opinions. It is that opinions arrive serially. One decision-maker gathers internal comments and hands you one list. Four stakeholders who never meet each other hand you four lists, at four different times, and two of them contradict. You do the work, then redo it, then reconcile.

Two extra review cycles at eight hours of rework each, plus four hours of scheduling and re-explaining, is twenty hours. The project is now 120 hours, and €9,000 ÷ 120 = €75 an hour. You have lost 17% of the rate before anyone has behaved badly.

The cheap fix is not refusing the project. It is naming one approver in the proposal and writing that consolidated feedback comes back in a single document by an agreed date. Clients accept this readily, because they also hate the four-list version.

"Done" defined by output costs less than "done" defined by mood

The second question is what the deliverable is. Not what the project is about — what the artifact is, in a form that can be pointed at.

"A homepage design, one desktop and one mobile artboard, delivered as a Figma file with components" is defined by output. It is either delivered or it isn't. "A homepage the team is happy with" is defined by satisfaction, and satisfaction has no completion state. You cannot invoice against it, because there is no moment at which the condition is objectively met, and there is no argument you can win about whether it has been.

This is the question that separates a project you can price from one you can only rent yourself out for. If the answer comes back as a feeling rather than a file, either convert it into an artifact list in writing before you quote, or price it hourly and let the client carry the estimation risk.

Assume you don't convert it, and acceptance drags on: ten extra hours of small changes nobody would have asked for against a defined deliverable. 130 hours, €9,000 ÷ 130 = €69.23 an hour.

Revision four is where fixed-price work stops earning

Ask how many rounds of revision are included. Then ask the question almost nobody asks, which is what happens at round four.

Most contracts stop at the first half. Two rounds included. Fine. But a revision cap with no stated consequence is not a cap, it is a preference. The client asks for a fourth round, you do it because the relationship is worth more than the argument, and the cap did nothing except make you feel worse about it. What makes a cap load-bearing is the sentence after it: further rounds are billed at your hourly rate, quoted and approved before the work starts.

The measurement matters as much as the clause. If you don't know how many hours went into each round, you cannot tell the client what round four costs, and you cannot tell yourself whether the cap was set in the right place. Time data that stops at a weekly total is not enough; you need hours attached to the phase they belong to, which is the honest argument for a Toggl alternative where tracked hours turn into paid invoices rather than a timer that reports into a spreadsheet you never open.

Uncapped, and with the two extra approvers from the first question all wanting a look, four further rounds at roughly four and a half hours each is eighteen hours. 148 hours, €9,000 ÷ 148 = €60.81 an hour. That is 68% of what you quoted, and no one has done anything you could call unreasonable.

Worth flagging the crossover point early: at the moment a project's committed hours pass its budget, the decision to re-quote is still available. A week later it isn't, because the work is already done. Worklyn's Scope Watch exists to catch that moment while it is still a decision.

Payment terms are a price, and someone else's calendar sets them

Two sub-questions, and the second one is the one that costs money.

First, what are the terms. Under Directive 2011/7/EU, the default B2B payment period in the EU is 30 calendar days from receipt of the invoice, and anything beyond 60 days needs to be expressly agreed and not grossly unfair to you. A client proposing net 60 is inside the law and outside your interest. The gap between net 15 and net 60 on a €9,000 invoice is 45 days during which you have financed the client's working capital at 0%.

Second, and more useful: who processes the payment, and on what cycle? Many companies pay on a fixed run — the 1st and 15th, or a single monthly batch. An invoice that misses Friday's cut-off does not wait three days, it waits until the next run. Net 30 against a monthly run can settle at day 55 with nobody at fault. Ask which day the run happens and who in accounts payable receives the invoice, and put that person on the invoice email from the first one.

Chasing has a cost you can count. In a 2022 survey of New York freelancers run by the Freelancers Union with several creator guilds, 76% said they spent one to two hours a week chasing payment. That survey covers one state, one year, and does not disclose its sample size, so treat it as an order of magnitude rather than a rate. Six hours of chasing on this project takes it to 154 hours: €9,000 ÷ 154 = €58.44 an hour, before you count the cost of the money arriving 45 days later than it needed to. The mechanics of shortening that gap are the subject of the payment system post, and they are largely decided at proposal stage, not at invoice stage.

Being load-bearing for the client's deadline cuts both ways

The last question is the one that sounds least commercial and predicts the most: what does the client need from you, and when do they need it for?

If your delivery sits on the critical path of something the client has already committed to, such as a launch or a funding round or a conference date, your negotiating position on price and terms is strong, because a delay at your end delays their thing. Reviews come back fast, because someone senior is being asked about the date.

If you are useful but not urgent, the opposite happens, quietly. Feedback takes eleven days instead of two. You reopen a project you last touched three weeks ago and spend two hours reconstructing where you were. The work does not get worse; it gets more expensive, in an accounting nobody is doing.

Call it twelve hours of re-immersion and stalled restarts across the project. 166 hours, €9,000 ÷ 166 = €54.22 an hour — 60% of the quoted rate.

Neither answer disqualifies a project. High dependency comes with expedite demands and weekend messages, and should be priced for that. Low dependency should be priced for drift, or structured with milestone dates that expire so the schedule stops being infinitely elastic.

The same project, under two sets of answers

Nothing in the right-hand column is a bad client. Every one of those answers is normal corporate behavior, and together they cost 40% of the rate. That is the argument for asking: the erosion is entirely legal, entirely polite, and entirely predictable in advance.

The same arithmetic run backwards after the fact is what an effective hourly rate calculation gives you, and it is worth doing on your last three finished projects before you trust the forecast on your next one. If you have never separated the rate you quote from the rate you get, that gap is the whole subject and it is bigger than most people expect.

The rule: reprice by the ratio, or decline

Here is the threshold at which the answer flips.

Ask the five. Count how many come back in the right-hand column. If it is zero or one, quote normally; the estimate will hold within the noise. If it is two or more, you are not looking at a €90-an-hour project, and quoting one is a decision to work for less, made in advance, in silence.

So make it explicit instead. Forecast the hours using the additions above, substituting your own numbers for mine, and divide the fee by the forecast total. Take the ratio of your target rate to that forecast rate, and multiply the fee by it. In the illustration, €90 ÷ €60.81 = 1.48, so the €9,000 project is a €13,300 project, and the choice in front of you is whether the client will pay it. Sometimes they will, because they know internally that their sign-off process is expensive. When they won't, you have learned that at proposal stage rather than in week seven, which is the entire point.

The one thing not to do is quote €9,000 while privately hoping the answers were wrong.

Worklyn's proposals convert straight into the contract and then the invoice, so the approver you named and the revision cap you priced survive into the document the client actually signs.