Quoting a range without losing the job
A single number communicates a confidence you do not have. A range communicates the truth and is harder to sell.
Every estimate is uncertain, and a single figure hides that entirely. The client hears a commitment, you meant a midpoint, and the difference becomes an argument in month two.
The honest alternative is a range, and the reason people avoid it is straightforward: clients read the top of a range as the price and the bottom as the promise. Writing assumptions down also reduces the load on memory; further details gives useful background on that effect.
Attach the range to the unknowns
A bare range looks like hedging. A range with its causes named looks like competence.
'Between forty and sixty-five hours: forty if the content is supplied as agreed and two rounds are enough, sixty-five if content needs writing and the integration behaves the way it did last time.' The client can now see what moves the number, and — usefully — several of those things are within their control.
A range with conditions attached becomes a conversation about the conditions. That conversation is where the real scope gets discovered, before the work starts. For a broader framework, PMI's overview of project management summarises the standard concepts behind planning and delivery.
Narrow the range by doing a paid discovery
Where uncertainty is genuinely large, the honest answer is that it cannot be estimated yet. Quoting anyway means picking a number that is either too high to win or too low to survive.
A short, separately quoted discovery phase — a few days to establish requirements, look at the existing system, produce a specification — resolves this. It is billable, it produces a deliverable the client owns, and it ends with an estimate you can stand behind.
Cap it if you must, and price the cap
Clients frequently want a fixed price. That is a request to transfer risk, and transferring risk has a price.
The workable version is a not-to-exceed figure set near the top of the range rather than the middle, stated as such. What does not work is agreeing the midpoint as a cap, which is accepting the risk and pricing it at zero.
Fixed price is a bet on the scope, not the work
Fixed-price projects fail on scope rather than on effort. The work was understood; what was not understood was that the client considered three additional things obviously included.
Fixed price is therefore viable exactly when the scope can be written down precisely and change is charged separately. Without both halves, a fixed price is an open-ended commitment with a closed-ended fee.
Show the maths
An estimate presented as a breakdown — twelve items with hours against each — is negotiated differently from a lump sum. The client removes items rather than pushing the total down.
That is a better conversation for both sides: the price drops because the scope dropped, which is the only kind of discount that does not come out of your margin.