Send the same brief to three studios and the quotes will not be close. The instinct is to assume the cheap one is cutting corners or the expensive one is padding. Usually neither is true: they have read your sentences differently, and nobody has said so out loud.
The four places the readings diverge
How much is in the first version. "A booking system" can mean one journey that works, or every screen the eventual product needs. A quote assuming the first and a quote assuming the second are not comparable, and both are honest.
Who is doing the deciding. If your side has a person who can answer domain questions within a day, the estimate is smaller. If every question waits a fortnight for a committee, the calendar stretches and someone has to price the waiting.
What "done" includes. Deployment, tests, an admin panel, a handover, a month of fixes afterwards. Or: a repository and good luck. The second is genuinely cheaper and genuinely worth less.
What happens to the unknowns. A studio that has built something similar knows where the sharp edges are and prices them. One that has not either omits them or adds a blanket margin. Same number, different reasoning.
Ask for assumptions, not for a number
Ask each studio for their assumptions in writing, in their own words: the list of things they believe to be true about your project. Three lists side by side tell you more than three prices ever could, because you will see immediately which studio understood you and which answered a different brief.
Then ask what they would cut first if the budget were thirty percent smaller. The answer reveals what they think is load-bearing. If everything is essential, nobody has thought about it.
The number that is actually comparable
A first version with one journey working end to end — accounts, the main flow, deployment and handover — is roughly eight weeks of a small team. Anything meaningfully cheaper has dropped something from the "done" list; anything meaningfully dearer is building more than one journey. That is a comparison three quotes can be held against.
And the estimate that moves
Every estimate is a guess with a confidence attached, and honest ones say so. What matters is not whether it moves but when you hear about it. An assumption that breaks in week two should reach you in week two, with what it costs. An estimate that holds silently until week seven and then explodes was never an estimate — it was a hope with an invoice attached. Our own estimates come with the assumptions printed under them, and the first version is fixed at $16,000 for eight weeks so the argument happens before the work rather than during it.