There's an old taxi rule, and it has nothing to do with trust: ask the price before you get in.
A metered taxi and a driver who names one price to your door will both get you home. But the moment you close the door, they are pulled in opposite directions. The meter rewards the long way round and the slow lane. The agreed fare rewards getting you there. Same driver, same road — the only thing that changed is how you agreed to pay.
How you pay someone isn't an administrative detail. It's the strongest instruction you'll ever give them, and most of it goes unspoken.
What each way of paying quietly rewards
There are three common ways to pay a firm to build something. Each one points the work in a direction, whatever the contract says on top.
- Hourly is the meter. Every hour worked is income earned, so finishing quickly costs the firm money. Nobody has to be dishonest for this to bend the work — the meter is running, and the meter doesn't reward done.
- Retainer is the monthly fee. You pay to keep the firm around. The reward is to stay necessary: to remain the people you call, indefinitely. A firm on a retainer has no reason to make itself unnecessary, and every reason not to.
- Fixed price is the agreed fare. One number, settled before the work starts, to reach a named destination. The reward is to get there and be finished. Speed helps the firm instead of costing it. Leaving is the win.
We bill the third way, and we do it on purpose.
We bill fixed scope because we want to leave. A retainer pays a firm to stay necessary. We'd rather be unnecessary, and paid in full.
Why fixed price is harder for us, not you
Fixed price sounds like the customer-friendly option, and it is — but the discipline lands on our side, not yours. To put one number on a piece of work, we have to understand it before we start. No discovering halfway through that it's twice the size and sending a bigger invoice. If we get the size wrong, that's ours to absorb, not a surprise on your bill.
That's the honest trade. Hourly billing lets a firm be vague and then bill the vagueness. Fixed price forces us to be specific or admit we can't be. The number is a promise about how well we understood your problem.
The only thing that moves the number
A fixed price is fixed. It doesn't creep, and it doesn't get revised upward because the work turned out harder than we thought — that risk is ours. The one thing that changes the number is you asking for something new: more scope, a second process, a feature that was never in the proposal. Add to the scope and the price moves, openly, with a new number you agree to before we build. Everything inside the original boundary, we hold ourselves to.
This is the opposite of how much of this trade is paid, and we think the trade has it backwards. A firm that profits from taking longer will, on average, take longer. A firm that profits from staying will, on average, find reasons to stay. We'd rather be paid to finish, and then go.
For the technical reader
The mechanism behind a fixed price is a scope boundary drawn before any code is written. We scope an engagement in a short, time-boxed pass — a 15-minute call, one shared process document, a look at the systems you already run — and from that we draw a hard line around exactly one process: the most common path end to end, plus a defined escalation route for the exceptions we deliberately won't automate on day one. Scope is a boundary, not a wishlist.
That boundary is sized against a fixed tier rather than an hourly estimate: SIGNAL, one to two weeks; SPRINT, two to four; OVERHAUL, four to eight. The tier sets the number, the number goes on a one-page proposal, and it's fixed before we start.
Two things then hold the line:
- Estimation risk sits with us. If we underestimated the build, we carry the overrun — the price doesn't move because we were wrong about the size. So the incentive to scope accurately is ours, which is where it belongs.
- Change is explicit, never silent. New scope — a second process, a feature outside the boundary — is a new line with its own number, agreed before we build it. Scope never creeps in unpriced, and it's never billed as a surprise.
There's one place fixed price is the wrong instrument: genuinely open-ended research, where the deliverable is "find out whether this is even possible." There the honest structure is a small, fixed-price spike with a defined question and a defined end, not an open meter. And if a firm can't define the question, that's the signal the work isn't ready to be quoted — by us or anyone.