Fixed price, hourly, or retainer. Whoever sets the model owns the incentive.

Most founders treat pricing as a number to negotiate. The model behind the number matters more. Hourly rewards slow work. Retainer rewards stretching. Fixed price rewards finishing. Pick the one whose incentive points where you want the agency pointed.

Hourly: where the math breaks for founders

Hourly billing only feels honest. "You only pay for what we use." What it actually means is: every hour the agency spends learning your codebase, debugging their own assumptions, sitting in meetings, or rewriting after a junior's pass is hours you pay for.

A 6-week build at one full-time senior plus a part-time frontend is roughly 320–400 hours. At blended $130/hr that is a sticker quote of $42k–$52k. By the time the build ends, real-world hourly engagements drift 30–60% over the original estimate, sometimes more, because no agency that bills hourly is incentivized to scope tightly.

Retainer: right model, wrong moment

A monthly retainer is the right shape for ongoing engineering capacity: post-launch maintenance, feature pipeline, on-call coverage. It is the wrong shape for a one-shot build because there is no finish line and no incentive to find one.

The retainer trap looks like this: month one feels productive, month two slows, month three is "almost ready," month four is invoiced anyway. By month six the agency has billed $60–120k for what should have been a six-week deliverable. You did not pay for a build. You paid for occupancy.

Reasonable retainer use cases:

• Post-launch maintenance with a defined scope and SLA
• Embedded engineering on an in-house team with a CTO running the work
• On-call audit response and incident handling for live contracts
• Multi-quarter feature pipeline with a roadmap you control

Fixed price: aligned by design

A fixed-price contract makes the agency's problem the same as yours: ship the deliverable. If the build runs long, the agency loses margin, not you. If scope was underestimated, the agency eats it. The only way the agency makes money is to scope honestly up front and execute fast.

This only works under three conditions, all of which we enforce on our side:

• The agency must scope competently. Bad scoping turns fixed price into a financial trap. We refuse work we cannot scope confidently in a single call.
• The deliverable must be writable down. "Build something cool" cannot be fixed-price. "ERC-3643 token, KYC integration, issuer dashboard, investor portal, mainnet by date X" can.
• The agency must actually absorb overruns. Some "fixed price" contracts include weasel clauses for change orders, scope adjustments, or "additional discovery." Ours do not.

When fixed price is the wrong choice

In the spirit of not overselling: fixed price is not always right.

• If the scope is genuinely unknown (research project, novel L1 work, ZK circuit design with uncertain constraints), demand fixed price and the agency will either refuse or pad heavily. Pay hourly with a senior team and a strict cap.
• If you have an in-house CTO managing the work day-to-day, hourly access to a strong contractor is often cheaper than agency overhead.
• If you are running a multi-year ongoing program rather than a discrete build, retainer is right.

But for the canonical web3 founder shape — known scope, hard deadline, fundraise calendar — fixed price is the only pricing model whose incentives align with shipping.