When fixed-price beats time-and-materials

Time-and-materials (T&M) is the default contract structure for software engineering work. It is the default because it is easy to write: hourly rates × estimated hours = budget, with overage billed at the same rate. Both sides know what they’re signing.

What both sides usually don’t acknowledge is that T&M structurally rewards the vendor for being slow.

The mis-alignment

Under T&M, the vendor’s revenue is monotonically increasing in hours billed. There is no point on the curve where the vendor’s interests align with the client’s interest in shipping fast. This doesn’t mean vendors are dishonest — most aren’t. It means the system creates background friction that even honest vendors don’t fight against. Meetings that could have been emails. Discovery rounds that could have been single questions. Refactor sprints that the senior team would have rejected if they had personal margin to defend.

Fixed-price flips this. The vendor’s margin shrinks with every hour spent. The vendor’s incentive is now to ship the agreed scope as fast as possible at the agreed quality. The client’s incentive is the same. These two now point in the same direction.

When T&M actually fits

T&M is genuinely the right instrument in two cases:

1. The scope is unknowable. True research engagements where the answer to “what are we building?” is “we don’t know yet, that’s part of the work.” Pricing the unknown as fixed is dishonest.
2. The relationship is long, with continuously evolving scope. Some embedded-team arrangements work this way. The engagement is not a project; it is staff augmentation with vendor packaging.

If neither case applies, T&M is a default chosen out of risk-aversion, not fit. The risk it manages — that the vendor will lose money on a bad estimate — is not a risk the client should subsidise.

Why fixed-price requires a senior bench

Fixed-price moves the estimation risk onto the vendor. To absorb that risk, the vendor needs three things:

• Pattern recognition. Engineers who have shipped this kind of work before, often enough to estimate it within ±15% confidence.
• Tight scope. A written SOW that says exactly what ships, in what state, on what date. Anything outside the SOW is a paid change order, not “scope creep we’ll absorb.”
• Operational efficiency. AI-paired delivery, reusable architectures, internal templates. The productivity dividend is what makes the fixed-price absorbable.

A vendor without all three should not offer fixed-price. They will either lose money or pad the estimate so heavily that the client could have gotten a better deal under T&M.

What a good fixed-price SOW looks like

A defendable fixed-price SOW has six things:

1. In-scope deliverables. Bullet list. Concrete outcomes. No “support for X” without “support for X means: the system does Y, Z when triggered by A.”
2. Out-of-scope items. Explicit list. This is where most disputes get prevented.
3. Milestone schedule. Dated milestones with payment triggers. Not “monthly retainer.”
4. Acceptance criteria. Testable outcomes per milestone. If a milestone passes the criteria, it is accepted; if it fails, the vendor fixes it without billing.
5. Change-order process. Pre-priced or pre-process. Don’t leave it to “we’ll figure it out later.”
6. IP terms. Code ownership from commit one. Not at end-of-engagement, not pending final invoice.

If any of these six is vague, the SOW will rot in week six.

The economics, blunt

A typical traditional consultancy quotes a 6-month T&M engagement at $400K–$600K for a project that, well-scoped, could ship in 12 weeks at $80K–$120K with senior + AI-paired delivery. The 4–5× delta is real and provable. The reason buyers don’t see it is that T&M never quotes the alternative.

The alternative looks like a fixed-price SOW with a milestone schedule. Ask for one. If the vendor can’t write one, you’ve learned something useful.

Read more: /build/ · /method/ · /about/