Fixed price vs time & materials — which contract to sign

What each contract model actually is

• Fixed price (fixed scope, fixed cost) — You and the vendor agree on a written scope. The vendor commits to deliver it for a fixed dollar amount by a fixed date. Any change to scope becomes a change order — new price, new date. The vendor owns delivery risk.
• Time & materials (T&M) — The vendor bills for actual hours worked at an agreed hourly or daily rate. Scope can change every sprint. You own delivery risk — if things take longer, you pay more. Usually paired with a monthly cap or a not-to-exceed number.
• Dedicated team / retainer — A hybrid. You pay a fixed monthly rate for a fixed pod (e.g. 4 engineers + a lead). The vendor commits to staffing and continuity; you commit to a minimum term. Scope is flexible within the capacity of the team.
• Milestone-based hybrid — The version we usually recommend. Fixed price on the discovery phase (typically 2–4 weeks, $15K–$50K). Once discovery lands with a real spec, either roll into a fixed price build (if scope is clear) or T&M with a monthly ceiling (if it isn't).

When fixed price is the right call

Fixed price works when the deliverable is unambiguously definable in advance. That's rarer than most buyers assume, but it happens.

• Compliance builds with a defined checklist — SOC 2 remediation, HIPAA gap closure, PCI scope reduction. The spec is the auditor's report; there's no ambiguity.
• Migrations from A to B where both A and B are known — Postgres 12 to 15, on-prem to AWS, monolith to modular. The scope is 'the same thing, running somewhere else, working the same way.'
• Websites and marketing sites with signed-off designs — Design is fixed, content is provided, launch date is real. Standard fixed-price territory.
• Well-defined integrations — 'Sync data from our Salesforce instance to our Postgres warehouse, using this schema, with these transformations.' If both ends are stable, this is bidable.
• POCs and prototypes — The scope is intentionally narrow and time-boxed. Fixed price with a 2–4 week timeline is the industry norm and it works.

When T&M is the right call

T&M wins when you know the destination isn't known — either because the product is early and you'll be learning, or because the roadmap has to respond to real customer feedback. Trying to fix-price this kind of work forces the vendor to pad the price for risk, and the padding often exceeds the actual T&M cost.

• New product development past the MVP stage — You're iterating. Every sprint changes based on what you learned last sprint. Fixed price would force you to define six months of work up front — which is exactly what you don't want to commit to.
• Ongoing product engineering — Roadmap changes, priorities shift, one feature reveals another. Dedicated team on retainer is the mature version of this pattern.
• Bug fixing and maintenance on an existing product — You don't know what bugs you'll find. Fixed-price maintenance contracts either don't fit the work or price for the worst case.
• R&D and technical spikes — 'Figure out if this is feasible' has no fixable scope. Time-box it (e.g. 3 weeks, $30K NTE) and see what the team learns.

The traps in each model

• Fixed price: change order friction — Every mid-project scope adjustment becomes a negotiation. In practice, buyers underestimate scope by 20–40%, then pay change orders for every gap. The final total often exceeds a T&M engagement by 15–30%.
• Fixed price: quality corners on the last 20% — When the vendor's margin is being squeezed by unbudgeted scope, the last-mile polish (tests, error handling, docs) is where they cut. This is where the phrase 'fixed price got fixed quality' comes from.
• T&M: runaway hours with no ceiling — Without a monthly cap, a hostile or careless vendor can run up hours without visibility. Fix: monthly not-to-exceed number, weekly burn reports, and a hard rule that hours over the cap are the vendor's to absorb unless pre-authorized.
• T&M: no delivery commitment — 'We'll get it done when it's done' is unacceptable. T&M contracts should still include committed milestones and delivery dates, even if the underlying billing is hourly.

The hybrid we recommend

The pattern that wins for most product work: (1) Fixed-price discovery — 2–4 weeks, $15K–$50K, produces a real spec, wireframes, and a shortlisted architecture. (2) Fixed price for the parts of the build that are locked (setup, marketing site, well-known integrations). (3) T&M with a monthly ceiling for the product-development portion (main app, features, iterations). (4) Retainer for post-launch continuous work. Almost every serious engagement we've run ends up with this shape by month three anyway — starting there saves the negotiation.