You have a quote in front of you and a choice to make: pay a fixed price for a defined scope, or pay for the hours the team actually spends. The fixed price feels safer. The hourly option feels like a blank cheque. Both feelings are partly wrong, and choosing on instinct is how founders end up either overpaying for padding or arguing over invoices.
The short answer: fixed price works when the scope is small and genuinely well defined, such as a first milestone or a focused MVP. Time and materials (T&M) works when the product will change as you learn, which is almost every product after launch. Most good studios use a hybrid: a short paid discovery phase, then a fixed-price first milestone, then T&M or a monthly retainer. Whatever the model, the contract should give you code ownership, repository access from day one, clear acceptance criteria and a warranty period.
Fixed price vs time and materials at a glance
| Fixed price | Time and materials | |
|---|---|---|
| You pay for | An agreed scope | Hours or days actually worked |
| Budget certainty | High, if scope holds | Lower; controlled by caps and sprints |
| Flexibility to change | Low; every change is a change request | High; reprioritise every sprint |
| Who carries estimation risk | The vendor (priced in as a buffer) | You |
| Upfront effort | Heavy specification before work starts | Light; can start quickly |
| Typical buffer in the price | 20–40% for unknowns | None, but you pay for detours |
| Best for | MVPs, first milestones, clear deliverables | Evolving products, post-launch work, R&D |
| Main failure mode | Disputes over what was "in scope" | Budget drift without visible progress |
The detail that most founders miss: fixed price is not cheaper. The vendor carries the risk of underestimating, so they price it in. You are buying certainty, and certainty has a premium.
When fixed price fits
Fixed price is the right call when:
- The scope fits on a few pages and you can describe every screen and rule.
- You are working with a new vendor and want to test them on a bounded piece of work.
- You have a hard budget, such as a grant, a pre-seed round or an internal approval.
- The deliverable is concrete: a landing page, a focused MVP, an integration, a migration.
It goes wrong when the scope is vague. "A marketplace like Amazon but simpler" cannot be fixed-priced honestly. Either the vendor pads heavily or underquotes and then fights you on every detail.
When time and materials fits
T&M is the right call when:
- You expect to learn from users and change direction.
- The work is ongoing: iteration after launch, new features, maintenance.
- The technical path is uncertain, such as a tricky integration or performance work.
- You have someone on your side who can prioritise weekly.
Protect yourself with a monthly cap, two-week sprints with a demo at the end of each, and a shared board where you can see what is being worked on. T&M without visibility is where the blank-cheque fear comes true.
The hybrid most good studios use
In practice, the best structure for a startup build combines both:
- Paid discovery (1–3 weeks). Workshops, user flows, a data model sketch, a clickable prototype of the core workflow and a written scope with estimates. Discovery typically costs 5–10% of the expected build. You own the output and can take it to any vendor.
- Fixed-price first milestone. The core workflow, deployed to production. This tests the team on a bounded deliverable and gives you something real to show users or investors.
- T&M or retainer afterwards. Once the product is live, priorities change weekly. A monthly retainer or T&M with a cap fits that reality.
This is how we usually structure engagements, and our published bands are on /services. Discovery is also where you avoid the most expensive mistake, building the wrong thing well. Our spare-parts marketplace MVP checklist shows the kind of decisions that belong in that phase.
If you want a head start on discovery, the free Project Brief Generator turns your idea into a structured scope document in a few minutes.
What a good software development contract contains
Whichever model you choose, check for these clauses:
- Scope and change control. A written scope (or backlog for T&M) and a defined process for change requests: who raises them, how they are estimated, who approves.
- Milestones and acceptance criteria. Each milestone lists what will be delivered and how you will test it. "Done" means accepted by you within a set number of days, not "deployed by us".
- Milestone payments. A deposit (commonly 20–40%), then payments tied to accepted milestones. Avoid paying the majority up front.
- IP and code ownership. All code, designs and assets are assigned to you on payment. Open-source components are listed with their licences.
- Source-code access from day one. The repository lives in your GitHub or GitLab organisation, not the vendor's. Same for cloud, domain and app-store accounts.
- Warranty period. Bugs against the agreed scope are fixed free for a period after launch, commonly 30 to 90 days.
- Handover. Documentation, environment setup, credentials and a handover session if you move to another team.
- Support SLA. Response times by severity after the warranty ends, and the monthly cost of that support.
- Termination. Either party can exit with notice; you pay for work done and receive everything produced to date.
Red flags in software quotes
- One line, one number. "Marketplace app: $18,000" with no breakdown means nobody has thought about your scope yet.
- No discovery, no questions. A vendor who quotes a complex platform after a 20-minute call is guessing.
- The vendor hosts the code. If you cannot see the repository until final payment, you have no leverage and no visibility.
- Unlimited revisions. Sounds generous; in practice it means scope was never defined and arguments will follow.
- Very low fixed price for a big scope. Expect change-request inflation or a team that disappears halfway.
- No mention of testing, deployment or warranty. Those are where cheap quotes quietly save money.
- T&M with no cap, no sprints, no demos. You will find out what was built when the money runs out.
Frequently asked questions
Is fixed price or time and materials better for a startup?
For a first milestone or a focused MVP, fixed price usually fits because the scope is small and you want budget certainty. After launch, time and materials or a retainer fits better because priorities change weekly.
How much does a discovery phase cost?
Typically 5–10% of the expected build cost, over one to three weeks. For a small MVP that can be a few hundred to a couple of thousand dollars; for larger platforms more. The output should be yours to keep.
How should milestone payments be structured in a software project?
A common structure is a 20–40% deposit, then payments released as each milestone is accepted against written criteria, with a final payment after launch. Tie payments to accepted deliverables, not to dates.
Who owns the code when I hire a software agency?
You should, and the contract should say so explicitly. Insist that the repository is in your organisation from the first commit and that IP transfers on payment.
Your next step
Before signing anything, write your scope down in a form any vendor can price the same way. Start with the free Project Brief Generator, compare our engagement bands on /services, and if you want a second opinion on a quote you already have, book a free call. We will tell you honestly whether it looks fair and what is missing.