Project pricing

Quoting fixed-price projects safely: the buffer math that keeps a fixed price fixed

Nov 16, 2026 · ~7 min read · Last verified: 2026-11-16

A fixed price is a promise: this outcome, this number, no surprises. Clients love the certainty, which is exactly why it is worth selling — and why quoting one without a buffer is gambling with your margin. Every overrun, every fuzzy requirement and every "quick tweak" lands on you. The safety comes from three moves: layered estimates, explicit buffer math, and scope-creep priced as line items instead of absorbed as favors.

Key takeaways

Estimate task by task, best case and reality; a single top-down guess hides the risk inside optimism.
Price = hours × rate × (1 + risk %), plus a separate contingency line for the unknowable.
Quote one number; keep the buffer internal — but write scope boundaries and change pricing into the agreement.
Scope creep is priced, not argued: "happy to do that, here's the line item" ends the negotiation.
Track estimate vs actual; your own error history is the only risk buffer that calibrates itself.

Step 1: estimate in layers, not vibes

Break the project into deliverables and each deliverable into tasks. For each task, two numbers: the time if everything goes right, and the time it realistically takes including feedback loops, revisions and waiting. Sum the realistic column — that is your base estimate. The discipline matters more than the precision: a layered estimate tells you where the uncertainty lives, which is exactly what the buffer will be sized against.

Step 2: the buffer math

Take your base hours (say 60) and your hourly floor from the rate calculator (say 65). Base fee = 3,900. Now two distinct loads go on top:

Risk buffer — covers estimation error on known work: new client, vague brief, unfamiliar tech. At 15%: 585, subtotal 4,485.
Contingency — a separate line for genuinely unknowable events (the stakeholder disappears for three weeks; a dependency breaks). At 10% of the subtotal: 448.50. Final quote: 4,933.50 — an effective rate of 82.23/hour.

Keeping the two lines separate is not pedantry. When a project goes sideways you can see which reserve absorbed the hit, and over a dozen projects you learn whether your problem is estimation (risk buffer too small) or environment (contingency too small). One blended "padding" number teaches you nothing.

How big should the percentages be? There is no universal figure — the honest answer is that uncertainty sets them. Repetitive work for a known client needs little; a first-of-its-kind build needs a lot. Start conservative (many freelancers begin in the 10–25% risk range and adjust), then let your tracked history take over. That is the "industry-standard reasoning" in its truthful form: buffers exist because estimates are distributions, not points.

Step 3: show scope, hide the buffer

The client sees one price and a crisp definition of what it buys: deliverables, revision rounds, what is explicitly out of scope, and the process for changes. Professional associations formalize exactly this structure — the AIGA Standard Form of Agreement for Design Services pairs a master agreement with explicit schedules for scope and pricing, and AIGA's business and freelance resources cover proposals and project management practice more broadly. You do not need their paperwork — you need their principle: the agreement defines the boundary; the price stands on the boundary.

Step 4: scope creep becomes a line item

Somewhere mid-project, the request arrives: "could you also…". Unpriced, each yes is a small donation; each no is a small conflict. Priced, it is neither: "Happy to — that's outside the current scope; as a change order it would be X." No apology, no tension, and often the client simply decides it wasn't important. The magic is that the mechanism was announced in the agreement, so invoking it is administration, not confrontation. When changes arrive in clusters, that is a signal the project has really changed shape — re-quote it rather than patching it with five change orders.

When not to quote fixed price at all

Two honest cases. First, unknowable endpoints — genuine research, discovery, or "let's see what we find" work: quote hourly or day rate and sell the process. Second, clients who won't define scope — if they cannot describe what done looks like, no buffer protects you; qualify them out (saying no). And whatever structure you sell, get the payment terms working for you: deposits and milestones, as covered in late-payment terms that work.