Fixed-Price vs Time-and-Materials: Which Protects Your Margin on Development Work

There's a version of the pricing conversation that happens in the sales meeting, and a different version that happens in the margin review six months later. In the sales meeting, fixed-price versus time-and-materials sounds like a question of what the client is comfortable signing. In the margin review, it turns out to have been the single decision that determined whether the project's overrun came out of your pocket or theirs.

Fixed-price vs time-and-materials is not, at root, a sales question. It is a question of who carries the risk that the estimate is wrong — and estimates are wrong often enough that the answer matters more than the rate on the contract. I've written before about why your estimates are thirty percent off; this piece assumes that's true and asks the next question: given that estimates drift, which contract structure protects your margin when they do.

This is for the owner or finance lead choosing a default contract model, or deciding which model a specific engagement should use. It's deliberately separate from the estimating posts — those are about making the number more accurate. This one is about who absorbs the error when, inevitably, it isn't.

The contract model is a margin decision

Start with what each model actually does to the risk of a wrong estimate.

Fixed-price quotes one number for a defined scope. The client pays that number regardless of how many hours the work takes. That means every hour over the estimate is a cost the agency eats, straight out of margin, and every hour under the estimate is margin the agency keeps. The agency has sold a guarantee; the price of issuing a guarantee is carrying the risk behind it.

Time-and-materials bills actual hours at agreed rates. An overrun is invoiced, so it lands on the client, not the agency. The agency's margin per hour is locked in by the rate card and barely moves. What moves is the total — and the upside is capped, because you can only bill the hours you work. You can't be more profitable than your rate; you can only be busier.

So the two models don't just price differently, they shape the margin differently:

• Under fixed-price, margin is variable and driven by execution against the estimate. Run to plan and it's healthy; run over and it falls fast.
• Under time-and-materials, margin is stable per hour and driven by your realised rate and utilisation. It rarely collapses, and it rarely soars.

That single difference — variable-but-high-ceiling versus stable-but-capped — is the whole decision. Everything else is detail about when each shape is the one you want.

The same build, priced both ways

Take a build (anonymised, scaled to round numbers): a ~£100,000 development engagement, estimated at 1,000 hours of delivery across a lead, two mids, and QA. Assume the agency's real cost to deliver those planned hours — labour at named-person cost plus allocated overhead, the way project margin is actually calculated — comes to ~£68,000 when the project runs to plan. That's a planned real margin of about 32%.

Now run it two ways: on plan, and 15% over (an extra 150 hours, which is well inside the normal range of estimate error).

Fixed-price

• On plan. Revenue £100,000, real cost £68,000, margin £32,000 / 32%. The estimate held; the agency keeps exactly what it priced.
• 15% over. Revenue is still £100,000 — the client agreed a fixed number. But the extra 150 hours cost real money to deliver: at a blended real cost of ~£55/hour that's ~£8,250 more. Real cost rises to ~£76,250. Margin falls to £23,750 / ~24%. The whole overrun came out of margin, an 8-point drop, because the agency carried the risk.

Time-and-materials

• On plan. 1,000 hours billed at the agreed rates produce ~£100,000 of revenue against ~£68,000 of cost. Margin £32,000 / 32% — the same as fixed-price on plan, because the plan is the plan.
• 15% over. The extra 150 hours are billed. Revenue rises to ~£115,000; real cost rises to ~£76,250. Margin is £38,750 / ~34%. The overrun didn't hurt the agency — it was invoiced. If anything the margin percentage held and the absolute margin grew.

The table the example hides is the symmetric one: what happens when the project comes in under estimate. Fixed-price keeps the saving — finish in 850 hours and the margin climbs above 32%. Time-and-materials gives it back — you bill 850 hours, revenue falls, margin in absolute terms shrinks. Fixed-price has the higher ceiling and the lower floor. T&M has neither.

When each one protects margin

The worked example shows the mechanism. The decision is about which mechanism suits the engagement in front of you. Three variables decide it.

Scope certainty

This is the dominant one. If the scope is genuinely nailed down — clear deliverables, low discovery risk, a build you've done close cousins of before — fixed-price protects margin, because your estimate is likely to hold and you keep the upside of running efficiently. If the scope is vague, exploratory, or likely to change mid-flight, fixed-price is a trap: you're guaranteeing a number against work you can't yet see, and the overrun risk sits entirely with you. That's where T&M protects margin, because the meter follows the scope wherever it goes.

Client maturity

T&M only protects margin if the client can live with a running meter without relitigating every timesheet. A mature client who understands that they carry scope risk in exchange for flexibility will let T&M work. A client who treats every invoice as a negotiation will turn T&M into a series of disputes that cost you senior time to defend — and that scrutiny is itself a cost the model invites. With that kind of client, a fixed-price number they signed once is often the calmer, more defensible position, even if it carries more risk for you.

Project shape

Short, well-defined pieces of work suit fixed-price — the estimate window is small and the risk is bounded. Long, evolving programmes suit T&M or a hybrid, because no honest estimate survives twelve months of changing priorities. The longer and more uncertain the engagement, the more the risk of a fixed number compounds against you.

There's also the quiet failure mode of pure T&M: it rewards slowness. If every hour is billed and nobody is watching realised efficiency, a team has no structural incentive to finish early — finishing early lowers the invoice. That's not dishonesty, it's incentive design, and it's why unmanaged T&M can erode the client relationship even while the agency's per-hour margin looks fine. You manage it the same way you'd manage any utilisation question: by watching delivered value against hours, not by trusting the meter.

The hybrids most agencies actually use

In practice, the binary choice is rarely the right answer. Two hybrids carry most of the engagements at a well-run agency.

• Capped T&M. Bill by the hour, but agree a ceiling the invoice won't exceed without a change conversation. The client gets the flexibility of T&M and the protection of a known maximum; the agency carries overrun risk only up to the cap, not infinitely, and the cap conversation forces scope discipline. This is the default I'd reach for on most uncertain engagements with a client I trust.
• Phased fixed-price. Break a long programme into phases, fix the price of each phase only once its scope is clear, and run a short paid discovery before each fix. You get the margin upside of fixed-price on work you can actually see, without guaranteeing a number for work that's still twelve months of assumptions away. The risk window shrinks to one phase at a time.

Both hybrids do the same thing: they keep the fixed-price upside where scope is certain and shift to a billed-hours model where it isn't, phase by phase. That's the honest version of the decision — not "which model do we prefer", but "where is the scope certain enough to guarantee, and where isn't it".

One more variable is reshaping this faster than most contracts have kept up with: AI-assisted delivery. When a build that used to take 1,000 hours now takes 700, fixed-price quietly captures that gain for the agency while T&M hands it straight back to the client as a smaller invoice. If your delivery is getting faster, your contract model decides who keeps the saving — which is one more reason to treat it as a margin decision rather than a sales habit.

The shorter version

Fixed-price and time-and-materials are two answers to one question: who carries the risk that the estimate is wrong. Fixed-price puts it on you, with a high ceiling and a low floor; T&M puts it on the client, with a stable margin and a capped upside. The right choice depends on how certain the scope is, how mature the client is, and how long the engagement runs — and for most real engagements the answer is a hybrid that fixes the price where scope is certain and meters it where it isn't.

What doesn't change is that you can only make this call well if you know your real margin under each model, on each project type, on your own data. If you'd like to see how your engagements have actually behaved — fixed-price overruns that ate margin, T&M projects where the realised rate told a different story than the rate card — Saldo computes it continuously on your real Jira. Pricing, including the per-engagement maths, is on the pricing page. Pick the model that protects the margin, not the one that's easiest to sell.