Fixed-Price vs Time-and-Materials in 2026: A Decision Framework That Actually Holds Up

Every agency founder has a scar from the wrong pricing model. Either a fixed-price deal that ate three weekends and half the margin, or a T&M engagement where the client stopped picking up the phone around invoice four. In 2026, with AI-assisted estimation making bids feel more confident than they should, choosing the right contract shape matters more than ever.

This is the framework we use at 72Technologies to decide between fixed-price, time-and-materials, and the hybrid variants that sit between them. It's not academic — it's what we walk into scoping calls with.

The Three Models, Honestly Described

Every pricing conversation collapses into one of these three shapes, no matter what the SOW calls them.

Fixed-price means you've committed to a deliverable for a number. The client's risk is capped; yours isn't. You eat overruns, you keep underruns (in theory — most clients renegotiate if they smell one).

Time and materials means the client pays for hours worked at agreed rates. Your risk is capped; theirs isn't. They can pull the plug, but they also carry the cost of every discovery, pivot, and rewrite.

Hybrid models — capped T&M, phased fixed-price, retainer-plus-scope — try to share risk. Done well, they align incentives. Done badly, they combine the downsides of both.

That's it. Everything else is packaging.

The Four Variables That Actually Decide It

Before quoting, we score a prospect against four things. Not on a spreadsheet — usually on a whiteboard during a partner call.

1. Scope clarity

Can we describe the finished thing in a paragraph a junior engineer would understand? "Rebuild the checkout flow with Stripe, keep the current cart model, add Apple Pay" is clear. "An AI-powered platform for our industry" is not.

Low clarity is a T&M signal, or at minimum a paid discovery phase before any fixed number gets written down. We've seen agencies quote fixed on the second type of brief and lose six figures within a quarter.

2. Client sophistication

Has this client shipped software before? Do they have a product person, or a CTO who's going to review PRs? Non-technical clients almost always want fixed-price because it feels safer. It isn't — for either side — but it's what they'll ask for.

Sophisticated clients often prefer T&M because they understand that software estimates are lies with error bars. If you're working with a technical founder or an in-house engineering team, T&M or a capped T&M is usually the honest option.

3. Change velocity

How often will requirements change? A marketing site with a launch date is low velocity. A pre-PMF product still finding its shape is high velocity. Anything with "we'll figure it out as users respond" baked in is high velocity.

Fixed-price on high-velocity work turns your team into a change-request accounting department. We've had months where a project manager spent a third of their time writing and pricing CRs instead of shipping.

4. Relationship horizon

Is this a one-shot build or the start of a multi-year relationship? First engagements with new clients often benefit from a smaller fixed-price piece — call it a proof of trust — even if the rest of the work would be better as T&M. Once the relationship exists, you can move to T&M or retainer without spooking anyone.

A Simple Decision Tree

Here's roughly how the logic runs in our heads. It's not scripture, but it survives contact with reality.

if scope_clarity == HIGH and change_velocity == LOW:
    return FIXED_PRICE

if scope_clarity == LOW or change_velocity == HIGH:
    if client_sophistication == HIGH:
        return TIME_AND_MATERIALS
    else:
        return DISCOVERY_PHASE_THEN_REEVALUATE

if relationship_horizon == FIRST_ENGAGEMENT:
    return SMALL_FIXED_PRICE_PILOT

return CAPPED_TM  # the honest default

The interesting case is the middle branch. When a non-technical client wants fixed pricing on a fuzzy scope, saying "we do T&M" gets you rejected. Saying "let's do a paid two-week discovery, then I can give you a real number" gets you the deal about seven times out of ten in our experience.

Where Each Model Quietly Bleeds You

Every model has a failure mode that doesn't show up until month two.

Fixed-price: the specification arms race

To protect a fixed bid, you write increasingly detailed specs. Then you spend engineering time policing them. Then every ambiguity becomes a negotiation. The project stops being about shipping software and starts being about interpreting a document written before anyone knew what they were building.

The tell: your PM's calendar is full of "scope alignment" meetings by week four.

T&M: the trust tax

Clients on T&M scrutinise every hour. Reasonable ones ask smart questions. Anxious ones want daily standups they attend, weekly burn reports, and explanations for why a login screen took twelve hours instead of eight. You end up with a second full-time job justifying work.

The fix is transparency by default — shared boards, honest weekly summaries, no surprises on invoices. But it's a real cost, and you should price it in.

Hybrid: the worst-of-both trap

Capped T&M sounds great: client gets a ceiling, you get flexibility. In practice, the cap becomes a fixed price the moment the project gets close to it, and the client remembers the ceiling but forgets the "capped" was theoretical. If you use caps, make the reset conditions explicit and in writing, and revisit them at defined checkpoints.

What to Put in the Contract

Regardless of model, a few clauses do most of the heavy lifting when things go sideways.

• Assumptions block. List every assumption the estimate rests on. Third-party APIs work as documented. Client provides design assets by X date. If any assumption breaks, the estimate is void and gets re-quoted. This one clause has saved us more than any other.
• Change control mechanism. How changes are logged, priced, and approved. Written, not verbal. Same email thread, same format every time.
• Payment cadence tied to milestones, not calendar. Monthly invoicing on fixed-price deals is a slow-motion cashflow disaster. Tie payments to demos or deployable artifacts.
• Kill clauses on both sides. What happens if either party walks. Notice period, work-in-progress payment, IP handover conditions.
• AI-generated content clause. New in the last couple of years — who owns what, what tools were used, and what warranties you can actually give. Worth a paragraph.

Reading the Room in the Sales Call

The fastest signal isn't in the RFP — it's in how the client answers one question: "What happens if halfway through we discover we need to build this differently?"

A good answer sounds like: "We'd want to understand the tradeoffs and decide together." That client can handle T&M or a hybrid.

A bad answer sounds like: "That's why we're paying you a fixed price — so we don't have to worry about that." That client needs either a very tight fixed scope with an ironclad assumptions block, or they need to be politely declined.

We've walked away from deals on that answer alone. It costs revenue in the short term and saves a project in the medium term. If you want to see how we structure the discovery work that precedes these decisions, our services page breaks it down, and there's more on scope conversations across our blog.

Where We'd Start

If you're rebuilding your pricing approach this quarter, don't rewrite your SOP first. Do this instead:

1. Pull your last ten projects. Score each on the four variables above at the time of signing, not with hindsight.
2. Mark each as profitable, break-even, or loss.
3. Look for the pattern. Almost every agency we've talked to finds the same thing: fixed-price wins on high-clarity, low-velocity work, and loses everywhere else — but they keep quoting fixed because clients ask for it.

The fix isn't refusing fixed-price. It's getting honest about when it's the wrong tool, and building the discovery muscle to move clients toward the model that actually fits the work. That conversation is uncomfortable the first few times. It's cheaper than the alternative.