The short answer
For any deliverable you can describe in two or three sentences, sprint-fixed wins on total cost of ownership. Hourly is a research model, not a delivery model. The mistake most clients make is treating the headline rate as the comparison; the right comparison is the final invoice on the same scope.
The hourly billing incentive problem
An agency’s revenue under hourly billing is hours times rate. The agency’s margin grows with hours. Therefore the agency optimizes for hours rather than for shipping. This is not malicious; it’s structural. Even ethical, well-intentioned agencies under hourly billing find themselves in the position where every conversation about whether to spend more time polishing a feature has a financial pull toward “yes”.
The client is on the other side of that equation. The client’s incentive is to ship. Under hourly billing, the client and the agency are misaligned by default. Good agencies counteract this with discipline, transparency, and explicit scope reviews; weaker agencies don’t, and the project drifts.
The sprint-fixed incentive problem
Sprint-fixed billing inverts the alignment. The agency carries scope risk: once the contract is signed, every additional hour the agency spends on the sprint is a hit to the agency’s margin. The agency therefore optimizes for tight scope upfront and fast delivery during the sprint.
The risk this introduces is that an agency under sprint-fixed pressure might cut corners. The mitigation is structural: senior-only delivery (less likely to ship shortcuts), weekly demos (forcing function for transparent progress), and a written change-order process (so genuine scope changes are negotiated rather than absorbed). Without these three, sprint-fixed is unworkable. With them, it’s the cleanest contract shape for defined-scope work.
The TCO breakdown
Here is the cost of a representative AI MVP across both billing models, over twelve months, including the post-launch iteration phase. The numbers are averaged across a sample of recent engagements; your mileage will vary, but the relative shape is consistent.
| Stage | Hourly agency (avg) | Sprint-fixed agency |
|---|---|---|
| Discovery (paid, both) | $5,000 | $1,500 |
| MVP build | $90,000 to $180,000 (overrun) | $18,000 |
| Post-launch fixes (30 days) | $25,000 | Included |
| Iteration sprint 1 | $35,000 | $12,000 |
| Iteration sprint 2 | $35,000 | $12,000 |
| Total | $190,000 to $280,000 | $61,000 |
The 3× to 4.5× gap is not because hourly engineers are slower. The same engineers, working on the same scope, ship the same product. The gap is entirely accounted for by three structural causes: scope drift, optimistic estimation, and the absence of a change-order discipline.
Why hourly overruns by 1.5× to 2×
Three forces compound to produce the typical overrun. None of them is individually large. Stacked across a 12-week engagement, they explain the entire gap.
Scope drift. Every conversation that starts with “while you’re in there, can you also…” adds hours. Under hourly billing, those hours bill at full rate. There is no mechanism that forces the client to decide whether the addition is worth it; the only mechanism is integrity on both sides, and integrity is not a substitute for structure.
Optimistic estimation. Initial estimates under hourly billing are floors, not ceilings. The agency has no contractual incentive to be conservative — overruns get billed. So estimates skew optimistic, and reality re-prices them upward. By the time the project is halfway done, the original estimate is a memory.
No change-order discipline. Without a written process for scope changes, every change is a verbal negotiation. Verbal negotiations favor whoever talks more, which under most engagement structures is the agency. Sprint-fixed forces every change into a written change order with a new price, a new timeline, and mutual sign-off. The bureaucracy is the feature.
How sprint-fixed forces scope discipline
Under sprint-fixed billing, scope is the contract. The agency cannot ship feature creep without a change order because every new feature reduces the agency’s margin on the existing sprint. The client cannot expand scope without paying because the change-order process is the gatekeeper.
This sounds adversarial. In practice it’s the opposite: both sides know the rules going in, and the conversation moves to substance instead of procedure. “Should we add this feature?” becomes a clear question with a clear price tag, not an ambiguous request that may or may not show up on next month’s invoice.
The change-order pattern in practice
Here is the rule we use at YATE Web, and we recommend the same shape regardless of agency. Any change exceeding 5% of the original scope requires a written change order with three components: the new total price, the new timeline, and mutual signature. Smaller changes (typo fixes, copy adjustments, color tweaks) flow through a lightweight approval process; substantive changes go through the formal one.
Two patterns matter. First, the change order is signed before work begins, not after. Second, the change order resets the relevant sprint price, not the project price — so the agency can’t accumulate small changes into a large invoice surprise. Both patterns protect both sides.
When hourly is genuinely the right call
We charge hourly in some engagements, and we recommend it to clients in specific cases. Three signals indicate hourly is correct:
True research with no defined deliverable. When the goal is to learn rather than ship — exploring what an agent can do, prototyping architectures, evaluating model options before scoping a build — hourly maps to the actual exchange. Sprint-fixed presupposes a deliverable; if there isn’t one, sprint-fixed is the wrong contract shape.
Maintenance retainers under 20 hours per month. Trickle work where any sprint cadence would over-charge. We do this on optional retainers after launch when the client wants ongoing senior availability without a committed scope.
Pair-programming with an internal team. When the deliverable is your team’s growth — knowledge transfer, code review of internal work, architectural mentorship — hourly maps to the actual engagement shape. The engagement is engagement-shaped, not project-shaped.
Common mistakes on both sides
Five recurring mistakes account for most of the unhappiness on hourly engagements, and a smaller but real set of unhappiness on sprint-fixed ones.
Treating hourly as flexible. Hourly is non-committal, not flexible. Real flexibility is a written change-order process; the absence of one is just ambiguity wearing a friendlier label.
Treating sprint-fixed as inflexible. Sprint-fixed has change orders. The change-order process is what makes scope evolution honest. The mistake is treating the contract as a wall rather than a structure.
Choosing hourly because the rate is lower. The rate is one number; the final invoice is the only number that matters. Hourly rates are often 20 to 30% lower than sprint-fixed equivalents and the final invoice is still 30 to 60% higher.
Choosing sprint-fixed without confirming scope alignment first. Sprint-fixed only works if the scope is well-understood. Sign sprint-fixed against an ambiguous scope and you push the ambiguity into the change-order process, where it becomes a series of small painful negotiations.
Not requesting both quotes. When you’re evaluating agencies, ask the same agency for both an hourly and a sprint-fixed quote on the same scope. The comparison teaches you a lot. Agencies that can’t quote sprint-fixed often can’t scope tightly enough to commit.
A worked example
A Series-A SaaS company adding RAG-based customer support. Defined scope: integrate with Help Scout, four retrieval sources, evaluation set, 30-day monitoring. Both quotes received in writing.
| Path | Quoted | Actual | Variance |
|---|---|---|---|
| Hourly agency | $60K to $90K, 8-12 weeks | $135K, 16 weeks | +50% cost, +33% time |
| Sprint-fixed agency | $45K, 6 weeks | $45K, 6 weeks | 0% cost, 0% time |
The hourly agency’s engineers were genuinely good. The agency itself was genuinely well-meaning. The structural pull of the billing model carried the engagement off-track, and the absence of a change-order discipline made the drift invisible until it was too late.
Frequently asked questions
Is sprint-fixed pricing always cheaper than hourly?
On the same scope, in our experience, yes — by 30 to 60%. The exception is exploration-only work where there is no defined deliverable; for that, hourly maps better to the actual exchange. We charge sprint-fixed for the discovery workshop precisely because the output (signed scope, cost model, go-or-no-go) is concrete, not open-ended.
What if scope changes mid-sprint?
If we missed it during scoping, it is on us. If you change it, we either pause the sprint and re-scope, or carry the change into the next sprint. The current sprint price doesn’t move; the next sprint price reflects the new scope. This is what a written change-order process looks like in practice.
What about overruns? Doesn’t sprint-fixed encourage shortcuts?
It can, which is why senior-only delivery and weekly demos exist as the counterweight. Senior engineers don’t ship shortcuts as readily, and a Friday demo with metrics surfaces shortcut behavior before it ships. The same risk under hourly billing exists; it just shows up six months later in the maintenance bill.
Why isn’t sprint-fixed standard at every agency?
It pushes risk onto the agency. Most agencies prefer to externalize that risk; hourly billing makes that easy. Sprint-fixed requires confident scoping, senior-only delivery, and a methodology library to compress estimation variance. Without those three, sprint-fixed is unworkable, which is why most agencies don’t offer it.
Do I see hour breakdowns under sprint-fixed?
If you ask, yes. We log internally so we can keep our estimates honest. The price you pay does not depend on those numbers, but you can see them. There is nothing to hide.
When does hourly genuinely win?
Three cases: pure exploration where the deliverable is unknown, maintenance retainers under 20 hours per month, and pair-programming engagements where knowledge transfer is the deliverable. For new builds with a defined scope and a deadline, hourly almost never wins on TCO.
The bottom line
For any deliverable you can describe clearly, sprint-fixed wins on TCO, calendar time, and predictability. Hourly is the right model for genuinely undefined research and for long-tail maintenance — both engagements where there is no fixed deliverable to anchor a sprint price against. The mistake most clients make is choosing the wrong contract shape because the rate looks lower; the contract shape, not the rate, determines the final invoice.
If you’re trying to understand which billing model fits your specific project, the free Product Audit returns a scoped sprint-fixed quote with a written timeline, three integration options, and one “don’t build this” recommendation in 48 hours.