Pricing & engagement
Five ways to work together, compared honestly
The model matters more than the rate. Below is what each one is good at, what it is bad at, and who carries which risk — including the cases where we will talk you out of the one you asked for.
The short version
Side by side
| Model | Best for | Who carries the risk | Flexibility | Typical duration | Indicative rate |
|---|---|---|---|---|---|
| Fixed-scope sprint | A well-understood problem you want solved without opening a long engagement. | We carry delivery risk within the agreed scope; changes to scope change the price. | Low during the sprint, by design — the scope lock is what makes the price firm. | Two to six weeks | Pending commercial sign-off |
| Time and materials | Evolving work where the next priority depends on what the last one taught you. | You carry scope risk; we carry the risk of the rate being honest about seniority. | High — priorities can change at any sprint boundary. | Three months or more | Pending commercial sign-off |
| Dedicated team | Sustained product development where continuity of context is worth more than flexibility of headcount. | Shared — you own the roadmap, we own delivery capability and staffing continuity. | Moderate — team shape changes at agreed notice, typically one month. | Six months or more | Pending commercial sign-off |
| Staff augmentation | A capable in-house team that needs specific capacity or a specific skill, not a vendor. | You own delivery and process; we own the calibre of the person and replace poor fits at our cost. | High — engineers roll on and off at agreed notice. | Three months or more | Pending commercial sign-off |
| Outcome-based | Problems with a measurable target — latency, deployment frequency, conversion, recovery time. | We carry real financial risk on the outcome, which is why we qualify these engagements hard. | Low on the metric, high on the method — how we hit it is our problem. | Three to twelve months | Pending commercial sign-off |
The rate column reads the same on every row on purpose. A band goes on this page once it has an internal owner and a signature behind it, and not before — a number we would have to revise is worth less to you than an empty column.
In detail
What each model is actually like
Each one names its failure mode as well as its strength. Where two look plausible for the same problem, the difference is almost always who should be carrying scope risk.
- 01
Fixed-scope sprint
A bounded piece of work with a defined outcome, price, and end date.
Discuss this modelBest when you can describe the finished state precisely: a payment provider integrated, a performance target met, a service extracted from a monolith. We scope it, price it once, and deliver against that.
If discovery reveals the problem is materially different from the brief, we stop and re-scope rather than quietly absorbing the difference or billing through it.
- 02
Time and materials
Billed for the hours worked, with a rate card and a monthly ceiling you set.
Discuss this modelThe default for product work where the roadmap is genuinely uncertain. You get a rate card by seniority, a monthly spend ceiling you control, and a burn report each week showing what the money bought.
The failure mode of time and materials is a vendor with no incentive to finish. We publish velocity and spend against outcome, not hours logged, so the trade-off stays visible.
- 03
Dedicated team
A standing cross-functional team, allocated to you and no one else.
Discuss this modelA named team with a tech lead, allocated to your product full time. They attend your ceremonies, hold your domain context, and are not rotated onto other accounts.
We publish the named people and their allocation percentage in the contract. If someone has to change, you get notice and an overlap period rather than a silent substitution.
- 04
Staff augmentation
Individual senior engineers embedded in your existing teams, under your process.
Discuss this modelYour standups, your board, your definition of done. We supply engineers who can operate at your level without supervision, and we screen for the specific stack rather than sending whoever is on the bench.
If a placement is not working, tell us inside the first four weeks and we replace them without charging for the ramp.
- 05
Outcome-based
A share of the fee is contingent on hitting a metric we agree and instrument together.
Discuss this modelThe most distinctive way we work, and the one we say no to most often. It only functions when the metric is unambiguous, measurable before we start, and genuinely within our control.
A workable structure looks like: a baseline measured jointly over four weeks, a target with a date, a base fee covering costs, and a bonus tranche released on the metric. A metric that depends on your sales team or a third party's roadmap is not a metric we can carry.
We will decline outcome terms where the baseline cannot be measured, where the target depends on decisions outside the engineering scope, or where the data needed to verify it is not something both sides can see.
The distinctive one
How outcome-based contracts actually work
Most agencies that advertise outcome-based pricing mean a discount structure. We mean a fee tranche that we do not receive if an agreed, jointly instrumented metric does not move. That is a real risk to carry, which is why the qualifying conditions below are strict rather than decorative.
What has to be true
- The metric is measurable today, using instrumentation both sides can see.
- A baseline can be established jointly before any code changes, typically over four to six weeks.
- Movement in the metric is genuinely within engineering control.
- The target has a date, and both sides agree what “achieved” means in advance.
A worked example
On the Zafarin reconciliation engagement, the metric was the count of end-of-day breaks left for manual clearing — emitted by the existing run log every business day, and already the figure finance operations signed the safeguarding attestation against. We spent the first six weeks measuring and classifying rather than building so that “before” was a number both sides had signed off, then put a fee tranche against the target.
When we say no
We decline outcome terms where the baseline cannot be measured, where the target depends on decisions outside the engineering scope, or where verifying it needs data one side cannot see. That is most requests, and saying so early is cheaper for everyone than discovering it at the first invoice.
Not sure which model fits?
Describe the problem and the constraint you care most about — budget certainty, speed, or flexibility — and we will tell you which of the five we would propose, and why.