How we engage
A commercial model is not an administrative detail. It decides what both sides are rewarded for, and most disappointing engagements can be traced back to a structure that quietly paid someone for the wrong thing.
The model shapes the behaviour
Fixed price rewards a supplier for narrowing scope and resisting change. Time and materials rewards duration. Per-ticket support rewards recurring tickets. None of this requires bad faith — it is simply what the incentive is, and over a long engagement incentives win against good intentions.
The useful response is to pick the model that fits the shape of the uncertainty, and to say out loud what it will push us toward. Where a model's incentive works against you, that is worth naming in the contract rather than discovering in month four.
The four we use
Fixed price
Suits genuinely well-defined work — a migration with a known inventory, a bounded integration. Requires enough discovery to price honestly, and makes change requests adversarial by design. We will decline fixed price on work that is not actually specified.
Time and materials
Suits work where the requirements will move, which is most product development. Needs a cap or a review cadence so it does not become open-ended, and it puts the estimating risk on you rather than on us.
Embedded team
A monthly rate for named people working inside your process, managed by your leads. Best when the work is continuous and priorities shift week to week. Predictable cost, and you carry the direction.
Managed service
A fixed monthly fee for operating something to agreed measures, rather than per ticket. Priced so that eliminating a recurring failure is in our interest too.
How we choose one with you
The question we ask first is how much is genuinely unknown. High uncertainty and fixed price is a trap for both sides — we pad the estimate, you pay for the padding, and every discovery becomes a negotiation. Low uncertainty and time and materials means you carry a risk you did not need to.
Where uncertainty is high but a fixed cost is genuinely required, the honest structure is a paid discovery phase that ends in a real estimate, with no obligation to continue with us. That is usually two to four weeks and it is the only way a fixed price on unfamiliar work means anything.
Discovery phase priced separately, with no obligation to continue
Named people, not a rotating pool
Change handled by an agreed cadence rather than by exception
Notice period and handover scope agreed at the start
Your repositories, your accounts, your infrastructure throughout
How an engagement runs
A call, not a deck
You describe the system and the constraint. We say what we would do first, what we would not do at all, and whether we are the right people. Sometimes the answer is no, and that call is free.
Discovery, if it is warranted
Two to four weeks, paid, ending in a written plan with a shape and a number. If the plan concludes the project should not go ahead, the engagement can end there and you have saved a great deal more than it cost.
A small first slice
Something real, deployed to your infrastructure inside the first few weeks. It proves the pipeline, the access and the working relationship while the stakes are still low.
Cycles in the open
Weekly, against real data, with your engineers reviewing our pull requests from the first one. Nothing is demonstrated on a branch that has not been deployed.
A handover with a date
Agreed at the start, not negotiated at the end. Your team runs a change end to end while we are still available to answer questions.
Common questions
For genuinely specified work, yes. For product development with moving requirements we will usually decline, because pricing it fixed means either padding the number or arguing about scope later. Where you need cost certainty on unfamiliar work, we would propose paid discovery first, then a fixed price on what that produces.
Related work
Tell us what you are trying to ship.
One call, no deck. Describe the system and the deadline, and we will tell you what we would do first — including when the answer is that you do not need us.