Short answer: outcome-focused agencies are paid for a defined result, so their incentive is to build things that stay built. Effort-billed agencies are paid for time, so nothing in the model punishes rework. Both look the same in month three. They look very different in month eighteen, when someone who was not there has to change the system.
Follow the incentive
Two agencies, same brief, same rate card, different billing model.
The effort-billed agency has no structural reason to prefer the version that will be easy to change. It has no reason to invest a day in a test harness that saves five days later, because those five days are revenue. Nobody is being dishonest; the machine simply has no gradient pointing at maintainability.
The outcome-focused agency eats its own rework. If the data model cannot express the second pricing tier, that is their week, not yours. So they spend the extra half-day on the data model, and they ask the awkward question in scoping instead of discovering it in build.
You do not get maintainable software by asking for it. You get it by paying for a result rather than for time.
The five observable differences
1. Scoping is adversarial in a good way. Outcome-focused teams interrogate the brief before quoting, because ambiguity is their risk. Effort-billed teams accept vague briefs cheerfully, because ambiguity is billable.
2. Acceptance criteria exist. "Done" is written down before work starts, screen by screen and state by state, including the empty, offline and error states. Without that, done is a negotiation at the end.
3. Tests where they change the economics. Not coverage theatre — tests around the money paths, the data migrations and the integrations. An outcome team writes these because they will be the ones woken up.
4. Documentation aimed at a stranger. The reliable signal: a README a new engineer can follow to a running local environment, runbooks, and architectural decision records saying why, not what. Written for someone who has never met the author.
5. A real last week. Outcome-focused engagements end with a handover: knowledge transfer sessions, updated documentation, access transferred, a list of known issues written down honestly. Effort-billed engagements often end when the invoices stop.
What long-term reliability actually consists of
Reliability in year two is not uptime. It is changeability:
- Can a new engineer make a small change safely in their first week?
- Does the data model express the business as it is now, or as it was assumed to be at kickoff?
- Are dependencies patched, or is there a framework three major versions behind that nobody dares touch?
- Does anything alert before customers notice?
- Is there a document explaining why the strange bit is strange?
None of those are visible in a demo. All of them are set during the build by an agency's incentives.
Where outcome-focus has genuine limits
It is not a universal good, and pretending otherwise is a sales pitch:
- Genuinely exploratory research cannot be scoped to an outcome, and forcing it produces either padding or a fight. Bill that by time, deliberately.
- Fixed price with vague criteria is worse than time and materials, because it creates an incentive to deliver the minimum defensible interpretation.
- Outcome framing can undercut refactoring if the outcome is defined only as user-visible features. Put maintenance capacity in the contract explicitly — a standing share of each month for dependencies, debt and observability.
Three questions before you sign
1. "Who pays when an estimate is wrong?" One sentence, and it tells you the whole model.
2. "Show me a handover package from a finished project." Names removed. If one exists, they finish projects properly. If it takes a week to produce, it does not exist.
3. "What happens in the last week of an engagement?" Listen for knowledge transfer, documentation and access. Listen for silence.
How we structure it
Fixed scope against written acceptance criteria — €9,500 for a bounded first slice. Rework inside those criteria is ours. Changes are €450 a senior day, published, so changing your mind is a number and not a negotiation. Continuing work is a €7,000 a month dedicated senior team with a standing share reserved for maintenance, dependencies and observability rather than only new features.
Everything on the pricing page, the record in the work, and a contact form that reaches an engineer.
Frequently asked
What does outcome-focused actually mean?
The agency is paid for a defined result against acceptance criteria rather than for hours consumed. The practical consequence is that inefficiency costs them, not you — which changes every small decision about how carefully something is built.
Why does it show up in year two?
Because year one hides everything. Effort-billed work looks identical to outcome-billed work while a team is present. The difference appears when the system must be changed by someone who was not there, which is a test of documentation, tests and data modelling.
How do I check for it before signing?
Ask who pays when an estimate is wrong, ask to see a handover package from a finished project, and ask what they do in the last week of an engagement. The three answers together are hard to fake.
Does outcome-focus mean fixed price?
Not necessarily, but fixed price for bounded scope is the clearest expression of it. A retained team can be outcome-focused too, if it commits to results and publishes what it delivered each month.