There are only two classic ways to buy outsourced delivery, and most agency owners have tried both. You rent people, or you buy projects. Each fails in a way you could have predicted before you signed.
The interesting part is that both models fail for the same underlying reason: they force you to commit to a shape of work months before you know what your clients will actually ask for.
Where both classic models break
Dedicated people: you buy the gaps too
A dedicated developer or designer gives you certainty about who does the work. It also gives you their idle days. Three quiet days in a slow week are still invoiced, and you cannot lend that person to a different supplier or a different month. In practice most agencies keep a dedicated resource around 65–75% utilised. You are paying full price for three quarters of a person.
Per project: you re-sell your own supplier every month
Buying by the project fixes the idle-time problem and creates a worse one. Every new scope is a new quote, a new negotiation, a new set of assumptions about revisions. Your delivery cost becomes a variable you cannot forecast, and the person doing the estimating is usually your most expensive person.
Both models make you decide in advance whether next month needs a designer or a developer. Your clients never cooperate with that decision.
The hybrid: buy the month, not the job title
The alternative that has quietly become standard in mature outsourcing is a capacity retainer. You buy a fixed block of delivery capacity every month — measured in specialist days, not names — and you decide during the month what it gets spent on. The supplier decides who does it.
The mechanics are simple:
- Capacity is measured in a neutral unit. One credit equals one specialist day. Heavier roles carry a multiplier so the maths stays honest.
- You choose the mix as the month unfolds: two landing pages in March, SEO for three clients in April, a rebrand and a migration in May.
- The price does not move. Revisions, QA and project management are already inside the credit.
- Unused capacity rolls over within a cap, so a slow month is not a total loss and a fast month cannot be gamed.
Why it is better for you specifically
Three things change the moment delivery becomes a flat monthly line:
- Your cost of delivery becomes forecastable. You can price a retainer to your client knowing exactly what it costs you to service, because the number does not move with scope.
- You stop negotiating internally. No more debating whether a small fix is billable. It comes out of the pool, the pool is already paid for, and the conversation never reaches your client.
- You can say yes faster. The most expensive thing an agency does is decline signed work because the team is full for six weeks. Capacity you already own removes that decision.
The partner running the capacity pool absorbs the utilisation risk. That is the actual trade. You pay a flat fee; they keep their people busy across several partners instead of parking them on yours.
When capacity is the wrong purchase
It is not universal. Two situations where a different model wins:
- You want a specific person. Once your team knows a particular developer by name and wants them permanently, buy dedicated. Capacity retainers work precisely because you are not attached to individuals.
- Your work is one very large, very technical build. If a single project will consume the entire pool for four months, price it as a project. Capacity is designed for a stream of varied work, not one monolith.
For everything in between — a stream of client work with unpredictable shape and a very predictable need for it to ship — capacity is the cheaper, calmer purchase. You stop buying people and start buying the thing you actually sell: finished work, on a date.
