You have a technical outcome to achieve — an integration, a rebuild, a workflow fixed properly. Do you employ someone, hire an agency, or bring in fractional senior engineers?

Most advice on this question comes from people selling one of the three answers. This comparison is not. Each option is right in some situations and wasteful in others, and the difference is rarely quality — it is the shape of what you are buying: fixed cost or variable, capacity or outcome, continuity or speed. Work out your situation's shape, and the decision mostly makes itself.

Option one: employ someone

Employing a developer buys control and continuity. You own the person's time, they learn your business deeply, and they are there every day. For sustained engineering workload, this is often the right answer, and an honest comparison should say so first.

The costs are real, though less visible than a salary:

  • Recruitment cost and risk. Finding a good hire takes time; a mis-hire costs months. Judging candidates is a skill a non-technical employer may lack.
  • Single-skill risk and key-person dependency. One person holds one skillset and, over time, all the knowledge; the day they leave, you are back to zero.
  • Management overhead. Someone must set direction and review work; if nobody can, an employee can drift unnoticed for a long time.
  • Under-utilisation between projects. Work comes in waves; between pushes, a salaried developer still costs the same — and good people leave when the work is thin.

Employ when the workload is sustained and predictable, most weeks of the year.

Option two: a traditional agency

An agency buys capacity and range: a team appears with varied skills and disappears when the project ends. For large, well-specified builds — a full product, a substantial platform — it is often the right answer too.

The shape of the arrangement deserves scrutiny:

  • Account layers between you and the doers. The people you meet at the pitch are rarely the people building; questions come back translated through account management.
  • Blended teams. Senior and junior staff blend into a single rate. You pay senior rates for junior learning, and often cannot tell who did what.
  • Scope shaped to the agency. Proposals are scoped in units the agency sells — phases, retainers, team-months — not units you buy, which are outcomes.

None of this makes agencies bad. It makes them the right tool for large, well-specified work where blended capacity is an acceptable price for range you cannot build yourself.

Option three: fractional, FORT-style engagement

A fractional engagement buys senior capability on defined outcomes rather than time. At Cooper Agile this takes the form of FORT cycles: a named senior engineer takes one bounded outcome per fortnight, works to acceptance criteria agreed up front, shows working evidence throughout, and hands over with documentation and training. Larger programmes are several linked FORTs.

What this shape buys you: senior people doing the work, with no blended rate and no account layer; outcome-shaped scope — "the integration works, tested, handed over", not a block of days; fixed cycles with working software and traceable decisions at the end of each; and handover by design, with capability landing in your team rather than in the engagement.

Its limits deserve the same honesty: it is not a 24/7 product team — sustained daily engineering over years calls for employment, not fractional. And it depends on good scoping: a FORT is only as good as the conversation that bounded it, which is why scoping is a deliverable in its own right.

The same brief, three ways

Illustrative example: a 30-person firm wants website enquiries to flow automatically into its CRM, deduplicated and assigned, with an acknowledgement email. Handled by an employee: recruit first, then the hire learns the estate, then the work takes a fraction of their month — after which you carry a salary through quieter periods, and the knowledge concentrates in one person. Handled by an agency: the brief becomes a phased proposal with an account manager; done well, but by a blended team at a blended rate. Handled as one FORT: the outcome is scoped with acceptance criteria, built and tested by a named senior engineer in two weeks, demonstrated live, handed over with a runbook and training. The point is not that one is cheapest — without prices, that claim would be invented. It is that the three shapes buy three different things, and only one of them is the outcome itself.

Which is right when

  • Sustained, predictable workload you can direct? Employ — control and continuity win.
  • Large, well-specified build needing varied capacity? Agency — range and scale win.
  • Bounded outcome with a measurable end-state? Fractional — senior capacity on the outcome, without the hire.
  • Workload too thin to keep an employee busy? Fractional or agency first; under-utilisation is a slow leak.
  • Nobody internally able to direct or judge technical work? Solve that first, whichever route you take.

Constraints and trade-offs

Three honest caveats before you decide:

  • Fixed cycles dislike vagueness. If you cannot say what "done" looks like, no fortnight-shaped engagement will save you; spend first on definition, not delivery.
  • Handover needs a receiver. Capability transferred to a team with no one to hold it decays; every option needs someone on your side who understands what exists.
  • Cheapest depends on timeline. A hire pays back over years, an agency over a large build, a FORT over weeks; one month's cost is the wrong comparison.

Your action checklist

  • Describe the outcome you need, in business terms, before choosing a route.
  • Judge whether the workload is sustained or lumpy — that alone eliminates one option.
  • Check whether anyone internally can direct and judge technical work; fill that gap first if not.
  • For bounded outcomes, write acceptance criteria before talking to anyone selling time.
  • Decide on shape — capacity, continuity, or outcome — then negotiate within it.

If you would rather not decide on instinct, our technical review is deliberately neutral: we pressure-test your workload shape, internal capability and acceptance criteria, and tell you honestly which route fits — even if the answer is not us. If it turns out to be fractional, our fractional capability page shows how FORT cycles are structured.