ยท 7 min read
Fractional CTO, agency, or full-time hire: which one an SME actually needs.
Most companies of 40 to 400 staff that depend on software need a fractional CTO before they need anything else: someone who owns the roadmap a few days a month, with an agency or a small team doing the building. A full-time CTO becomes the right answer once software is the product or the in-house team heads past 15 to 20 engineers. An agency is never the answer to the ownership question, because it cannot grade its own work. Here is how the three compare on cost, ownership, failure, and exit.
The short answer
The three options sell different things, so they are not really alternatives. An agency sells capacity: people who build what you specify. A fractional CTO sells judgement and ownership for part of the week: someone who decides what gets built, grades the people building it, and answers for the result. A full-time CTO sells the same judgement every day, at the price of a permanent senior salary and a search that takes most of a year.
The question that sorts them is simple. Who owns the roadmap? If the honest answer is "the agency", nobody on your side does. If it is "a full-time CTO", check that there is a full-time job to own. For most owner-led companies of 40 to 400 staff, where software runs the business but is not what the business sells, the answer that fits is a fractional CTO a few days a month, with an agency or a small internal team doing the building.
| Agency | Fractional CTO | Full-time CTO | |
|---|---|---|---|
| Cost shape | Variable. Billed per sprint, per profile, or per project. Grows with the scope it sells you | Day rate times days. 2 to 4 days a month up to 2 to 3 days a week. Stops when you stop it | Fixed. Salary, employer costs, bonus, car, pension. Plus a search of 6 to 9 months before it starts |
| Who owns the roadmap | The agency proposes, you approve without being able to grade it | You do, through the fractional CTO, who writes it and defends it | The CTO |
| When it goes wrong | You find out late, from a missed date or an invoice. The fix is a contract dispute | You find out within a quarter, because the work is written down and reviewed. The fix is a few weeks' notice | You find out after 6 to 12 months. The fix is a severance and a new search |
| Exit cost | Code access, deployment keys, and knowledge that lives in the supplier. A rebuild if nobody owned the architecture | Low by design. The handover is part of the scope | Notice or severance under Belgian or Dutch employment law, and an empty seat while you search again |
| The situation it fits | Well-specified building work, with an owner on your side | Software is load-bearing but not the product, a supplier needs grading, a decision needs making | Software is the product, or the in-house team is heading past 15 to 20 engineers |
What each option actually sells you
An agency sells capacity. What it cannot sell you is an independent view of its own work. Its commercial interest is more scope, and that is not dishonesty, it is the model. That works for a clearly specified build. It fails for a roadmap, because a roadmap is a series of choices about what not to build, and nobody bills for what they talk you out of.
A fractional CTO sells ownership on a schedule. A few days a month, or 2 to 3 days a week, of someone who has run engineering before and now runs yours. The work is concrete: one roadmap, the architecture decisions that span suppliers, grading the agency against outcomes, and the hiring or promotion calls. It is the role the one lead developer, two agencies, no CTO setup is missing. It is also not a builder. If you need code written, that is the agency or the team.
A full-time CTO sells presence. Someone in the building every day, in every meeting, carrying the context in one head. That is worth a great deal when there is enough work to fill the week: a product company, a team of 20 engineers, a platform that ships weekly. When there is not, the presence turns into something else. A full-time leader with a part-time problem builds a team to lead.
What happens when it goes wrong
The agency failure is slow and expensive. A pattern I have seen more than once: an owner signs a rebuild of a customer portal with a reputable agency, the first demo looks good, and 18 months later the project has cost twice the budget and does half of what the old system did. Nobody inside could have said in month 2 that the architecture was wrong, because nobody inside could read it. The agency was not lying. It was building what it was asked for, and nobody was asking the right questions.
The full-time failure is slow and personal. The common version at an SME is a CTO hired from a large company, who arrives with the habits of a 200-person engineering department and starts building one: a platform team, a hiring plan, a migration. For a company with 6 developers and 2 suppliers, that is a cost structure, not a strategy. It takes 6 to 12 months to see clearly, and the exit is a severance negotiation and a second search.
The fractional failure is quicker and cheaper, which is the point. A fractional CTO who is not delivering shows it within one or two quarterly reviews, because the work is written down: the roadmap, the decisions, the grading of the suppliers. You end it with a few weeks' notice. The risk is a different one: someone who stays on with no end state in sight. Ask on day 1 what ends the engagement.
Exit cost is the number nobody asks for
Owners compare the monthly cost of the three options and almost never the cost of leaving each one. Leaving an agency is cheap if you own the code, the cloud accounts, and the deployment pipeline, and a rebuild if you do not. Read the contract for who holds the keys before you need them. Leaving a full-time CTO costs a notice period or a severance, and then an empty seat that senior engineers read as uncertainty. Leaving a fractional CTO should cost almost nothing, because a good one builds the exit into the scope: decisions written down, a successor named, a handover agreed. If someone offering fractional work cannot describe their own exit, that is your answer.
Which one fits your situation
- Software runs the business, a supplier builds it, and nobody inside can grade the work. Fractional CTO, a few days a month. Keep the agency, put an owner over it, and rewrite the scope to pay for outcomes. This is the most common shape at an owner-led SME, and the one the SME practice is built for.
- One internal system quietly became load-bearing and nobody owns it. Fractional first, to decide whether it gets an owner, a supplier, or a replacement. The pattern is the SME that became a software company by accident.
- Software is the product, or the in-house team is heading past 15 to 20 engineers. Hire full-time. A fractional CTO can hold the seat and shape the role while the search runs, but this is a job that fills the week.
- A large build has to happen and the team inside is small. Agency for the capacity, under a fractional or internal owner who wrote the specification and grades the delivery. An agency without that owner is the expensive version of this table.
- You have a strong lead developer who wants more. Promote, and add a fractional CTO a few days a month as a counterweight and a coach. That combination often replaces a full-time hire for years.
For most SMEs the answer is a combination: an agency or a small team for the building, a fractional CTO for the owning, and a promoted internal lead for the daily decisions. Move to a full-time hire when the fractional days no longer fit in the week, not because the last agency project went badly. A bad agency project is an ownership problem, and ownership is a few days a month.
Frequently asked
Should an SME hire a CTO or use an agency?
Usually neither on its own. An agency sells capacity, not an independent view of its own work, so it cannot own your roadmap. A full-time CTO can own it, but most SMEs of 40 to 400 staff where software supports the business do not have 5 days of CTO work in a week. The combination that fits most often is an agency or a small team for the building and a fractional CTO a few days a month for the owning.
Is a fractional CTO cheaper than a full-time CTO?
Per day, no. Market day rates for a leadership-level fractional CTO in Belgium and the Netherlands sit around EUR 1 100 to EUR 1 600. Per year, usually yes, because you buy 2 to 4 days a month or 2 to 3 days a week instead of a full salary with employer costs, a bonus, and a search of 6 to 9 months. It also stops when the work is done, without a severance.
Can an agency act as my CTO?
Not without a conflict of interest. An agency's commercial interest is more scope, and a CTO's job includes saying no to scope, grading the supplier, and deciding what not to build. An agency can provide a technical lead for its own project. Ownership of the roadmap, and the judgement on whether the agency is delivering, have to sit on your side, with an internal lead, a fractional CTO, or both.
When should an SME move from a fractional to a full-time CTO?
When the fractional days no longer fit in the week. In practice: software becomes the product, the in-house team heads past 15 to 20 engineers, or the roadmap needs a leader in the building every day. A good fractional CTO tells you when that point arrives, writes the job description, and hands over to the permanent hire.
Written from patterns I've seen across engagements. If any of it sounds too specific to be general, that's why.
Related reading: one lead developer, two agencies, no CTO, the SME that became a software company by accident, and hiring a CTO versus bringing in a fractional one.