The three arrangements get compared as though they are competing bids for the same job. They are not. They solve different problems, and most of the regret we see comes from a company buying capacity when it needed judgment, or hiring a permanent executive to answer a question that a good advisor could have settled in a month.
What each one actually is
| Fractional CTO | Full-time CTO | Agency | |
|---|---|---|---|
| You are buying | Senior judgment, on a schedule | Ownership and leadership capacity | Execution capacity |
| Typical cost | $8,000 to $15,000 a month | $180,000 to $320,000 plus equity | $15,000 to $60,000 a month while engaged |
| Time commitment | One to two days a week | Full time | Project-shaped, then gone |
| Good at | Architecture calls, hiring, vendor decisions, delivery discipline | Building and running an engineering organisation | Shipping a defined scope |
| Bad at | Volume of hands-on work | Being affordable before you need one | Telling you the project is wrong |
| Reversibility | High, typically 30 days | Low, a bad hire costs a year | High, contract ends |
When a fractional CTO is the right answer
The fractional arrangement fits a specific and very common stage: your decisions have outgrown your in-house depth, but your headcount has not yet outgrown your ability to manage it informally. You need someone qualified in the room a few times a week, not someone at a desk every day.
The signals are recognisable.
- A non-technical founder or CEO is making architecture and vendor calls alone, and knows it.
- You are about to hire engineers and nobody internally can assess a candidate's technical depth.
- A platform contract or a build-versus-buy decision is in front of you that would be painful to unwind.
- Delivery predictability has degraded and nobody can explain why.
- An investor or acquirer is asking questions about technical risk that you cannot currently answer.
In each of those, the scarce resource is judgment applied at the moment a decision is made. Adding engineers does not help. Hiring an executive is disproportionate. A senior person present at the decision points is exactly the shape of the need.
When you should hire a full-time CTO
Two conditions make the permanent hire correct, and either one on its own is usually enough.
The first is that technology is the product. If what you sell is software, the person accountable for it needs to be in every commercial conversation, own the roadmap, and be measured on the same outcomes as the rest of the leadership team. Part-time attention will not survive that.
The second is organisational scale. Somewhere around eight to ten engineers, running the engineering organisation becomes a full-time job in its own right: hiring, performance, career paths, on-call, the cadence of delivery. A fractional CTO can design that system. They cannot run it day to day, and pretending otherwise produces a team that feels leaderless four days out of five.
The cost is not only salary. A CTO search typically runs three to six months, carries recruiter fees around 20 to 25 percent of first-year compensation, and a mis-hire at that level costs the better part of a year in lost momentum. That is a strong argument for having a qualified person help you write the role and run the technical interviews, which is itself a common fractional engagement.
When an agency is the right call
Agencies are the correct answer to a specific question: the decisions are made, the scope is defined, and you need more hands than you have. Building a defined product surface, replacing a legacy system whose replacement has been specified, delivering a project your own team has no capacity for. That is honest, valuable work and a good agency does it well.
The failure mode is structural rather than moral. An agency's business model is capacity, which means it is not well positioned to tell you the project should not happen. If nobody on your side can evaluate the architecture being proposed, assess whether the estimate is realistic, or push back on a technology choice that suits the vendor better than it suits you, then you have bought execution with no oversight attached.
The question is not whether the agency is good. It is whether anyone on your side of the table can tell.
This is why the agency and the fractional CTO are often complements rather than alternatives: the fractional seat provides the oversight that makes the agency relationship safe.
The sequence most companies should follow
- Start with a bounded diagnostic. A few weeks of senior attention to establish what the actual constraint is. This is cheap and it frequently changes what you buy next.
- Take a fractional arrangement for the decision-heavy period: architecture, vendor selection, the first serious hires, delivery discipline.
- Use agency or contract capacity for defined build scope, with the fractional seat reviewing architecture and estimates.
- Hire the permanent CTO when technology becomes the product or the team passes roughly eight engineers, using the role specification the fractional engagement produced.
Companies that skip step one usually buy capacity for a problem that was not a capacity problem. Companies that skip step four end up with a senior advisor holding a role that has quietly become full-time, at which point everyone is being underserved.
Questions to ask before you commit
- What decision am I trying to get right, and does it need judgment, leadership, or hands?
- If this arrangement ended in ninety days, what would break, and is that acceptable?
- Who on my side can evaluate the technical quality of what I am buying?
- What is the written plan for not needing this arrangement any more?
- Am I solving a problem, or am I buying reassurance about one?