The honest pitch for a fractional CTO is narrower than the way it is usually sold. It is not “senior leadership without the cost”. It is a specific answer to a specific constraint, and the constraint expires.
The problem it actually solves
Early on, the binding constraint is rarely how much engineering gets done. It is how many expensive, hard-to-reverse decisions get made badly.
Which parts to build and which to buy. What the data model has to survive when the second market opens. Whether the compliance requirement that appears in every enterprise deal is a two-week problem or a two-quarter one. What “good” looks like in an engineering interview when nobody on the founding team has run one.
Those decisions do not need someone present five days a week. They need someone senior enough to be right about them, present often enough to have the context, and accountable enough that they are not just opinions. That is the fractional shape, and it is why it works as a monthly retainer rather than ad hoc hours. Someone who shows up when summoned does not have the context to be accountable for anything.
When it is too early
If there is no product and no users, the constraint is not technical judgement. It is finding out whether anyone wants the thing. A founder who can build a rough version themselves should build a rough version themselves, and buy an hour of advice when a specific decision comes up.
That is what an office-hours tier is for, and it is the honest recommendation at that stage even though it is the smaller sale.
When it is time to hire
Three signals, and one is usually enough:
Throughput has become the constraint. Once there are four or five engineers, the bottleneck stops being decision quality and becomes coordination, review, and unblocking: daily work that a defined two days a week cannot absorb.
The technology is the differentiator. If the hard part of the company is the technical part, that judgement has to live inside the company permanently. Renting it is fine while you are proving the market; it is not fine once the moat is the thing being rented.
The team needs someone present. Engineers do not just need architecture; they need someone whose job is their careers. A fractional CTO cannot do that well and should not pretend to.
Planning the exit rather than discovering it
The awkward version of this conversation is the one that happens by accident, usually when a board member asks why there is no CTO. The better version is agreed at the start: the fractional engagement includes writing the job spec, running the loop, and handing over to the person who replaces us.
That should not be a difficult thing for a supplier to say. If a fractional CTO is reluctant to define the conditions under which they become unnecessary, that is worth noticing.