A fractional CTO who still writes code

Retainers are sized in days per month, agreed up front and adjusted as the engagement settles. Most owner-operator businesses find they need less than expected, because the value concentrates in availability at decision points rather than in a fixed number of hours logged. A typical shape is a regular check-in, review of anything significant in flight, and being reachable when a quote arrives or something breaks. If the real need turns out to be a full-time engineering leader, the honest recommendation is to hire one — and helping scope and fill that role is legitimate work for the engagement.

Both, and the combination is the point. An advisor who has not shipped software recently gives strategy that sounds reasonable and estimates that are wrong, because the real cost of a piece of work is not something you can reason about from memory. In practice some problems are faster to solve directly than to specify, delegate, and review — and where that is true, the work gets done. The rest of the time the role is genuinely advisory: reading proposals, reviewing architecture, and being the technical counterparty in decisions.

Yes, and it is one of the most common single pieces of work. A quote review covers whether the scope matches what you asked for, whether the estimate is plausible for that scope, what is conspicuously missing, and where the lock-in sits — hosting, proprietary frameworks, data portability, contract terms. This is available as a one-off engagement rather than requiring a retainer. The value comes precisely from having no stake in which vendor wins, which is the one thing the vendors quoting cannot offer you.

Yes — writing the role definition, screening applications, running the technical portion of interviews, and reviewing early work. Technical hiring is genuinely difficult without an engineer, and the common failure is not hiring someone bad but hiring the wrong shape: a specialist where you needed a generalist, or a senior engineer for work that will bore them within a quarter. Where it helps most is being honest about whether you need a hire at all, since a defined project is often better served by a fixed-scope engagement than by a permanent role.

That is largely what you are paying for. The recurring examples: the product should be sunset, the platform being proposed will lock you in, the developer is not doing good work, or the thing you want to build is not worth what it costs. SendMyLove is the public version of this — a project that shipped, earned nothing, and got sunset with the post-mortem published rather than quietly buried. An advisor whose incentive is to keep the engagement going will not reliably tell you to stop spending.

Mostly continuity and incentive. A consultant is typically engaged for a defined piece of work, delivers a recommendation, and leaves — which means they never see whether the advice was right. A fractional engagement is ongoing, so decisions get made by someone who knows what was decided six months ago and why, and who is still around when the consequences land. The other difference is having no product or platform to sell you, which removes the structural bias in advice from anyone whose revenue depends on which answer you choose.