What to ask before signing a fractional CTO contract.

Eleven questions a founder should put to a fractional CTO operator before signing the retainer. Asked in the last 15 minutes of the discovery call, they cut a wrong-fit retainer in 30 minutes instead of 3 months. Each question lists what a real answer sounds like and what a hand-wave answer sounds like.

Most founders treat the fractional CTO discovery call as a sales call where they are the buyer being sold to. That framing produces wrong-fit retainers because the founder asks soft questions ("what's your background", "what kinds of companies have you worked with") that any operator can answer well regardless of fit.

The framing that produces right-fit retainers is the opposite. The founder asks operational questions that are hard to bullshit. The operator's answers tell you whether they have actually held the seat before. Eleven questions, in roughly the order they should come up.

The questions that actually surface fit.

I am going to give you the two questions that decide it. Then four more that close the loop. Then a sentence each on the boring stuff.

Question one: what is NOT in the retainer.

If you only have time to ask one question on the discovery call, ask this one.

Real fractional CTOs have a clear, short, slightly tired list of things they will not do, because they have been asked to do all of them and learned that the engagement breaks when they say yes. Mine: line management of engineers, daily standups, and IC code beyond evaluation spikes. Those are not the role I sell. Different shapes, different price points, and I refer them to people who do that work.

The hand-waver answers this question with "we figure it out as we go" or "I am flexible, whatever you need." Both versions are saying the same thing, which is that the operator has not held the seat long enough to know where it ends. Walk on this one alone if the answer is vague. Everything else on the list amplifies what this question already told you.

Question two: tell me about a retainer that did NOT work out.

The second question that actually decides it.

Every fractional CTO with two or more years of experience has had a retainer end badly. If they cannot describe one in specific terms, with what they learned and what they changed, they are either lying or they have not done this enough times. I have one I tell when asked: a retainer that ended early because the founder used the weekly 1:1 as a status meeting rather than a thinking conversation, and the work that needed senior judgment never came to the call. What I changed: the four-week pilot now explicitly tests whether the founder will use the seat as a partner, and either side can close cleanly at week 4 if not.

The shape of the answer matters more than the content. An operator who has reflected on a failure has the kind of self-correction that compounds across engagements. An operator who says "all my retainers have gone great" has not been asked the question often enough.

Four more that close the loop.

How many other retainers, capacity question. The honest answer is a number, usually two, with the seat counted. "Depends on the client" or "I am flexible" is the same problem as question one wearing a different label.

References. The honest answer is yes, two of them, one current and one past, with an offer to email them today. The honest qualifier is "they will speak to working style; engagement specifics are under NDA." The dishonest answer is "most of my clients prefer not to be contacted." Those clients exist; one of them will say yes for a serious buyer.

Continuity if the operator is sick or unavailable. Already covered in the India hiring essay; the answer here is the same. Named peers who can pick up on one week notice, decisions and architectural state kept in writing throughout, and the bus-factor plan in the SOW. Anything less than that is "best efforts" with a different label.

What does the price not cover, and what triggers an uplift. Retainers price the steady-state operating rhythm. Specific events do justifiably trigger uplifts: investor technical diligence, M&A technical due diligence, a regulator audit, a major architectural pivot. The operator should be able to list those events from memory and have the uplift rate written into the SOW. "It is a flat fee" or "I would never nickel-and-dime you" is the answer of someone who will end up either resenting the engagement or quietly under-serving.

The boring ones, one sentence each.

Cadence: a real operator can describe a typical week in specific time chunks (weekly 1:1, async response window in business hours, monthly written architecture review). A hand-waver says "as needed."

Hiring support: the answer covers JD draft, technical screen design, panel interview seat, offer calibration, reference checks, with rough hours per senior hire.

Pilot: four-week paid pilot is what I run; either side walks clean at the end. Some operators run two-week or six-week pilots; whatever the length, the structure should be the same.

Off-ramp: a real operator describes how the engagement ends, and the description is concrete (transition to a full-time hire, recruit help, written handover plan in the last six to eight weeks).

IP and exclusivity: work-for-hire on deliverables, no broad IP assignment if the operator is not writing production code, industry non-compete around a specific competitor for a fee uplift, no general exclusivity. Operators who promise exclusivity to every client are lying to one of you.

How to use the list.

Spend the last 15 to 20 minutes of the discovery call on these. Lead with question one, then question two; if the answers there are vague, the rest does not save the call. If those two land, run the four close-the-loop questions, and then a fast pass through the boring ones.

The operator's reaction to the framing is itself a signal. A real operator says "good, that is how I would do it too." A hand-waver pushes back or tries to redirect to general chemistry.

The asymmetry of the cost is severe. A wrong-fit retainer caught in the discovery call costs you sixty minutes. A wrong-fit retainer caught in month four costs you three months of retainer plus the opportunity cost of not having had the right operator in the seat. Slow down on the discovery call.