How to Interview a Fractional CTO: 12 Questions That Separate Operators from Talkers

Hiring a fractional CTO is an odd interview problem. The candidate has sat through hundreds of sales calls. You have probably never assessed a CTO in your life. And the thing you are buying, which is judgement, is exactly the thing that is easiest to fake for an hour.

So most founders fall back on chemistry. They pick the person who explained things clearly and seemed confident. The trouble is that confidence is the one trait every bad fractional CTO has in abundance.

The questions below are designed to produce answers you can write down, compare across candidates, and check against references. Ask all twelve. Take notes. The differences between candidates will be obvious by question four.

Why the usual interview fails

You cannot give this candidate a coding test, because the work is months of decisions, not an afternoon of typing. Their case studies are anonymised, their references are curated, and every one of them has a polished answer to "what do you do?".

The fix is to ask questions whose good answers are specific and costly. An operator answers with stories that cost them something: money, a client, an argument they lost. A talker answers with frameworks. When you hear a framework where a story should be, that is your signal.

One more rule before you start: ask every candidate the same questions in the same order. You are not making conversation. You are collecting data.

Track record: questions 1 to 3

Start with history, because it is the hardest thing to invent on the spot.

1. "Tell me about a technical decision you got wrong. What did it cost, and what changed afterwards?" A good answer has a number in it: months lost, money burned, a customer who left. A bad answer is a humble-brag, like picking a technology that was "too advanced for the team at the time".
2. "Tell me about a client you turned away, or told they did not need you. Why?" An operator names the situation immediately, because saying no is half the job. If every prospect this person ever met turned out to need a retainer, you are talking to a salesperson.
3. "Describe a codebase you inherited that was worse than described. What did you do in the first 30 days?" Listen for sequence: audit first, quick wins second, big structural decisions last. The bad answer jumps straight to a rebuild, which is the most expensive sentence in software.

Capacity and availability: questions 4 to 6

Fractional means shared. You need to know exactly what fraction you are getting, and what happens when the sharing breaks down.

1. "How many clients do you work with right now, and what is your ceiling?" There is no single right number, but there is a wrong reaction: hesitation. Five clients at two days each is a full calendar with no slack. Ask where you fit in it.
2. "What happens when two of us have a crisis in the same week?" A good answer describes real triage: who gets called, what gets rescheduled, who covers. "That has never happened" is a bad answer. It means it has not happened yet.
3. "What response time between sessions will you commit to in the contract?" You want a number: same-day acknowledgement, an answer within 24 hours, whatever it is. "I am always on Slack" sounds generous and commits to nothing.

Ways of working: questions 7 to 9

These three tell you what the engagement will feel like in month four, after the honeymoon.

1. "Show me a board update or technical audit you have written. Redacted is fine." One real document tells you more than an hour of conversation. If the thinking only ever existed in calls, you are renting a conversation, not leadership.
2. "How do you run hiring for us without staying the bottleneck?" The good answer: they design the interview loop, train your people to run it, and step back. The bad answer keeps them in every interview forever, because a bottleneck is billable.
3. "Tell me about disagreeing with a founder who wanted to build something you thought was wrong. What happened?" You want evidence they pushed back in writing, and evidence they lost some of those arguments without sulking. A candidate who has never lost an argument with a founder has never really had one.

Fit and the exit: questions 10 to 12

The last group checks whether this person suits your stage, and whether they have thought about how the engagement ends. Both matter more than the stack.

1. "Have you worked with a company at our stage and stack, or one stage ahead of us?" One stage ahead is the experience you actually want, because it means they have seen your next 18 months of problems. An enterprise-only background translates badly to a five-person team.
2. "What would make you recommend we hire a full-time CTO?" The honest answer describes their own exit: a team size, a funding stage, a point where two days a month stops being enough. Someone who cannot describe the end of the engagement is planning for it not to end.
3. "When the engagement winds down, what do we keep?" You want assets that outlive the retainer: architecture documentation, a hiring plan, an org design, written standards. "We will figure that out later" means you keep nothing.

Check references like it matters

Ask for references from founders at your stage. Not enterprise IT directors, not fellow consultants, not the CEO of a 400-person company where the candidate was one voice in a steering committee. Founders who ran a 6-person startup while this person was their fractional CTO.

Then ask the referees questions with teeth:

• What did they push back on, and were they right?
• What do you still use today that they built, wrote, or set up?
• Why did the engagement end, and would you restart it tomorrow?

A glowing reference that cannot answer the second question is describing a pleasant relationship, not a useful one.

Red flags that end the conversation

Some answers should stop the process on the spot, however good the chemistry:

• They recommend a rebuild in the first meeting, before reading a line of your code.
• They cannot name a single client they told "you do not need me".
• Everything is a retainer. There is no audit, no strategy day, no smaller way to test the fit first.
• They have no written deliverables to show you, even redacted.
• They ask for a 12-month lock-in. The whole point of the model is that it flexes with your stage.
• They talk about technology for an hour and never once ask about revenue, runway, or customers.

How to run the process

Interview two or three providers. Ask the same twelve questions in the same order and write the answers down. When you compare notes, score specificity, not polish: the candidate with the messier delivery and the more expensive stories is usually the operator.

If two candidates survive, buy a small piece of work before you commit to a retainer. A one-day strategy session or a paid audit shows you their written output, their questions, and how your team reacts to them. It is the cheapest hiring mistake insurance available.