Interviewing a fractional CMO is different from interviewing a full-time hire. You are assessing someone who will make senior decisions in your business for part of the week, often remotely, while working with other companies. Credentials and past logos tell you less than how the candidate thinks about your specific situation. The questions below are grouped into five areas: diagnosis, measurement, leadership, working model and exit. For each group, the article describes what a strong answer usually includes and what a weak one sounds like. Use a selection of them in a first conversation and the rest in a second, ideally with your sales leader in the room for the measurement questions.
Questions about diagnosis
- Based on what you know so far, where do you think our main constraint might be, and what would you need to see to confirm it? 2. What would you look at in your first two weeks, and in what order? 3. What would make you recommend that we not hire a fractional CMO at all? 4. Tell me about a company where your first diagnosis turned out to be wrong. What did you change? Strong answers form a hypothesis but hold it loosely, ask for specific data such as spend by channel, lead-to-opportunity conversion and sales cycle, and are willing to say the engagement might not be the right answer. Weak answers present a full strategy before seeing any data, or describe the same first steps they use with every client regardless of business model.
Questions about measurement
- How would you measure marketing's contribution to pipeline and revenue in a company like ours? 6. How do you handle disagreements with sales about lead quality? 7. What would you put in the monthly report to me, and what would you leave out? 8. How do you decide whether a channel is working when attribution is imperfect? Strong answers distinguish sourced from influenced pipeline, talk about agreeing lead definitions with sales, mention checks such as holdout tests or comparing trends when attribution is unreliable, and describe a report built around pipeline, cost per qualified opportunity and decisions taken. Weak answers centre on impressions, clicks, followers or a single attribution model presented as the truth.
Questions about leadership and agencies
- How do you lead a team when you are only here part of the week? 10. How do you decide whether to keep, change or replace an agency? 11. Tell me about a time you had to stop a project or channel that someone senior liked. How did you handle it? 12. How would you work with our sales leader? 13. When would you recommend we hire, and who first? Strong answers describe a clear weekly rhythm, written briefs, decision rights and asynchronous communication; judge agencies on pipeline and ownership of accounts rather than activity; and show they can disagree with a founder respectfully using evidence. Weak answers avoid the conflict question or suggest they would simply do more of the work themselves.
Questions about the working model
- How many clients do you work with at the moment, and how would you protect time for us? 15. What time commitment would you recommend for us, and why? 16. Which meetings would you attend, and which would you replace with written updates? 17. What do you need from me to be effective, and what would slow you down? Strong answers are open about their client load, recommend a commitment based on your situation rather than their standard package, and list specific needs such as access in week one, decision rights and the sales leader's time. Weak answers promise unlimited availability or cannot explain how they split time between clients.
Questions about exit and handover
- What will exist in our company when you leave? 19. How do you make sure we are not dependent on you? 20. Who owns the work, data and accounts during and after the engagement? Strong answers describe documentation of systems and decisions, a trained in-house owner, and all accounts sitting under company ownership from the start. Weak answers are vague about handover or assume their own tools or accounts will hold your data. The exit playbook article covers what a clean handover includes.
Scoring sheet
Score each area from 1 to 5 after the interview, and compare candidates on the same sheet.
| Area | Questions | Strong answer signals | Weak answer signals |
|---|---|---|---|
| Diagnosis | 1 to 4 | Hypothesis plus data request, willing to say no | Generic plan before seeing data |
| Measurement | 5 to 8 | Sourced vs influenced pipeline, sales-agreed definitions | Vanity metrics, one model as truth |
| Leadership | 9 to 13 | Weekly rhythm, evidence-based disagreement | Avoids conflict, does the work themselves |
| Working model | 14 to 17 | Honest client load, tailored commitment | Unlimited availability promises |
| Exit | 18 to 20 | Documentation, trained owner, company-owned accounts | Vague handover, own tools hold your data |
Reference checks
References matter more for fractional roles than for many full-time ones, because you are relying on someone's judgement with limited oversight. Ask for two or three references from companies at a similar stage and business model. Ask each reference what the fractional CMO changed in the first three months, how they handled a disagreement, whether the time commitment matched the contract, and what was left behind at the end. Ask whether they would hire the person again, and for what. Specific, consistent answers across references are a good sign.
A short paid task, if you want one
Some companies ask finalists for a short, paid piece of work: for example, a review of one funnel stage or the current marketing report, with recommendations. This shows how the candidate works with your real data and how they communicate findings. Keep it small and pay for it, because unpaid strategy work tends to attract generic answers and discourages experienced candidates. Judge the output on specificity and prioritisation, not on how polished the slides are.
Running the interview process
A simple process works well. Start with a first call of around an hour, covering diagnosis and the working model, with the founder or CEO. Shortlist two or three candidates for a second conversation that includes the sales leader, focused on measurement and leadership. Take up references for the finalists. If you use a paid task, set it after the second conversation and review it together in a short call. Decide within a few days of the last step, because strong fractional candidates usually have limited capacity and may commit elsewhere. Throughout, share enough information for candidates to give specific answers: rough spend, lead volume, sales cycle and the main problem as you see it. Withholding everything until after signing makes every candidate sound generic.
Mistakes to avoid when interviewing
A few mistakes recur. Choosing on presentation polish rather than on the quality of thinking about your business. Asking only about past results, which are hard to verify and depend on context, instead of how the candidate would approach your situation. Excluding the sales leader, who will work closely with the fractional CMO and can test measurement answers better than anyone. Treating a confident promise of specific growth as a strength rather than a warning sign. And skipping references because the candidate seems impressive. A structured set of questions and a scoring sheet, used the same way for every candidate, keeps the comparison fair.
FAQ
Ask how they would diagnose your business in the first two weeks and what data they would need. The answer shows whether they think specifically or rely on a standard plan.
Yes, at least for the measurement and leadership questions. The fractional CMO will work closely with sales, and the sales leader can test how they handle lead quality disagreements.
A short view of likely constraints is reasonable. A full strategy should be paid, because unpaid work tends to be generic and discourages experienced candidates.
Two or three from companies at a similar stage and business model. Ask what changed in the first three months and what was left behind at the end.
A guaranteed revenue result before any diagnosis, a full strategy before seeing data, vagueness about client load, or an expectation that your data will live in their accounts.
Yes, ideally two or three. Comparing answers to the same questions on a shared scoring sheet makes differences in thinking much easier to see than a single conversation.