How to Evaluate an Interim CMO Before You Hire

The short answer: judge an interim CMO on evidence you can check, not on logos or on how well they present. Write the brief around a decision first. Then verify every result a candidate claims, ask each reference the same questions, give your last two or three candidates a paid planning exercise, and score everyone on one sheet.

Disclosure: I sell interim and fractional CMO work, so I am not a neutral party here. I have 15+ years in marketing across 50+ organizations, and I compete for the kind of role this post is about. Run every check below on me too.

A weak choice costs budget, time and the team's trust. An interim engagement is short by design, so there is little room to recover from one. Checks 1 and 2 cost only time and should cut the list down, so that just two or three candidates reach the paid exercise in check 3. For each check I say what to ask, what a good answer looks like and where the check falls short.

Step zero: write the brief around a decision

Do this before you meet anyone. A candidate can only be judged against something you have written down. Put three things on one page:

• Decision rights. What the interim may change without asking you: budget moves, hires, agency and tool contracts.
• The exit condition. What must exist when they leave, as a list of things (a plan, a dashboard, a hired head of growth, a handover document), not as a feeling.
• Who inherits the work. A named person or role. If nobody is named, the CEO becomes the default owner of everything the interim built.

The recruitment page shows a weak and a usable brief side by side, and the six interview questions I would ask first: how to write the brief and what to ask.

Check 1: verify the claimed results

Ask each candidate for two or three pieces of work they would stand behind. For every result they quote, get five facts:

1. The measure. Which number, in which system.
2. The baseline. What it was before.
3. The period. How long the change took.
4. Their role. Led it, contributed or advised, and who else was involved.
5. A person who can confirm it.

"Increased leads by a third" fails on all five until someone supplies them. Ask what else changed in the same period (budget, pricing, a new sales team) and what happened further down the funnel. More leads with the same pipeline is not a result.

Ask for one project that went wrong. Listen for what they would do differently, not for how polished the story is. A logo on a CV shows who the client was, not what the candidate did there. Agency work, a workshop and a full mandate can all sit under the same name.

Check 2: reference calls, same questions for everyone

Take two references per finalist. One comes from the most recent engagement. The other the candidate did not offer first: ask the first reference who else you should speak to. Make the calls yourself, and ask every candidate's references the same questions so the answers can be compared:

• What was the mandate, and what existed at the end that did not exist at the start?
• How did they handle a budget cut or a sudden market shift?
• Describe a time they had to change the mind of a senior executive. What was the context, the approach and the outcome?
• Who did they hire or develop, and are those people still there?
• What did they leave unfinished, and why?
• Would you give them the same mandate again, and what would you change in the brief?

Reference calls are the main check that does not depend on the candidate's own account. The candidate still chose the references, so a warm but vague answer is weak evidence and a specific one is strong.

Check 3: the paid planning exercise

Ask your final two or three candidates for a written 30/60/90 day plan for your situation, and pay them for it. After verified results, this is the step I would weight most, because it is a sample of the real work. That is my opinion. I would be tested by it too, so apply it as strictly to me as to anyone.

What to give them. A two-page written brief (the decision, the decision rights, the exit condition), a summary of the last 12 months of marketing numbers, the org chart, the budget range and the three questions you most want answered. Every candidate gets the same pack and the same deadline.

What to ask for back, in no more than five pages:

• A diagnosis: what they think is wrong, what evidence supports it and what evidence they still need.
• A 30/60/90 day plan with outputs: what exists at day 30, 60 and 90, and who signs each off. What a strong plan holds.
• What they would stop doing.
• The first three decisions they need from you.
• Their assumptions, and the questions they would ask in week one.

How to read it. Prefer the plan that names its assumptions and its outputs over the one that sounds most confident. A plan written from a two-page brief should contain questions. A plan with none is a template.

Pay for it. Agree a fixed fee in writing before you send the brief, cap the time you ask for, and pay every finalist the same. I recommend a cap of two days of the candidate's time, with the fee set from their own daily rate: at a rate of R per day, a two-day cap costs 2 x R.

Know its limits. The exercise measures planning from a short brief, not execution. It favors candidates who write well and have practiced the format, and it asks finalists to do real work before they have a contract. Keep it to two or three people, weigh it against the references, and skip it when the mandate is small or when references and a structured conversation already tell you enough. It is not a trial engagement, which puts someone inside your company before you have chosen them. It is a few days of written work done before you choose. The 30-day exit clause I recommend for the engagement itself still applies.

Why a work sample. Selection research supports testing candidates on a sample of the real job. In a 2022 re-analysis of decades of studies (Sackett, Zhang, Berry and Lievens, restated by the same authors in a 2023 article), structured interviews, where every candidate gets the same questions scored the same way, had the highest average validity (.42). Work sample tests came in at .33. Validity here is the correlation between an assessment score and later job performance: 1.0 would be perfect prediction and 0 would be none. The authors caution that the .42 should be read as ".42, plus or minus .24", because it varies widely between studies. These are averages across many kinds of job, and none of them tests a planning exercise for an interim marketing executive. I use them to support two practices, not to prove this one: sample the real work, and ask every candidate the same questions.

Check 4: working fit with the team and the executives

An interim CMO succeeds or fails on how fast the team and the other executives start working with them. Test that directly.

• Give two or three people from the marketing team and one peer executive (often the head of sales or the CFO) 30 minutes each with the candidate. Ask each of them to answer the same three questions in writing before they talk to each other. Would you bring a problem to this person? Did they ask about your work before they talked about theirs? What would you want them to change in how they work with you?
• Give the candidate a realistic conflict. The head of sales says the leads are poor, marketing says the follow-up is slow. What do they do in the first week?
• Ask how they will treat the current team's work in the first month: keep, change or stop, and how they decide.

Define fit as behavior you can name: how they disagree, how they share credit, how they treat existing staff. Someone you would enjoy a drink with is not a criterion.

Check 5: technical depth in your channels and stack

An interim CMO does not need to build anything. They do need to read your numbers and your tools. Ask them to:

• Review a one-page summary of your stack and say what they would keep, change and stop, and what they would need to see before deciding.
• Describe the dashboard they would want after 60 days: five to eight numbers, who owns each and what decision each one informs.
• Explain how they would tell whether marketing is working when attribution is imperfect, and what they would not conclude from the data.
• Say what they have in production with AI or automation, who owns it and what happens when it fails. A pilot slide is not a workflow that runs.
• If the brief includes a new product or venture, explain how they test demand before building, for example with a smoke test.

Ask for specifics: a tool, a number, a decision. Fluent general answers are easy to give and hard to check.

Score candidates on one sheet

Set the weights before you meet anyone, write them down and do not change them afterwards. Score each criterion from 1 (no evidence) to 5 (strong, checked evidence), multiply each score by its weight, add them up and divide by 100.

The weights are an example. Here is an illustration with made-up scores, in the order of the table. Candidate A scores 4, 3, 4, 5, 3 and 3. Candidate B scores 5, 4, 2, 3, 4 and 2.

• A: (25 x 4 + 20 x 3 + 15 x 4 + 15 x 5 + 15 x 3 + 10 x 3) / 100 = (100 + 60 + 60 + 75 + 45 + 30) / 100 = 3.70
• B: (25 x 5 + 20 x 4 + 15 x 2 + 15 x 3 + 15 x 4 + 10 x 2) / 100 = (125 + 80 + 30 + 45 + 60 + 20) / 100 = 3.60

A gap this narrow tells you to look at the weakest scores, not the totals. B's 2 on outputs and exit plan matters more than the 0.10 difference, because the brief depends on the exit. I would also set a floor in advance: a score of 2 or below on verified results or on the exit plan ends the process, whatever the total.

Frequently Asked Questions

What is the single most important step when evaluating an interim CMO?

No single step is enough, and I have no evidence that one beats the others for this role. My own weighting, which is an opinion, puts most on results that a reference has confirmed and on the written plan from the paid exercise. In the research cited above, structured interviews had the highest average validity and work sample tests were in the top five, so use several methods and apply each one the same way to every candidate.

Should I pay candidates for a 90-day plan, and how much?

Yes, for the final two or three. This is my recommendation, not a rule. Set the fee from the candidate's own daily rate times the days you cap the exercise at, agree it in writing before you send the brief, and pay every finalist the same. A short written exercise is different from a trial engagement, which I would not use.

Can I rely on my network to find an interim CMO?

Use your network to build the candidate list, not to skip the process. A recommendation tells you someone was liked. It does not tell you what they changed. Put a referred candidate through the same results check, references, exercise and scoring sheet as everyone else.

How is evaluating an interim CMO different from evaluating a permanent one?

The time horizon. An interim has months, not years, so weigh how fast they can diagnose and decide, and weigh their exit plan as heavily as their entry plan. Ask what their handover document contains and who they have handed work to before. The 30/60/90 post lists what a handover document holds.

What are the biggest red flags?

Results with no baseline, no period or no named role. A reference who is vague, or a candidate who cannot supply one. A plan that would fit any company, with no assumptions and no questions. No curiosity about your business, your numbers or your team. Anything the candidate says that a reference contradicts. And a candidate who describes how an engagement starts but not how it ends.

References

[1] Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040-2068. https://doi.org/10.1037/apl0000994

[2] Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2023). Revisiting the design of selection systems in light of new findings regarding the validity of widely used predictors. Industrial and Organizational Psychology, 16(3), 283-300. https://doi.org/10.1017/iop.2023.24