How to hire a fractional CMO
Most bad fractional CMO engagements were decided before the search began, by a company that had not settled what the role was for. The evaluation gets the blame, but the failure happened earlier.
What follows is the sequence that avoids that: three things to settle internally, where candidates actually come from, how to evaluate one in a working session rather than an interview, what belongs in the agreement, and the 30-60-90 scorecard that tells you early whether it is working. If you are still deciding whether you need this role at all, start with what a fractional CMO is.
Published August 21, 2026· 9-minute read
Settle three things before you talk to anyone
Every one of these is answerable in a single leadership meeting, and each one that goes unanswered becomes a month of the engagement.
- 1. Which number are they accountable for?
- Not "growth." A specific figure with a baseline and a date. If leadership cannot agree on it in a room without the candidate present, the engagement will spend its first two months resolving that argument on your budget.
- 2. Who will they direct?
- Name the people. A fractional CMO with nobody to direct becomes an expensive coordinator, and you will feel it by week six. If the honest answer is nobody, hire a senior marketing manager first and revisit this in a year.
- 3. What can they decide without asking?
- Budget reallocation up to some threshold, firing an agency, killing a channel, changing the brief. Write the threshold down. Executives who must escalate every decision are being paid executive rates to produce recommendations, which is the consultant arrangement you were trying to avoid.
Where candidates actually come from
Your own network, and your investors
Highest hit rate by a wide margin, because the referrer has watched the person work. Ask specifically for someone who has operated in your model — multi-location, DTC, B2B — rather than someone who is broadly impressive.
Fractional marketplaces and talent networks
Fast and well-vetted for credentials. Weaker on fit, because the matching is done on résumé keywords rather than on the shape of your problem. Useful for building a comparison set.
Operators who write publicly about your problem
Slower, and the highest signal available. Someone who has published a detailed view of the exact problem you have is demonstrating judgment before you have paid for any. Read what they publish and judge whether it is specific enough to act on.
Agencies offering a fractional CMO product
Verify who actually shows up and how much of their week you get. This model sometimes places a junior operator against a senior brochure, and sometimes routes your budget toward the agency own services. Ask both questions directly.
Evaluate in a working session, not an interview
Interviews measure how well someone interviews. For an executive you will work beside every week, that is close to irrelevant. Run a working session instead.
- Send real data in advance — the last two quarters of performance, the current budget, the org chart. Under an NDA if you prefer. A candidate who will not read it before the session will not read anything after it either.
- Ask them to come with what they would kill and what they would fund in the first sixty days, and the evidence they would want before doing either.
- Put a person from your team who is skeptical in the room. Watch how the candidate handles being pushed on their reasoning. That is the interaction you will have every week for a year.
- Ask them what they would need from you and what would make the engagement fail. Candidates who name your organisation as a risk factor are describing reality, not being difficult.
- Pay for the session if it runs beyond an hour. You will get materially better preparation, and the candidates who decline paid work at that stage are telling you something useful.
What belongs in the agreement
- Days per week, named
- One to two days is typical, three when a build is underway. Vague "as needed" arrangements always compress toward the lower bound at the moment the engagement is busiest.
- A minimum term of at least three months, more often six
- The first month is diagnostic and nothing compounds before roughly ninety days. Shorter terms reliably produce a strategy document and no change in the number.
- Written decision rights
- The thresholds you settled before the search. Put them in the agreement, not in a conversation.
- An artifact-ownership clause
- Strategy documents, vendor contracts, measurement setup, playbooks, prompts and configurations if AI systems are involved, and any accumulated training data. All of it should be yours during the engagement, not handed over at the end. If a counterparty resists this, you have learned what their retention strategy is.
- A named exit
- What happens when you hire a permanent CMO, and what transfers. A good operator has a rehearsed answer, because making themselves unnecessary is the job.
The 30-60-90 scorecard
Agree this with the operator before day one. Both sides benefit from the same definition of working.
Day 30
What good looks like: They can describe your business back to you more precisely than your own team does, and they have named something everyone knew was broken but nobody had said out loud. At least one thing has been stopped.
Warning sign: They are still "getting up to speed" and have produced a deck.
Day 60
What good looks like: Something has been killed and something has been funded, both with a stated reason. Your team knows what they are supposed to do differently. Measurement that did not exist now exists.
Warning sign: Activity has increased and nothing has stopped. This is the most common failure and the hardest to see, because it looks like progress.
Day 90
What good looks like: The number has moved, or you can see precisely why it has not and what changes next. Your team is better at their jobs. You could explain the strategy to your board in three sentences without notes.
Warning sign: You cannot tell what is different. At this point the honest move is to end it rather than hope month four is different.
Red flags
- They agree with everything you say in the first meeting. You are paying for judgment, and judgment disagrees sometimes.
- They cannot describe an engagement that failed, or blame the client entirely for the one they describe.
- They will not tell you how many other engagements they are running.
- Their proposal is a list of deliverables with no stated number attached.
- They want to bring in "their" agency before understanding what you already have. Ask whether they receive anything for that referral.
- They resist the artifact-ownership clause.
The last one is worth insisting on regardless of who you hire. Whoever runs this role accumulates the assets that make the next decision cheaper — the measurement setup, the playbooks, the record of what was tried. If those leave when the operator leaves, you have rented progress rather than bought it.
If we are one of the candidates
Run the checklist above on us. The deliverables, the timeline, and the ownership terms are already published, so you can hold them against anyone else you are considering before a conversation happens.
Ready to talk instead? Book the 30-minute consultation, or take the three-question diagnostic first. No email required for the diagnostic.
Budgeting first? What a fractional CMO costs covers market bands, the five factors behind them, and the total-cost comparison against a full-time hire.
The assessment is the smaller first step if you would rather test the working relationship before committing to a term.