The board wants a CMO who understands data. The candidate wants a role that gives them authority over the commercial system. The job description promises both. Neither the board nor the candidate is being honest about the structural gap between what the role offers and what the business needs.
Data fluency in a CMO candidate is assessed in the interview through questions about attribution models, incrementality testing, and marketing mix modelling. The candidate who can speak confidently about those methodologies gets the job. The candidate who can actually read the commercial system those methodologies are supposed to describe is a different and rarer person and the interview process is not designed to distinguish between them.
Speaking about attribution models and reading what an attribution model is actually producing from the data underneath it are not the same skill. The first is vocabulary. The second is diagnosis. Your interview panel can assess vocabulary. Diagnosis is only visible in the work.
The CMO who arrives with impressive data vocabulary and finds a measurement infrastructure that has been producing confident narrative rather than commercial intelligence is not the person who built the infrastructure. They are the person inheriting it. And the distance between their vocabulary and the reality they have inherited is where the first eighteen months of the role disappear.
Related reading