Contents
- Why unstructured interviews fail at marketing roles
- Four groups of criteria worth measuring
- A 1–5 scale that different interviewers read the same way
- Three test formats that produce real signal
- Scoring and making the decision
- Biases that distort marketing interviews
- Designing the interview loop
- Applying the same framework to people already on the team
- Frequently asked questions about assessing marketers
Hiring decisions in marketing go wrong for a predictable reason: the interview measures how well someone talks about work rather than how well they do it. A structured approach to marketer competency assessment replaces impressions with signals you can compare across candidates.
This guide sets out a four-group criteria framework, a 1–5 scale with observable anchors, three test formats that produce high-signal evidence, how to score without letting one strong answer dominate, and the biases that most often distort a marketing interview.
Why unstructured interviews fail at marketing roles
Marketing attracts people who present well — that is part of the job. An unstructured conversation therefore selects for the skill of describing work, which correlates far less with performance than most hiring managers assume.
- No comparison basis. If each candidate is asked different questions, ranking them is guesswork dressed up as judgement.
- Credit is unverified. “I grew the channel 3x” may describe a team of eight, a market tailwind, or genuine individual work.
- Fluency reads as competence. The candidate who knows the vocabulary sounds stronger than the one who knows the work.
Four groups of criteria worth measuring
Most marketing roles can be assessed on four groups. Define them before you write a single question, and weight them differently for each role rather than treating all four as equally important everywhere.
Craft, judgement, measurement, collaboration
- Craft: the specific execution skill the role needs — writing, media buying, technical SEO, lifecycle design. Directly testable.
- Judgement: choosing what not to do. Best surfaced by asking a candidate to allocate a constrained budget and defend what they cut.
- Measurement: whether they can define success before starting, and whether they understand attribution’s limits rather than quoting dashboard numbers.
- Collaboration: how they work with sales, product and finance — the source of most real marketing friction.
A junior role weights craft heavily and judgement lightly. A lead role inverts that: you are buying decisions, not output. Writing the weights down in advance prevents the interview from drifting toward whatever the candidate happens to be good at.
A 1–5 scale that different interviewers read the same way
A numeric scale is worthless unless each point has an observable anchor. “4 out of 5 on judgement” means nothing across two interviewers; “identified a trade-off unprompted and named the cost of the option they rejected” means the same thing to everyone.
- 1 — Cannot perform. Answers describe the concept but not any execution of it.
- 2 — Performs with supervision. Correct steps when told which steps; no independent diagnosis.
- 3 — Performs independently. Handles the standard case end to end, including its usual failure modes.
- 4 — Handles the non-standard case. Adapts when the situation breaks the playbook and can explain why.
- 5 — Improves the system. Changes how the team works, not just their own output.
Most hires for most roles should land at 3. Insisting on 4s and 5s across every group is how a search stays open for six months and ends in a compromise nobody planned.
Three test formats that produce real signal
A test earns its place only if a strong and a weak candidate would visibly do it differently. Anything both could complete adequately is a formality.
- The diagnosis. Give real, anonymised data showing a problem — a funnel with an unexplained drop-off. Ask what they would check first and what would change their mind. This tests reasoning without requiring hours of work.
- The critique. Hand them an existing campaign or landing page and ask what they would change and why. Cheap to run, and it separates people with a point of view from people with a checklist.
- The constrained plan. A fixed budget, a fixed timeframe, one goal. The valuable part is not the plan but the defence of what they excluded.
Keep any take-home under two hours and never ask candidates to solve a live business problem you intend to act on. Unpaid work disguised as assessment costs you the strongest applicants first — they have other options.
Scoring and making the decision
Score immediately after each interview, before any discussion with other interviewers. A group conversation held first produces convergence, not accuracy — the most senior or most confident voice sets the anchor and everyone drifts toward it.
- Record evidence, not adjectives. Next to each score, write the thing the candidate actually said or did.
- Set a floor per group. A candidate at 5 on craft and 1 on collaboration is not an average of 3; they are a specific and predictable problem.
- Discuss disagreements, not averages. Where two interviewers differ by two points, one of them saw something the other missed. That conversation carries the information.
- Decide against the bar, not the field. “Best of the five we saw” is not the same as “meets the bar”, and a weak applicant pool makes the two feel identical.
Biases that distort marketing interviews
Each of these is common enough to be worth naming out loud before a hiring panel starts, because they are much easier to catch in someone else than in yourself.
- Brand halo. A well-known former employer says a lot about that company’s hiring bar and very little about what this person did there.
- Tool fluency. Naming platforms confidently is not evidence of results produced with them.
- Similarity bias. Marketing teams that hire people who think alike lose exactly the perspective diversity the function depends on.
- Recency in the panel. The last candidate interviewed is systematically over-rated. Score before the next one walks in.
Designing the interview loop
A criteria framework only works if the process collects evidence against it. Most loops fail by asking four interviewers roughly the same questions and calling the overlap a consensus.
- Assign each interviewer a group. One covers craft, one judgement, one collaboration. Overlap wastes the candidate’s time and produces four readings of the same signal.
- Keep the loop to three or four sessions. Beyond that, marginal information falls and candidate drop-off rises — you lose the people with alternatives first.
- Put the practical exercise early. It is the highest-signal step, so running it late means spending senior time on candidates the exercise would have filtered.
- Include someone who will work alongside the hire. Peers surface collaboration problems that a hiring manager, whom candidates are actively trying to impress, will not see.
- Tell candidates the structure in advance. Assessing improvisation under uncertainty measures how well someone handles a bad process, not how well they do the job.
Write the questions before you meet anyone. Questions invented mid-conversation tend to follow whatever the candidate has just made sound interesting, which is precisely the drift a structured assessment exists to prevent.
Applying the same framework to people already on the team
The framework is more useful internally than in hiring, because you have something no interview provides: months of observed work. It also removes the main weakness of a standard performance review, which is that it evaluates outcomes the person did not fully control.
- Score the same four groups. This separates “the channel underperformed” from “this person made poor decisions”, which a revenue-based review conflates.
- Have the person score themselves first. The disagreements are the conversation. Where both sides agree, there is nothing to discuss.
- Use the gaps to plan development. A 2 on measurement is a specific, addressable brief; “needs to be more strategic” is not.
- Re-score twice a year at most. Competence moves slowly, and frequent scoring turns the framework into surveillance rather than development.
Running the framework on the existing team before a hiring round has a second benefit: it shows which groups the team is already thin on, so the next role can be specified by gap rather than by whichever title feels overdue.
Frequently asked questions about assessing marketers
How do you assess a marketer when you are not a marketer yourself?
Assess reasoning rather than technique. Ask what they would measure, what would tell them the approach was failing, and what they would stop doing. You do not need domain expertise to judge whether someone can define success and change their mind on evidence. For the technical craft portion, bring in one external practitioner for a single session.
Should the assessment differ for agency versus in-house candidates?
The criteria stay the same; the expected profile shifts. Agency backgrounds usually score higher on breadth and speed, in-house on depth and cross-functional work. Neither is better in the abstract — decide which your situation needs before the interviews, not after meeting someone persuasive.
Is a probation period a substitute for careful assessment?
No. Marketing outcomes usually take longer than a probation period to show, so the decision to keep someone often rests on the same impressions the interview did. A structured assessment up front is cheaper than discovering the mismatch a year later.
You can find additional research on structured hiring at Harvard Business Review.
You can find more hands-on marketing guides at https://marketing365.vn.
Content from marketing365 is created for SMEs, online shop owners, and new marketers.
You can also read more articles on the same topic in the Analysis Corner category.
