Marketing interviews test whether you can connect activity to numbers — channels, funnels, attribution, and unit economics. Below are the most common questions with model answers. Middle: deeper understanding, optimization, and real-world situations.
A typical funnel runs impression, click, visit, lead, activation, paid, retained — each with its own conversion rate. The value is not the diagram but locating the biggest drop-off, because improving a 2 percent step to 4 percent doubles everything downstream. Measure each stage separately by channel, since a channel with cheap clicks and terrible activation looks good only in aggregate.
ROMI is revenue attributable to marketing minus its cost, divided by cost — it answers whether a channel returns more than it consumes. Judge a channel on cost per activated user rather than cost per click or lead, because cheap traffic that never converts is expensive. You also need enough time for the payback window: cutting a channel after two weeks judges it before its revenue arrives.
Attribution assigns credit for a conversion to the touchpoints that led to it. Last-click gives everything to the final interaction, which is simple but systematically overvalues branded search and undervalues awareness channels that started the journey. Multi-touch distributes credit across touchpoints and is closer to reality but harder to implement and easier to argue about. Many teams pair a simple model with incrementality tests, which measure lift by holding out a group.
Start with a hypothesis and the metric it should move, calculate the sample size needed to detect the effect you care about, then run until that sample is reached rather than until the result looks good. Statistical significance is the probability the observed difference is not chance, conventionally at 95 percent. The most common mistakes are stopping early, testing several changes at once, and running for less than a full business cycle.
Break it down before acting: rising CAC can come from auction competition, audience saturation, worsening creative, or a change further down the funnel that has nothing to do with acquisition. Check whether the drop is channel-specific or global, whether traffic composition shifted, and whether a product or pricing change coincided. Raising budget on a channel with deteriorating economics accelerates the problem.
Retention is the share of users still active after a given period, and it is the metric that decides whether acquisition compounds or leaks. Cohort analysis groups users by when they joined and tracks each group over time, which separates real improvement from a mix shift — an overall retention number can rise simply because you acquired more of a better segment. A flattening retention curve indicates product-market fit far better than growth does.
In a real interview you speak under pressure. Cam asks these same questions, scores every answer, and shows exactly what to fix.
Practice a Middle Marketing interview →