Marketing analytics

Marketing analytics is measuring whether marketing worked. Attribution is genuinely difficult, and being straight about that difficulty is itself a hiring signal — the people most certain about which channel caused a sale are usually the least worth believing.

What the work consists of

Deciding before a campaign what would count as evidence it worked, instrumenting that, and then reporting the result including when the result is inconclusive. Much of the day-to-day is less glamorous: checking that tracking still works after someone changed the site.

What employers screen for

Whether you set up measurement in advance or reverse-engineer a story afterwards. And whether you know the difference between a correlation visible in a dashboard and a result you can act on. A candidate who can describe a campaign that did not work, and what they learned, is demonstrating something rare.

• Defining a metric that cannot be gamed by the person being measured
• Attribution models, and being able to say plainly what each one assumes
• Separating a seasonal effect from an actual change
• Building a report somebody reads, rather than a dashboard nobody opens

The honest limit

Two to four months to competence on the tooling. The analytical judgement underneath takes considerably longer, and it is what divides someone who reports numbers from someone who is asked what to do next. The tooling also changes constantly, which makes the judgement the durable half.