Marketing measurement planning

What marketing measurement planning is.

Marketing measurement is how a business works out what its marketing produced. A measurement plan writes down the approach in advance, including the goal, the numbers that show progress, the events that need to be tracked, the systems that hold the data and the person responsible for each piece.

The timing matters. Once a campaign is live, you can't go back and collect data you never set up. A plan written first also forces people to agree on definitions, such as what counts as a qualified lead, while nobody has a result to defend yet.

When a measurement plan makes sense.

This work pays off before a launch, a new website, a change of agency or a large increase in budget. It also helps when your ad platforms, analytics and CRM each report a different number and meetings keep turning into arguments about which one is right.

A very small account with one conversion and one channel doesn't need a formal framework. A short list of what to track and a check that it works will probably cover it.

What a measurement framework contains.

We build a framework from the business outcome down. It starts with what the company needs, such as revenue, qualified leads, donations or registrations. Under that, we set the marketing numbers expected to drive that outcome and a small set of supporting numbers that help explain why they move.

Each number gets a written definition, a source, an owner and a reporting schedule. We also write down the target and where it came from, because a target with no reasoning behind it can't be challenged.

The technical side follows from there. We list each event to be tracked, what should trigger it, the information it should carry, which platforms need it and how we'll test it. Your developer or tag manager can build from that document directly, or we can scope the build.

Choosing how to measure.

No single method answers every question, so a good plan says which one to use for which decision.

Platform reporting is quick and it's what the bidding systems learn from, but each platform uses its own attribution rules and can claim the same sale as another platform. Adding them together usually overstates the total.

Web analytics gives one view across channels, though it depends on consent and misses people it can't connect across devices. A CRM is closer to the money, since it can show which leads closed, as long as the source of each lead is captured and sales keeps the records up to date.

Incrementality tests compare people or regions who saw the ads with those who didn't, which gets closest to what the ads actually caused. They need enough volume and budget to read, so they aren't right for every account. Marketing mix modeling looks at spend and results over a long period, which makes it a better fit for larger advertisers with years of history.

We match the method to your budget and your data, and we say what each one can't tell you.

Consent and privacy.

A measurement plan needs to account for what you're allowed to collect. Consent rules vary by market, and in Quebec, Law 25 applies to personal information. We plan which data goes where and what happens when someone declines tracking. Legal questions stay with your qualified reviewer.

How we check the plan is working.

We look at three things. Do the systems agree closely enough to trust, and can the differences be explained? Is each important event still arriving, from the right source? And are people using these numbers to make decisions? When the answer to any of those is no, we revise the plan.