Data-driven marketing is not a dashboard

The phrase has been used so widely that it no longer distinguishes anyone. Here is the definition we hold ourselves to, and the three tests that separate it from reporting.

Reporting is a description of the past. A dashboard is reporting with colours. Neither one is a decision. Data-driven marketing, if the words mean anything, has to describe how decisions get made, not how results get displayed afterwards.

Our working definition is short. A marketing decision is data-driven when it was ranked against the alternatives by evidence, baselined before the money was spent, and re-scored after, with the evidence attached to the result. Every clause does work.

Test one: was the decision ranked, or just justified?

Most marketing plans are assembled and then supported with data. The channel mix was decided in a meeting; the numbers were found later. A ranked decision starts the other way round: here are the twelve things we could do, here is what the evidence says each one would move, here is the order. The plan is whatever comes out of the ranking, including when it is inconvenient for the agency.

Test two: was there a baseline before the spend?

You cannot know what marketing changed if you do not know where the brand stood the day before. A baseline is not last month's dashboard. It is a scored, dated position across every dimension the plan might move, against the peers the brand is actually competing with. This is why every Colleviate engagement begins with an audit and not with a campaign: the audit is the baseline.

Test three: was the result re-scored with evidence, or reported with a narrative?

"Brand awareness improved" is a narrative. "Share of search moved from 23 to 27 percent against the same four peers, measured the same way as the baseline, while paid spend was flat" is a re-score. The second one can be wrong, which is exactly what makes it useful. A claim that cannot be wrong is not evidence of anything.

What this rules out

• Vanity metrics without a peer comparison. Impressions, followers and sessions are inputs, not outcomes, and they mean nothing without a competitive baseline.
• Attribution by assertion. If a channel is credited with revenue, the model that credits it should be inspectable. Attribution is a measurement problem before it is a marketing one.
• Benchmarks without sources. "Industry average CTR" is an anecdote unless the industry, the period and the sample are stated.
• Plans that cannot fail. Every line in a data-driven plan says what would show it did not work.

Why most agencies stop at the dashboard

Because a dashboard is cheap to build and impossible to be wrong about. A ranked, baselined, re-scored plan is expensive to build and easy to be wrong about. The second one is the only one that lets a client learn anything, and the only one that lets an agency improve. We built our practice so that the expensive version is affordable: the outside-in read, the peer cluster and the re-score are automated, which leaves the humans to do the ranking and the work.

A simple way to check your current agency

Ask for last quarter's plan and last quarter's results side by side. If the plan says what each line was expected to move and the results say what it actually moved, against the same baseline, you have a data-driven partner. If the plan is a list of activities and the results are a dashboard, you have a reporting partner. Both can be fine. Only one can be held to account.