How to read a media plan

Every line should name the supplier that will execute it, because that determines almost everything downstream: what reporting you get, what evidence exists that the media ran, how a pause works and how a failure is reconciled. A line that says only 'programmatic' or 'digital' is hiding the answer to all of those.

It also determines whether the line can happen at all on the timeline you have. A line executing through a booking workflow — cinema, addressable television, premium audio deals — carries a lead time that a programmatic line does not.

AdBuyMCP names the executing adapter on every plan line. That is not unusual generosity; it is that a plan which cannot name the supplier cannot be reconciled afterwards.

Forecast reach, frequency and CPM are the numbers most likely to be treated as facts and least likely to be interrogated. The question to ask is what produced them: a live forecast from the supplier's own API, a rate card, a modelled estimate, or a planner's judgement.

All four are legitimate inputs and they are not equally strong. A modelled reach figure across channels is a formula rather than a count — no single seller can observe the others — and should carry the word modelled wherever it appears rather than in a footnote.

The specific number to distrust is a blended CPM computed by averaging averages across days. It is arithmetically meaningless and it appears constantly.

Third-party audience segments carry their own price on top of the media: roughly £2.50 per thousand impressions on intent topics and £3.00 on firmographic segments. On a £20 CPM that is 12.5 to 15% again, and it is routinely absent from the plan and present on the invoice.

The reason it matters beyond the money is that it changes the comparison between lines. A cheap-looking CTV line with heavy segment targeting may cost more per thousand than an expensive-looking one with none.

If the plan does not break it out, ask for it explicitly per line rather than as a total, because the totals hide which line the cost sits on.

Every plan implies a targeting claim and almost none quantifies it. The same brief compiles very differently by channel: a business audience can carry a person-level claim on connected TV where consent-chained segments match, becomes a household proxy on addressable television, and is only a property of a place on out-of-home.

AdBuyMCP publishes that as a score out of 100 with a grade and a written rationale, running from 85 on CTV down to 40 on cinema. The number is less important than the habit: knowing which lines in your plan are precise and which are approximations changes how you read the report afterwards.

Where a plan offers no such measure, the proxy question is: what exactly is being selected on this line? If the answer is a household attribute or a venue type, the claim in the covering email should not be about people.

A per-line reason is the field that distinguishes a plan from an allocation. It should say what job the line does — reach, frequency, demand capture, attention, incremental audience — and why that budget rather than more or less.

The test is whether removing the line would change anything you could name. If nobody can articulate what a line contributes, it is usually there because the channel was on the list.

This is also where under-funded lines get caught. A line too small to pace properly does not do the job it was added for, and it would be better folded into a line that can. Enforced plan floors exist for exactly that: below them, a line is dropped and its budget redistributed rather than left to fail quietly.