What a £30m media plan teaches you about a £30k one
The planning discipline that big budgets force on you scales down. Most small accounts skip it, and that is why they stay small.
Joseph has planned media at every scale, from four-figure test budgets to £30m a year for a betting and gaming group. The surprise is how little changes. The big plans are not more complicated. They are more disciplined, because at that scale a lazy assumption costs a million pounds. The discipline works just as well at £30k a month. Here is what it looks like.
Start from the business number, not the media number
A £30m plan starts with the revenue the business needs and the margin it makes. Media is worked backwards from there. How many customers, at what value, at what cost, in which markets, in which months.
Most £30k accounts start the other way round: “we have £30k, what can we buy?” The target ROAS is whatever it was last year. Nobody has checked it against the margin. Start with the business number. If you do not know the contribution margin on a sale, find out before you spend another pound.
Plan the year, not the month
Big plans are built for the year with the seasonality in them. The budget for the Grand National is agreed in November. Small accounts set a flat monthly budget and then wonder why they are capped in the weeks that matter.
Take last year’s revenue by week. Overlay the events that drive demand. Put the budget where the demand is. A flat budget across a seasonal business is a decision to under-spend in the good weeks and over-spend in the bad ones.
Separate the channels by job
At scale, every channel has a written job. Search captures demand. Social creates it. Affiliates extend reach into places the brand cannot go. Each has its own target because each does a different job.
Small accounts often give every channel the same ROAS target and then cut the one that looks worst. That is usually the one doing the hardest job. Give each channel a job and a target that fits it.
Measure incrementality, not attribution
With £30m at stake, nobody accepts the platform’s word for what it delivered. Geo tests, holdouts and marketing mix modelling are standard. The finding, almost always, is that brand search and remarketing are less incremental than they look, and prospecting is more.
You can run a holdout at £30k a month. Turn off remarketing for a month in half your regions. See what happens to sales. It costs little and it settles arguments that dashboards never will.
Write the assumptions down
A big plan has a page of assumptions: conversion rate, average order value, margin, seasonality index, the attribution model. When the results miss, you can see which assumption was wrong.
A small account has the assumptions in someone’s head. When results miss, the answer is “search is not working”. Write them down. One page. Review it quarterly.
Have a governance rhythm
At scale: a weekly trading meeting with the numbers, a monthly review of the plan against actuals, a quarterly re-plan. Decisions get made in those meetings and written down.
At £30k a month, a thirty-minute weekly review with the same three numbers (spend, revenue, contribution) and a written note of what changed will do most of the same work. The point is that someone looks, every week, with the plan in front of them.
The one thing that does not scale down
Big accounts can afford tests that small ones cannot. A £30k account cannot hold out ten per cent of budget to test a new channel every month. So test less often and test bigger. One proper test a quarter beats twelve small ones that never reach significance.
The short version
Start from margin. Plan the year. Give each channel a job. Measure incrementality. Write the assumptions down. Look every week. None of it needs £30m. It needs the discipline that £30m forces.
Talk to Joseph about your account
A thirty-minute call. You describe the account, Joseph tells you what he would look at first. No deck, no pitch.