How to allocate a media budget across channels

Start with the decision the budget needs to make

The useful question is not 'what percentage should go to social?' It is 'what job does each part of the budget need to do, and how much spend can each channel realistically absorb while still doing that job well?' That distinction matters because a channel can look attractive on historic averages and still be the wrong place for the next £50K.

Before allocating money, write down four things: the commercial objective, the audience, the campaign period and the primary KPI for each channel role. If those are unclear, the spreadsheet will produce precise-looking numbers without a defensible planning decision behind them.

Step 1: define channel roles before percentages

Separate channels by the job they are expected to do. Paid search may capture existing demand. Video may build reach. Paid social may create demand, retarget site visitors or both. Programmatic may add incremental reach or act as a controlled retargeting layer. The same platform can play different roles, so the role needs to be explicit at plan-line level.

Step 2: calculate whether each channel can absorb the proposed spend

This is the check most percentage-based allocation guides miss. A budget can only be placed efficiently where there is enough demand, audience or inventory to absorb it. Search is the clearest example: if the available non-brand query volume cannot support £200K without materially inflating CPCs or expanding into weaker intent, then £200K is not a sensible allocation simply because search has the best historic CPA.

For reach-led channels, capacity may be constrained by audience size and acceptable frequency. For performance channels, it may be constrained by qualified demand, conversion volume or the amount of expansion you are willing to accept before efficiency drops. This is why audience definition and budget allocation belong in the same planning process.

Worked example: a £500K eight-week acquisition plan

The following is an illustrative planning example, not Medusa customer data. Assume a UK advertiser has £500K for eight weeks, is focused on new-customer acquisition and has enough creative to run paid search, Meta, YouTube and a controlled test budget.

Now challenge that first pass. Search forecasting suggests the account can absorb around £110K over the period before the planner must materially broaden queries or accept higher marginal CPCs. That means £70K of the original search allocation does not yet have a defensible home.

What happens to the £70K?

Do not automatically move it to the channel with the second-best historic CPA. Re-run the capacity check. If Meta is already forecast to reach the core audience at high frequency by week six, another £70K there may buy repetition rather than useful reach. If YouTube can add incremental reach at an acceptable CPM and the business needs more demand creation, part of the money may move there. If neither case is strong enough, keep some budget unallocated until early delivery provides better evidence.

Step 3: record the assumptions behind every line

A strong media plan does not only show the final split. It records why the split exists. At minimum, capture the audience, role, cost assumption, expected delivery, KPI, measurement method and the condition under which the budget would be increased or reduced. This turns allocation from a one-off spreadsheet exercise into a decision that can be reviewed later.

• Audience assumption: how large is the targetable audience and what exclusions apply?
• Cost assumption: which CPM, CPC or CPA range was used, and is it based on account history or a benchmark?
• Capacity assumption: how much can this line plausibly spend before demand, reach or frequency becomes a constraint?
• Measurement confidence: is the KPI directly observed, modelled, attributed or dependent on a proxy?
• Reallocation rule: what evidence would justify moving the next £10K into or out of this line?

Do not confuse average efficiency with marginal efficiency

Historic average CPA tells you what a channel did across all previous spend. The reallocation question is different: what is the expected return from the next unit of budget? A channel that averaged a £40 CPA may produce the next block of conversions at £70 once the highest-intent demand has been captured. Another channel averaging £55 may still have room to scale at £58. In that case, moving budget toward the second channel can be rational even though its historic average looks worse.

This is also why a fixed 70-20-10 rule should be treated as a governance device, not an allocation formula. Ring-fencing test budget can be useful, but the correct share depends on the maturity of the account, the quality of existing evidence and the size of the campaign.

When the launch allocation should change

• Demand is lower than forecast. Search or other intent channels cannot absorb the planned spend without weakening targeting.
• Frequency rises too quickly. Reach-led channels are repeating impressions rather than adding enough incremental audience.
• The objective changes. A plan built for acquisition should not survive unchanged if the business moves toward reach, retention or a product launch.
• Marginal efficiency deteriorates. The next block of spend is materially less efficient than alternatives.
• Measurement proves unreliable. A channel should not receive more budget simply because a weak attribution method makes it appear efficient.

Use the approved allocation as the pacing baseline

Once approved, the allocation should become the benchmark for live delivery. Actual spend needs to be compared with planned spend to date, using the intended flighting rather than a generic straight line. That is the point where planning and operations meet: budget pacing tells you whether the campaign is delivering as approved, while in-flight optimisation determines whether the approved plan itself should change.