How should a marketing budget be split across channels?
There is no correct percentage split, and the benchmarks circulated as if there were describe averages of companies with different margins, different sales cycles and different amounts of existing demand. What does transfer is the reasoning: how much demand exists to capture, what it costs to capture it, and what happens to the money once capture is saturated. Those three questions produce a defensible split for a specific company, which is the only kind that matters.
Why don't benchmark splits work?
Because the split is downstream of variables that differ enormously between companies: how many people are searching for your category right now, what a customer is worth, how long the decision takes, and how well known you already are. Two companies of identical size in the same sector can justify very different allocations on those grounds alone.
The specific reason a borrowed percentage misleads is that it assumes your demand ceiling matches theirs. A company in a well-established category with heavy search volume can spend a large share on capture productively. A company selling something people do not know exists cannot, because there is nothing to capture, and copying the first company's split simply overspends into a small pool of searches at rising cost.
So the useful starting question is not what proportion others spend. It is how much money your existing demand can absorb before the cost of each additional customer starts climbing, which is a number you can find from your own account.
What is the demand ceiling and how do you find it?
The demand ceiling is the point at which spending more on capture channels stops producing proportionally more customers, because you are already reaching most of the people actively looking. It is the single most important number in a budget conversation and most plans never establish it.
You find it by looking at what happened the last few times spend increased. If a rise in budget produced a roughly proportional rise in customers, you are below the ceiling and capture spending is still the best available use of money. If the customer count rose much less than the spend did, and the cost per acquisition climbed, you have found the ceiling and additional money must go somewhere that creates demand rather than harvesting it.
In paid search accounts there is a more direct read: the impression share you are missing for budget reasons. If you are capturing only a fraction of the searches you could and the economics still work, the allocation argument is over and the money should go there.
What should the split look like at different stages?
Directionally, as below. These are starting positions to be moved by evidence from your own account, not targets to hit, and the right response to any of them is to test rather than to adopt.
| Situation | Capture channels | Demand creation | Retention and owned | Reasoning |
|---|---|---|---|---|
| New, category already searched | Most of the budget | Minimal | Enough to keep a list | Harvest existing demand before creating more |
| New, category not searched | Little | Most of the budget | Build the list from day one | Nothing to capture yet |
| Capture saturated, growth flat | Hold, do not increase | The growth money goes here | Increasing | Extra capture spend now buys the same customers dearer |
| Long sales cycle, several deciders | Moderate | Substantial and sustained | Large, this is where deals are won | Familiarity and follow-up beat interception |
| High repeat purchase | Moderate | Moderate | The largest share | Existing customers are the cheapest revenue |
| Budget cut mid-year | Protect the profitable core | Cut last, not first | Protect entirely | Cutting creation shows up two quarters later |
How should the split change as the company grows?
It should move from capture towards creation, and it usually does the opposite. Early on, harvesting existing demand is efficient and reports well, so the budget concentrates there. As the company grows past the size of that demand pool, the same allocation produces worse returns each quarter, and the natural response is to conclude the channel has degraded rather than that the pool has been exhausted.
The transition is uncomfortable because it means moving money from something with clean numbers to something without. This is why it tends to happen late, after two or three quarters of declining efficiency, and after the marketing team has been asked repeatedly why the cost per acquisition keeps rising.
The signal to watch is the trend in cost per acquisition at constant spend. If it climbs steadily while nothing about the market changed, the pool is thinning. That is a budget-allocation problem presented as a channel-performance problem, and treating it as the latter leads to switching agencies rather than switching strategy.
What should never be cut first?
Anything you own: the customer list, the website, the content archive and the retention programme. These are the cheapest revenue available and the slowest to rebuild, and they are usually the first casualties of a mid-year reduction because pausing them has no immediate visible cost.
The second thing to protect is any demand creation that has been running long enough to work. Awareness activity has a build-up period, and stopping it discards that accumulation entirely, so a six-month pause is not a six-month saving. It is a reset. If the choice is between running awareness at half the level and pausing it, halving it is almost always better.
What can be cut with the least damage is anything you cannot answer this question about: if we paused this, what would we lose and when would we notice. Run that question across every line before a cut is imposed, because it produces a better answer than an across-the-board percentage reduction, which cuts the working things and the wasteful things equally.
How often should the split change?
Quarterly for the proportions, monthly at most for the amounts inside a channel. Changing the strategic split more often than the sales cycle produces noise, because you are reacting to results from an allocation that has not finished playing out.
The exception is a clear, sustained signal: capture costs rising at constant volume, a channel that has stopped delivering for reasons you can name, or a demand-creation test that produced a measurable movement in branded search. Those justify a mid-quarter move. A single poor month does not.
Set the review date in advance and hold it, because the alternative is that the split changes whenever someone senior reads something, which is how budgets end up spread thinly across many channels, none funded enough to work. Concentration is usually better than balance at small budgets, and the discipline of a scheduled review is what keeps it concentrated.
Common questions
- Is there a standard marketing budget split across channels?
- No split transfers between companies, because allocation depends on how much demand already exists in the category, what a customer is worth, how long the decision takes and how well known the company is. Benchmarks describe averages of businesses with different economics. What transfers is the reasoning: establish how much demand exists to capture, what capturing it costs, and where money should go once capture saturates.
- What is a demand ceiling in marketing?
- The point at which spending more on capture channels stops producing proportionally more customers, because most people actively looking are already being reached. It is found by examining what happened the last few times spend increased: proportional growth means the ceiling is not yet reached, while rising cost per acquisition with flat customer growth means it has been.
- How should a marketing budget change as a company grows?
- It should shift from capturing existing demand towards creating new demand, because the pool of people already searching is finite and gets exhausted. Companies usually delay this, since capture reports cleanly and creation does not. The signal is cost per acquisition climbing steadily at constant spend with no market change, which is an allocation problem often misread as channel underperformance.
- What should be cut first when a marketing budget is reduced?
- Not the owned assets and not long-running awareness work. Customer lists, the website, the content archive and retention programmes are the cheapest revenue and the slowest to rebuild. Awareness activity accumulates, so pausing it resets rather than saves. Cut whatever nobody can answer this question about: if we paused this, what would we lose and when would we notice.
- How often should marketing budget allocation be reviewed?
- Quarterly for the split between channels, and no more than monthly for amounts within a channel. Changing the strategic split faster than the sales cycle means reacting to results that have not finished arriving. Mid-quarter changes are justified by sustained signals such as rising capture costs at constant volume, not by a single poor month.
