Resources
How to Set a Marketing Budget
A marketing budget should follow from your economics and your constraints, not from a percentage of revenue. Here is a method that works, with worked allocations for three business models.
Why is the percentage-of-revenue rule unreliable?
The 'spend X% of revenue on marketing' heuristic is widely repeated and widely misleading. It assumes revenue is knowable in advance, that marketing's effect on revenue is proportional and immediate, and that the right percentage is similar across businesses. None of those is true.
Two businesses at 5% of revenue can be in completely different positions. One has 30% gross margins and £200 average order values; the other has 80% margins and £20,000 contracts. The same percentage produces radically different acquisition costs and radically different payback periods. Worse, in a growth-stage business the correct budget is often higher than current revenue supports, which the rule cannot express at all.
What is the three-budget model?
Split the budget into three functions with different rules, because they behave differently and are usually managed by different people.
| Budget | What it covers | How to size it | Review cadence |
|---|---|---|---|
| Acquisition | Paid search, paid social, outbound, events | What your break-even CPA allows, tested in stages | Weekly |
| Retention | Email, lifecycle, CRM, loyalty, customer success support | Enough to retain a target proportion of customers acquired | Monthly |
| Measurement | Analytics, call tracking, attribution, reporting time | Enough that acquisition decisions are trustworthy | Annually, plus on every change |
The mistake is usually allocating acquisition first and letting whatever remains fund retention and measurement. Acquisition is the easiest budget to justify and the one with the fastest feedback loop, which is precisely why it absorbs everything. Retention work is unfashionable, slow to show results, and consistently under-funded — while being the cheapest revenue available.
How should budget change as the business grows?
| Stage | Budget priority | What to fund first | What to hold back |
|---|---|---|---|
| Pre-revenue / validation | Learning, not scale | Cheap tests of message and offer; enough content to be findable | Anything with a monthly commitment |
| Early traction | Proving a repeatable channel works | The one channel that has shown conversion; enough content to compound | Second and third channels |
| Scaling | Capacity and efficiency | The proven channel, plus retention and the ops to service it | New channels until the first is saturated |
| Established | Diversification and efficiency | New channels, automation, brand and market development | Underperforming campaigns and tools nobody uses |
Should budget go to people, media or tools?
For most businesses under about fifty people, people should be the largest line. The reasoning is straightforward: media and tools can be changed or canceled this month, but the capability to use them compounds. A business that outspends its capability in paid media acquires customers it cannot service or convert.
| Line | Typical share (indicative) | Note | Review trigger |
|---|---|---|---|
| People | Often the largest, 40–60% | Capability compounds; hard to hire quickly later | Understaffed before overstaffed |
| Media | 25–45% | Rented; only works with a working offer and tracking | Cut when break-even CPA is not met |
| Tools | 5–10% | Buy for a decision you will actually make | Audit annually; unused tools cancel |
| Measurement | 5–10% | Cheap insurance on every other spend | Cut last, never first |
These are indicative planning ranges only, not recommendations for your business. The durable insight is the ordering: people first, measurement last. A business with excellent tooling and no one to act on it has bought reporting, not results.
How do you ramp without wasting money?
Most waste comes from scaling before a channel has proven itself, then pulling back when performance normalises and concluding the channel failed. A staged ramp avoids both.
Prove
Scale only if cost per acquisition sits below break-even at test volume. If it does not, fix the offer or the landing page before spending more.Scale
Increase in increments, watching whether cost per acquisition rises faster than conversion rate improves. If it does, you are buying less with each additional unit.Efficient
At this point look for margin improvements — better conversion, automation, retention — rather than more spend.How do you cut budget without breaking growth?
Cut in this order. It is roughly the reverse of the order in which budget is usually added, which is why cutting feels so uncomfortable.
- Unused tools and subscriptions. Free, immediate, and frequently several thousand a year. Audit before anything else.
- Platforms with no attribution. You cannot improve what you cannot measure. Keeping an unmeasurable channel on a 'maybe it works' basis is not prudence.
- Underperforming campaigns. The obvious cut, and it should be last in practice because it is the only one that requires data.
- Weak channels generally. Where you cannot see a plausible path to break-even, stop and revisit when the economics change.
- Acquisition spend in the proven channel. Cut last. This is revenue you already know works.
- Never cut measurement. Cutting the thing that tells you what works is how budgets become unaccountable.
Frequently asked questions
What percentage of revenue should I spend on marketing?
Treat the percentage as a sanity check, not a rule. It assumes revenue is knowable in advance and that marketing's effect is proportional and immediate, neither of which is true. Budget should follow from your unit economics and strategic stage instead. Two businesses at the same percentage can face completely different acquisition costs and payback periods.
How much should a small business spend on marketing?
Enough to test properly and to retain what you acquire. A useful starting point is enough to generate statistically meaningful data on one channel — often tens of conversions per month — rather than a round number chosen for comfort. If that is unaffordable, the honest conclusion is to concentrate on organic, referral and retention rather than to spend a smaller amount on paid.
Should I increase budget when I get more traffic?
Usually no. More traffic does not automatically mean more revenue, and adding budget in response to traffic rather than to profit is a common and expensive reflex. Scale on the basis of cost per acquisition against your break-even figure, and watch whether cost per acquisition rises faster than conversion rate improves as you increase spend.
How do I decide what to cut when budgets are tight?
In this order: unused tools and subscriptions first, then channels you cannot attribute, then underperforming campaigns, then weak channels generally, then your proven channel last. Never cut measurement — it is what makes every other budget accountable, and cutting it is how spending becomes unaccountable.
How do I know if I'm over-spending on paid?
If cost per acquisition is above your break-even figure, which you derive from your margin and conversion rate. Rising traffic with rising cost per acquisition, and no improving conversion rate to offset it, means you are buying less with each additional unit of spend. That is the signal to stop scaling and fix the economics underneath.
Should I fund retention before acquisition?
Often yes, and it is frequently under-funded. Retention is cheaper than acquisition, produces faster results, and the budget is harder to justify because it is unfashionable and slow to show. The common failure is funding acquisition first and letting whatever remains fund retention — which usually means retention is not funded at all.
What is the right budget for a startup with no revenue?
Enough to test the offer and the message cheaply, and to be findable. Concretely: content that compounds, a small number of targeted experiments, and almost nothing with a monthly commitment. Paid acquisition before demand is proven usually buys data at an expensive price, and the data is worth having — but it should be bought knowing that is what you are doing.
How do I split budget between channels when I don't know what works?
Test one channel properly rather than testing five poorly. Most budget allocated to multiple small tests produces no result from any of them. Pick the channel that best matches your economics and your audience, fund it to the point where you can learn something, and only then consider a second.
Keep reading
Want this applied to your business?
Stop guessing what to prioritize. Get a straight answer on where your marketing is leaking money, and what to fix first.
Book a free 30-minute audit