How much should
you actually spend?
Budget · 9 October 2026 · 4 minute read
The standard advice, somewhere around five to ten percent of revenue, is a reference point rather than a rule, and it tends to mislead small businesses in both directions.
A more useful method is to work backwards: what is a customer worth to you, how many do you need, and what does it currently cost to get one. That gives you a number you can defend, and more importantly a number you can tell whether you are hitting.
Why percentage rules break at small scale
Percentage of revenue budgeting assumes marketing is a proportional cost, like rent. For an established business with steady demand, that is close enough to true.
For a small or growing business it breaks in two directions at once. If revenue is low because marketing is not working, a percentage of that revenue funds a budget too small to fix the problem, which keeps revenue low. And if you have just landed a large one off project, the same rule tells you to spend heavily in a month when the pipeline does not need it.
The percentage is a sanity check at the end, not the method at the start.
Work backwards from what a customer is worth
You need four numbers. Estimates are fine. Precision is less important than having thought about it at all.
- What a customer is worth. Not the first invoice: the whole relationship. Average value per job, multiplied by how many times a typical customer buys, less your delivery costs.
- How many you want. Over a defined period, and honestly: can you deliver that many.
- What you can afford to pay to get one. A common starting position is somewhere up to a quarter to a third of the gross profit on a customer, more if they buy repeatedly, less if margins are thin.
- What it currently costs you to get one. Total marketing spend over a period, divided by customers acquired in it. Most businesses have never calculated this and are surprised by the answer.
Multiply the number you want by what you can afford to pay for one. That is a budget derived from your business rather than from an article.
If you do not know what a customer costs you to acquire, you are not setting a budget. You are choosing an amount you are comfortable losing.
Split the budget into three
Lumping everything into one figure is how businesses conclude that marketing does not work, when what actually happened is that the media budget was eaten by production.
- Media. Money paid to platforms. Google, Meta, anywhere you buy attention. This is the only part that scales up and down quickly.
- Production. Making the things: photography, video, the website, creative. Lumpy, periodic, and usually underestimated.
- People. Whoever runs it, in house or outside. Fixed and predictable.
Track them separately. When performance changes, you will want to know which of the three moved.
The minimum viable media budget
There is a floor below which paid advertising cannot teach you anything. It is not a universal number, it depends on what a click costs in your market and how often a click becomes an enquiry.
The rough shape of it: you need enough monthly conversions for patterns to be distinguishable from luck. If your budget buys three enquiries a month, you will not be able to tell a good campaign from a bad one, and you will spend a year making confident decisions based on noise.
If the arithmetic says your budget is below that floor, the better move is usually to put the money somewhere with slower but more certain returns, rather than to run an underfunded campaign and conclude that ads do not work.
What to spend on before ads
Paid traffic multiplies whatever is already happening. If the page converts badly, ads buy you more of a bad outcome, faster and at cost.
In most cases the order is: make sure you can measure what happens, make the page clear about what you sell and what happens next, make sure enquiries are actually followed up, and then buy traffic. Spending on ads before those are in place is the single most common way small marketing budgets are wasted.
A free review of the website, the social presence and local search will usually tell you which of those is weakest before any money is committed.
Review the number quarterly
Set the budget, leave it alone long enough to produce readable data, then review it against two questions: what did it cost to acquire a customer, and is that number moving in the right direction.
If cost per customer is falling, spend more. If it is rising, find out why before spending more. If you cannot tell, fix the measurement, because that is the actual problem.
Common questions
What percentage of revenue should go to marketing?
Commonly quoted ranges sit around five to ten percent of revenue, higher for businesses in a growth phase. Treat it as a sanity check rather than a method: a percentage of current revenue funds the marketing that current revenue can afford, which is circular when the goal is to grow.
How much should I spend on Google Ads to start?
Enough to produce a readable number of conversions each month, which depends on your cost per click and your conversion rate rather than on a fixed figure. If the budget would produce only a few enquiries a month, the data will not be interpretable and the money is usually better spent elsewhere first.
Should I spend on ads or on the website first?
The website, in almost every case. Paid traffic multiplies whatever the page already does. If the page converts poorly, advertising buys more of that outcome at cost, and you will conclude that ads do not work when the page was the problem.
How soon should I expect a return?
Paid search can produce enquiries in days, though reliable data takes weeks. SEO, content and local search typically take months. A budget that mixes both should be judged on different timescales, which is why tracking media, production and people separately matters.
Where to start
Not sure which of these is costing you the most?
I'll look at your website, your social presence and how you show up in local search, and tell you plainly which one is the problem. Written findings in five business days. All I need is your website address.