The most common budget question is "what is the minimum I can spend", and it is the wrong question. The minimum is whatever produces enough conversions for the platform to optimise, and below that the money is wasted regardless of how little it is.
So the useful calculation runs backwards from what a customer is worth to you. Here is how to do it, what the realistic floors are, and how to scale without breaking what is working.
Quick Info
Work backwards from customer value
Four numbers give you a defensible budget. Most small businesses have never written them down together.
Example: a customer is worth £1,200, you can spend 20% acquiring them, so target cost per acquisition is £240. One lead in four closes, so target cost per lead is £60. At £60 a lead, £600 a month buys ten leads and two or three customers. That is a real budget with a real expectation attached to it.
A budget without a target cost per acquisition is just an amount of money you are prepared to lose.
The learning floor
Independent of your economics, there is a technical minimum. Meta's optimisation needs roughly fifty conversions per ad set per week to work properly. Below that it is guessing and your results will be erratic.
Spending £5 a day is not a cautious test. It is a payment for insufficient data.
What to expect at each level
£10 to £20 a day
Viable for a local business in a tight radius with a cheap conversion event. Expect slow learning, high variance and a need for patience. Judge over a month, not a week.
£30 to £70 a day
Where most small business accounts start behaving properly. Enough volume to learn, enough to test creative meaningfully, enough to draw conclusions in a fortnight.
£100 to £300 a day
Enough to run prospecting and retargeting separately, test multiple angles simultaneously, and see clear patterns. This is where creative production budget becomes necessary rather than optional.
Above that
Different discipline: incrementality testing, creative pipelines, and the account structure matters less than the volume of new creative you can produce.
How to scale without breaking it
The most common way to ruin a working campaign is doubling its budget on a good day. Large sudden changes reset learning and performance collapses, which people then attribute to the platform.
Slow scaling feels frustrating and is much cheaper than rebuilding a campaign that has lost its footing.
How to split it
That last line is the one most small businesses miss. Media budget without creative budget is a car with no fuel plan.
Testing budgets
A test needs enough spend to produce a conclusion, and the conclusion is about creative or offer, not about whether Facebook works.
When to stop or cut
Pausing is a legitimate decision. Continuing at a loss because you have already spent money on it is not.
The number to actually watch
Blended cost per acquisition: total marketing spend divided by total new customers, across all channels, monthly. It ignores attribution arguments entirely and it is the number your business lives or dies by.
Watch it alongside total new customers. If both spend and customers rise and blended cost holds, you are scaling successfully. If blended cost climbs, you have reached the edge of efficient demand and further spend is buying worse customers.
Frequently Asked Questions
Practically £20 to £30 a day for lead generation, because an ad set needs around fifty conversions a week to optimise. Less is not a cautious test; it is insufficient data.
Work backwards: customer lifetime value, the share of it you can spend acquiring them, your close rate. That gives a target cost per lead and a budget with an expectation attached.
20% to 30% every three to four days. Larger jumps reset learning and performance usually drops before it recovers.
At small budgets, no. Test within one consolidated campaign. Separate testing structures need volume most small accounts do not have.
Two to four weeks to stabilise, and profitability depends on your economics. If your target cost per acquisition is unreachable at any creative quality, the offer or the margin is the problem.
Auction prices rise and so does buying intent. If your product suits the season, yes, and budget for the higher cost rather than being surprised by it.
Before You Go
Set the budget from customer value, respect the learning floor, scale in small steps and reserve money for making new creative. That is the whole discipline.
Then judge it on blended cost per acquisition. The cost per lead levers are what you pull when the number is wrong.
Work backwards from customer value. Scale slowly.
I'm a marketing consultant, entrepreneur and content creator. I help businesses grow through practical marketing, websites, SEO, content and AI.
More About Tariq →