There is no universal ad budget, only a formula. Work out what a customer is worth to the business, decide what share of that value can be spent to acquire one, then check that the resulting spend is large enough for the ad platform's own algorithm to gather enough signal to optimise properly. A budget set below that threshold rarely fails because the product or the creative was wrong. It fails because the platform never collected enough data to know who to show the ad to. The honest starting point is customer value and an acceptable cost per acquisition, worked out before a single pound is spent, not a flat number borrowed from a blog post or a competitor's guess.
There's no universal number, only a formula
"How much should I spend on ads" has no single answer, because the number depends entirely on what a customer is worth to a specific business. The right question isn't "what should I spend," it's "what can I afford to spend to get one customer, and is that spend big enough for the platform to actually learn from."
Every "spend £X a month" answer floating around online is really a guess dressed up as a rule. A business with a £300 average order value and healthy margin can afford a very different cost per customer than one selling a £20 product on thin margin, even if both are "small businesses" on paper. The framework below replaces the guess with three questions, asked in order: what is a customer worth, what can be paid to acquire one, and is that spend large enough for the platform to do its job.
Start with what a customer is worth
Customer value, not competitor spend or industry averages, is the only honest starting point for a budget. That means knowing the margin on an average sale, and ideally the value of a customer over time, not just their first purchase.
A single sale's revenue is the wrong number to anchor on. Two businesses can sell the same £50 product, but if one customer buys once and the other reorders four times a year, the second business can justify paying far more to acquire that same customer and still come out ahead. This is why lifetime value, not order value, is the real anchor: a budget built on first-purchase revenue alone systematically underspends on the customers who are actually worth the most.
The broader discipline here has a well-known reference point. Venture investor David Skok, introducing a piece on his For Entrepreneurs blog, describes "having introduced the goal that this should be greater than 3 for a healthy SaaS business," referring to the ratio of lifetime value to acquisition cost. Skok raises it to qualify it: in the same note he says he "made a significant mistake in not telling my readers when it would make sense to compute LTV and CAC." That heuristic was built for recurring-revenue software businesses with long customer lifetimes, not for every ad-spending business, so it should not be copied over as a literal rule. What it does establish, reliably, is the underlying logic: acceptable acquisition cost is a function of what a customer is genuinely worth, not a flat percentage borrowed from an unrelated industry.
Work out what you can actually afford to pay for one
The maximum acceptable cost per acquisition comes from customer value minus the margin the business needs to keep, not from what feels affordable on a given day. Working this number out before spending anything is what turns an ad budget into a business decision instead of a guess.
Once customer value is known, the next step is subtraction, not addition. Take what a customer is worth, decide what margin has to be protected, and whatever is left over is the ceiling for what can be spent to acquire that customer. Spend above that ceiling and every new customer makes the business poorer, no matter how good the return-on-ad-spend number looks in a screenshot. This is also where the payback window matters: a business that needs cash back within 60 days can afford a lower acquisition cost than one that can wait a year for a customer to become profitable, even with an identical lifetime value. For the deeper version of this problem, including why a strong sales month can still be a bad month once acquisition cost is accounted for properly, see why your best month is really a CAC payback problem.
The minimum spend a platform needs to learn
Ad platforms need enough purchase or lead signals in a short window to learn who is likely to convert. Most platforms' own algorithms want somewhere in the range of fifty conversion or lead events in a rolling week before they have enough signal to optimise reliably, an industry rule of thumb, not a fixed guarantee.
This is the piece the customer-value math alone misses. A business can correctly calculate that it can afford £40 to acquire a customer, set a daily budget that reflects that number, and still get poor results, because the resulting spend never generates enough events for the algorithm to learn from. Below that threshold, the platform is effectively guessing at who to show the ad to, the same as it would on day one, indefinitely. Above it, the algorithm has enough data to start narrowing in on the audience that actually converts. The exact threshold moves by platform and by campaign objective, and no platform publishes it as a fixed, guaranteed number, so it should be treated as a planning rule of thumb, not a promise.
Why "test with a fiver a day" wastes money
"A fiver a day doesn't test anything," Alessandro Lombardo says. "It just spends five pounds a day for a month and calls the result a test. If the spend never crosses the threshold the platform needs to learn, you haven't tested the offer, you've tested how long you can watch nothing happen." After nine years running The Social Target across 600+ clients, with 50+ active today, the pattern holds: a small, "safe" test budget usually produces the least useful data of any spend level, because it sits below the point where the platform can do anything but guess. A genuinely efficient budget isn't the smallest number that feels comfortable, it's the smallest number that's actually large enough to generate a real answer. That's the difference between spending less and spending smart: making your cents work like dollars means putting them where they can actually earn their keep, not spreading them so thin they can't do anything at all.
Putting the framework together
The three steps run in this order, not in isolation:
- Work out customer value. Use lifetime value where the data exists, not just the first sale, and be honest about margin, not revenue.
- Set the acquisition ceiling. Subtract the margin that has to be protected, and treat what's left as the maximum acceptable cost to acquire one customer, adjusted for how quickly the business needs that cash back.
- Check the spend against the learning threshold. If the daily or weekly budget implied by the ceiling sits below the platform's rule-of-thumb learning threshold, either the business needs to raise the acceptable acquisition cost, narrow the targeting so fewer, better-matched events are needed, or accept that testing will take longer than a week to produce a reliable answer.
Once a budget clears all three steps, the question shifts from "how much should I spend" to "how consistently can this spend be maintained." See why ad spend behaves like a duty cycle, not a switch for what happens to an account that starts and stops instead of running steadily once the budget is right.
FAQ
Is there a standard percentage of revenue I should spend on ads? No defensible single percentage exists across industries and margins. A business with high repeat purchase and strong margin can justify a higher spend-to-revenue ratio than a low-margin, one-time-purchase business, so a flat percentage borrowed from a blog post will be wrong for most specific businesses.
How much should I spend to test a new campaign before scaling it? Enough to clear the platform's learning threshold within roughly a week, not an arbitrary "safe" small number. If the planned test budget sits well below that threshold, extending the test window or narrowing the audience is usually a better fix than simply spending longer at too low a level.
What's the difference between ad spend and a marketing budget? Ad spend is the money paid directly to platforms like Meta or Google for placements. A marketing budget is broader: it includes ad spend plus everything else, creative production, tools, and any agency or team cost to run the account.
How long before I can judge whether an ad budget is working? Long enough for the platform to exit its learning period and for at least one full customer payback cycle to complete, whichever is longer. Judging a budget before either has happened usually confuses a slow start with a failed strategy.
Is a bigger budget always better? No. A budget above what customer value can support turns every new customer into a loss, regardless of how efficiently the platform spends it. Bigger only helps once the underlying acquisition-cost ceiling has been calculated and respected.
Does a low daily budget stop Meta or Google from optimising properly? It can. If the resulting spend consistently falls short of the rough weekly threshold the platform's algorithm needs to gather enough signal, the account can stay stuck in a permanent learning state instead of settling into stable, informed delivery.
When should I increase my ad spend? Once the current spend is consistently profitable against the acquisition-cost ceiling and is comfortably clearing the platform's learning threshold. Increasing spend before either condition holds usually just scales the same guesswork to a bigger number.
If you want a second opinion on where your customer-value math lands, or you're trying to work out whether your test budget is actually big enough to mean anything, tell us about your business. No pitch deck required, just a straight answer on what the number should be for your situation. And if the budget question is settled and you want to see how that spend actually gets run day to day, the paid media work this framework feeds into covers how the pieces fit together.