'How much should I spend on marketing' is one of the most common questions a new or growing business owner asks, and one of the least usefully answered, because the honest response — 'it depends' — doesn't help anyone actually set a number.
Why generic percentages don't really work
You'll see rules of thumb like 'spend 7-10% of revenue on marketing' repeated widely, and they're not useless as a very rough sense check, but they ignore the two things that actually determine a sensible number for your business: how much a new customer is worth to you over time, and how quickly you need growth versus how much runway you have to be patient.
A business with a high-value, repeat customer can often justify spending far more than 10% to acquire one, because the return plays out over years. A business with thin margins and one-off purchases might find even 5% unaffordable. The percentage was never really the point — it's a shortcut that skips the actual maths.
The right marketing budget isn't a percentage of revenue. It's a function of what a customer is worth to you and how long you can afford to wait for the return.
A more useful way to think about it
Start from customer value, not revenue. Work out roughly what a typical customer is worth to your business over the time they stay with you, not just their first purchase.
That number tells you the ceiling for what you could reasonably spend to acquire one customer and still be profitable — everything below that ceiling is a genuine option, everything above it is a loss dressed up as growth. It won't be a precise figure the first time you calculate it, and that's fine; a rough ceiling is still far more useful than no ceiling at all.
A worked example
Say a typical customer spends £80 with you on their first order, and roughly six in ten come back for a second and third order over the following year, averaging another £140 over that time. Your rough lifetime value is around £220. If your margin on that revenue is 50%, the actual profit you're protecting is closer to £110 — and that, not the £220, is your real ceiling for what you can afford to spend acquiring one customer while still coming out ahead.
That doesn't mean you should spend right up to £110 — a healthy business usually wants a comfortable margin between what a customer costs to acquire and what they're actually worth, to cover the customers who don't repeat, the campaigns that underperform, and the simple fact that the first calculation is always a rough one. But it gives you an actual number to test spending against, rather than a vague sense that £15 a click 'feels like a lot'.
The mistakes that blow the budget anyway
The most common one is looking at a single blended cost-per-customer figure across all marketing activity and missing that it's hiding one channel performing well and another quietly losing money underneath the average. Break the number down by channel — paid social, search, referral, local advertising — before deciding the overall budget is working or not, because the average can look perfectly healthy while one specific channel eats most of the spend for almost none of the return.
The second is optimising for the metric that's easiest to see rather than the one that actually matters — clicks, impressions and website visits all feel like progress, and none of them are revenue. A campaign can generate an excellent click-through rate and a terrible number of actual paying customers, and it's the second number, not the first, that should decide whether the spend continues.
The third is scaling a channel up before understanding its payback period — how long it takes for a customer's spending to actually cover what it cost to acquire them. A channel that's profitable over eighteen months can still cause a genuine cash-flow problem if you scale it aggressively without the cash to cover the gap between spending the money and it coming back.
Starting small and learning fast
If you genuinely have no data yet, start with a small, deliberately limited budget treated as a learning exercise rather than a growth bet — enough to get real data on what it actually costs you to acquire a customer through a specific channel, not so much that a wrong early guess is expensive.
Increase the budget once you have real numbers to justify it, not before. The businesses that get burned by marketing spend are usually the ones that scaled a budget up before they'd actually confirmed the channel worked at a small scale first.
What to do this week
Work out your rough customer lifetime value using whatever data you already have, even if it's only a handful of repeat customers to go on — a rough number beats no number. Then set a small test budget on a single channel, track actual paying customers against it, not clicks, and give it a fixed window, four to six weeks is reasonable, before deciding whether to scale it up, change it, or drop it entirely.
Write the ceiling number down somewhere you'll actually see it before approving spend, not just in a spreadsheet tab you never reopen. The businesses that quietly overspend on marketing rarely do it in one dramatic decision — it happens a campaign at a time, each one only slightly over budget, none of them individually alarming, until the pattern is only visible in hindsight. A number you've genuinely committed to checking against is the simplest defence against that drift.
Common questions
How do I work out customer lifetime value with only a few months of data?
Use what you have and label it provisional rather than waiting a year for certainty. Take your average order value, multiply by the number of times a typical customer has actually bought so far, then apply your gross margin percentage so you end up with profit rather than revenue — the profit figure is the one that sets your ceiling. With three months of trading you cannot see repeat behaviour properly, so assume only the repeat rate you have genuinely observed, with nothing further after it. A cautious number you revisit each quarter beats a generous one built on a hoped-for repeat rate. Write down which inputs you guessed at, so you know what to check first when real data arrives.
Is 7-10% of revenue a reasonable starting point after all?
Only as a sanity check on a number you have already reached another way. The percentage rules get repeated because they are easy to remember, not because they fit any particular business — they take no account of your margin, your repeat rate, or how long you can wait for the money to come back. A business with 70% margins and customers who stay for years can rationally spend well above 10%; a business with 15% margins on one-off sales can be losing money at 5%. Set the budget from your customer-value ceiling first, then compare it with the percentage. If the two are miles apart, re-examine your lifetime value assumptions rather than defaulting to the rule of thumb.
How long should I run a test before deciding a channel does not work?
Long enough to produce a meaningful number of actual customers, which in practice means four to six weeks and enough spend to generate at least 20 to 30 conversions. Below that you are reading noise — three sales in week one tells you close to nothing about week four. The other half of the answer is payback period: a channel that acquires customers profitably over eighteen months will look like a failure at week four if you only count first purchases. Decide before you start what result counts as a pass, what counts as a fail, and how repeat revenue will be treated. Deciding that afterwards is how a channel survives because somebody liked it.
What should I track if I cannot tell where customers came from?
Ask them, and accept a messy answer. A single short, optional 'how did you hear about us?' field at checkout or enquiry gets a small business further than most attribution setups, because self-reported data catches word of mouth and offline channels that analytics never sees. Alongside it, run the bluntest test available: switch one channel off for a month and watch whether total enquiries move. It costs you a month and it answers the question no dashboard can when several channels overlap. Keep a simple monthly record of spend and new customers by channel, even by hand on one sheet — the trend over six months is what you actually need.
Should I cut marketing when money is tight?
Cut the spend you cannot attribute to customers, not the marketing you can prove is working. The reflex is to treat the whole marketing line as discretionary, and that is how a bad quarter becomes a bad year: you stop acquiring customers now, and the revenue hole appears two or three months later when cash is already tighter. Go through it channel by channel instead. Keep anything with a proven cost per customer inside your ceiling, pause anything you are funding on faith, and shorten the payback period you are willing to accept rather than abandoning acquisition altogether. If you genuinely have to stop everything, make it a dated decision with a review point, not a drift.
