How much to spend on Facebook ads comes down to one number: what it costs you to win a sale. Meta's system needs about 50 results a week to settle, so start with 50 times your expected cost per purchase. Then grow only as fast as your margin allows.
Start with the 50-results rule.
Every new ad set on Facebook and Instagram goes through a learning phase. Meta says ad sets usually leave it after "about 50 results" in the week after the last significant edit.
Until then, delivery is less stable and cost per purchase usually runs higher. If an ad set cannot reach that level, Ads Manager marks it "Learning limited".
So the useful starting point is a weekly budget of 50 times your cost per purchase, for each ad set you optimise for sales.
| Cost per purchase | Weekly budget for 50 sales | Roughly per day |
|---|---|---|
| £10 | £500 | £71 |
| £20 | £1,000 | £143 |
| £30 | £1,500 | £214 |
| £40 | £2,000 | £286 |
Two ad sets need twice that. This is why a small budget spread across many ad sets rarely works.
Know what a sale is likely to cost.
If you have run ads before, use your own cost per purchase from the last three months. If you have not, benchmarks give you a starting guess.
Triple Whale tracks Meta ad accounts for over 40,000 e-commerce brands. Its latest figures cover August 2025 to July 2026.
| Category | Median cost per purchase on Meta (US$) |
|---|---|
| All brands | $38.99 |
| Apparel and accessories | $36.98 |
| Food and beverage | $38.57 |
| Beauty | $39.31 |
| Sports and outdoors | $44.53 |
| Home and garden | $47.93 |
At the overall median, 50 purchases a week costs about $1,950. Costs are also rising: the median cost per 1,000 impressions went up 13.24% on the year before.
Treat these as a rough guide. Half of those brands pay less, and your price point, creative and offer will move the number more than your category.
Check what you can afford per sale.
Your break-even cost per purchase is your average order value multiplied by your gross margin. Gross margin here means what you keep after product cost, shipping, payment fees and returns.
Take a £60 order at a 40% margin. You keep £24, so £24 is the most a sale can cost before you lose money on it.
If your break-even figure is well below the benchmark for your category, fix that before you spend more. Raise the order value with bundles, improve the margin, or sharpen the offer. Our guide to what makes a good ROAS shows the same sum from the other side.
Repeat customers change this. If people come back and buy again, you can afford to pay more for the first order. Only count repeat sales you can actually see in your store data.
How much to spend on Facebook ads with a small budget.
If 50 sales a week is out of reach, you can still run Meta ads well. You just need to use the budget more carefully.
- Run fewer ad sets. Meta advises against high ad volumes and suggests combining similar ad sets, so each one learns faster.
- Optimise for a more frequent event. Add to cart happens more often than purchase, so an ad set can gather enough results on less money. Switch to purchases once sales build up.
- Keep budgets realistic and steady. Meta warns that very small budgets, or frequent changes, give its system a poor signal.
- Accept some learning-limited ad sets. A limited ad set that sells at a profit is still worth running. Judge it on results, not the label.
A low cost per sale helps most of all. Silverstone Museum runs on Meta alone at a cost per purchase of £1.26. That meant £3.9K of ad spend, spread across nine campaigns, returned £130K in tracked purchases.
Most brands pay far more per sale than a museum ticket costs to sell. But the principle holds: the cheaper each sale, the less budget you need to give Meta enough to learn from.
Scale in steps, not leaps.
Once an ad set is profitable, the question becomes how fast to add money. Meta's own advice is to avoid frequent budget changes, because they can send an ad set back into learning.
Raise budgets in modest steps and leave several days between changes. Watch cost per purchase after each step. If it climbs past your break-even figure, step back.
Expect returns to fall a little as you spend more. Your first pounds reach the people most ready to buy, and each extra pound reaches someone colder. A lower ROAS on a bigger budget can still mean more profit in total.
NINEYARD started on Instagram and Facebook to build a steady revenue stream before adding Google. Meta returned 14× in three months, against a 10× target.
When Meta stops growing at a profit, that is often the point to add another channel. Our comparison of Meta ads and Google Ads covers how to split a budget between them.
Set your first budget in six steps.
- Work out your break-even cost per purchase: average order value times gross margin.
- Estimate your likely cost per purchase, from your own data or the benchmark for your category.
- If the likely cost is above break-even, improve the offer or order value before you spend.
- Multiply the likely cost by 50. That is your weekly budget for one sales-optimised ad set.
- If that is too much, use one ad set, or optimise for add to cart until sales build up.
- Run it for at least a week without significant edits, then raise the budget in steps while sales stay profitable.
Common questions.
Can I run Facebook ads on £10 a day? Yes, but at most benchmark costs that buys only a handful of sales a week. Use one ad set and a more frequent event, such as add to cart, until purchases build up.
Should my ad budget be a percentage of revenue? A percentage is a useful ceiling, but it ignores what a sale costs you. Set the budget from your break-even cost per purchase, then check total revenue against total ad spend each month.
How long should I wait before judging results? At least a week after the last significant edit, which is the window Meta uses for learning. Judge profit over a full month, including returns.
Does the budget cover Instagram too? Yes. Meta ads cover Facebook and Instagram, and one campaign budget can run across both.
Sources.
- Meta Business Help Centre: About the learning phase, accessed October 2026
- Triple Whale: Facebook Ad Benchmarks by Industry, August 2025 to July 2026