Key takeaways
- Don't pick a round number. Work back from your margin to a target CPA, then from the CPA to a budget.
- A scaling ad set ideally gets about 50 purchases a week, so its daily budget is roughly target CPA × 50 ÷ 7.
- A fair creative test needs two to three times your target CPA in spend per concept before you judge it.
- Small budgets still work. Use fewer ad sets, judge over longer windows and accept Learning limited.
- Grow the budget in steps once cost per purchase holds, and check blended revenue against total ad spend every week.
How much you should spend on Facebook ads depends on what a customer is worth to you. Work out the most you can pay for one order (your break-even CPA), set a target below it, then fund roughly 50 purchases a week per ad set: target CPA × 50 ÷ 7 per day. Add a testing budget of two to three times the target CPA per new concept. If that is too much, spend less and judge more slowly.
Why shouldn't you just pick a budget?
Because the right budget is a consequence of your numbers, not a starting point. "$20 a day" is either plenty or hopelessly small depending on what your product costs and earns. A store selling a $15 phone case and a store selling a $400 mattress need completely different budgets to learn anything.
Three numbers decide it:
- Your target CPA, the most you are willing to pay for an order, which comes from your margin.
- How many purchases Meta needs to learn, about 50 a week per ad set according to Meta's learning-phase guidance.
- How many things you want to test, because every new creative needs enough spend to earn a verdict.
How do you calculate your target CPA?
Step 1: find your break-even CPA
Break-even CPA is the cost per purchase at which the first order neither makes nor loses money.
Break-even CPA = average order value × gross margin % − variable costs per order
Variable costs are the ones each order brings: shipping you pay, payment fees, packaging and any average returns cost.
Say your average order value is $70, your gross margin after product cost is 60%, and shipping, payment fees and packaging come to $9 an order:
Break-even CPA = 70 × 0.60 − 9 = $33
Step 2: choose your target CPA
Your target depends on what you want the first order to do:
- Profit on the first order: set the target below break-even, for example $25 here.
- Break even on the first order, profit on repeats: set it near $33, if customers reliably come back. Your customer lifetime value tells you whether that is safe.
- Buy growth deliberately: set it above break-even only if you know your repeat economics well and have the cash to wait.
The CPA calculator and break-even ROAS calculator do these sums for you, and our guide on how to calculate break-even ROAS explains the same idea expressed as ROAS.
How much daily budget does Meta need?
Meta says an ad set usually leaves the learning phase after about 50 results in the week after its last significant edit. If you optimise for purchases, that means 50 purchases a week.
Daily budget per ad set ≈ target CPA × 50 ÷ 7
With a $25 target CPA: 25 × 50 ÷ 7 ≈ $179 a day for one ad set to have a good chance of exiting learning.
That is the ideal, not a minimum. Plenty of ad sets are profitable below it, as covered in the small-budget section below.
How much should you spend on testing?
Testing money is separate from scaling money. A test needs enough spend per concept that a result means something.
Spend per concept ≈ 2 to 3 × target CPA
At a $25 target CPA, each concept needs roughly $50 to $75 before you call it. An ad that has spent $75 with no purchase is unlikely to become a winner. One with two or three purchases at or below target deserves more.
Weekly testing budget = concepts per week × spend per concept
If you test four new concepts a week at $75 each, that is $300 a week, about $43 a day.
How many concepts should you test? Enough to keep a steady supply of fresh winners, because every ad eventually tires. For most small stores, three to five a week is plenty. Our creative testing framework shows how to run the tests, and ad fatigue explains why you need the pipeline.
How do you set a monthly Facebook ads budget?
Put the pieces together:
Monthly budget = (scaling daily budget + testing daily budget) × 30
Or work back from a revenue goal:
Monthly budget = monthly revenue goal from ads ÷ target ROAS
where target ROAS = average order value ÷ target CPA.
Worked example: a skincare store
Say you sell a serum. All figures are hypothetical:
- Average order value: $70
- Break-even CPA: $33
- Target CPA: $25, so target ROAS = 70 ÷ 25 = 2.8
- Goal: $15,000 a month in ad-attributed revenue
Monthly budget = 15,000 ÷ 2.8 ≈ $5,360, or about $179 a day.
That happens to match one well-fed scaling ad set at 50 purchases a week. A sensible split:
| Use | Share | Daily | What it buys |
|---|---|---|---|
| Scaling | 75–80% | about $135–$145 | One ad set with proven ads |
| Testing | 20–25% | about $35–$45 | Three to four new concepts a week |
The scaling ad set alone won't quite reach 50 purchases a week at this split, which is fine. Expect Learning limited at times and judge weekly. The Meta ads budget calculator runs this for your own numbers.
Worked example: a higher-priced product
Now say you sell a $240 piece of furniture with a 45% gross margin and $35 of delivery and fees per order.
- Break-even CPA = 240 × 0.45 − 35 = $73
- Target CPA: $60
- Daily budget for 50 purchases a week: 60 × 50 ÷ 7 ≈ $430
Most small furniture brands can't fund that, so they make trade-offs. They optimise one ad set for purchases with a smaller budget and accept Learning limited, judge over two to four weeks rather than one, and test fewer concepts with more spend each, around $120 to $180 per concept.
For a store in India, the same method works in rupees. Say your average order is ₹1,800, your gross margin is 60% and shipping, payment fees and COD returns average ₹260 an order. Break-even CPA = 1,800 × 0.60 − 260 = ₹820. With a target CPA of ₹600, one ad set needs about 600 × 50 ÷ 7 ≈ ₹4,300 a day to reach 50 purchases a week, and each test concept needs ₹1,200 to ₹1,800 of spend before you judge it. For COD-heavy stores, count only delivered orders when you calculate CPA. Our guide to Meta ads cost in India has local benchmarks to sanity-check against.
What if your budget is small?
Most stores start below the ideal, and that is fine. A small budget doesn't stop ads working, it just slows down how quickly you can be sure.
Rules for small budgets:
- One campaign, one ad set. Every extra ad set divides your purchases. With $30 a day, a single broad ad set learns far more than three $10 ones. The CBO vs ABO guide has the formula: ad sets you can feed = (daily budget × 7) ÷ (target CPA × 50).
- Broad audience. Narrow audiences cost more per impression and leave less room to learn. See broad vs interest vs lookalike targeting.
- Fewer, bigger tests. Test two concepts a week properly rather than six that each get $10.
- Longer judging windows. Judge on 14 days rather than 7, and on cumulative spend rather than daily results.
- Accept Learning limited. It is a warning that results will be noisier, not a sign your ads have failed.
Is there a minimum budget for Facebook ads?
Meta sets a technical minimum daily budget that depends on your currency and what you optimise for, and Ads Manager warns you if you go below it. It is usually small, often a dollar or a few dollars a day. Clearing it means the ad set can run, not that it can learn. For a purchase-optimised campaign, a daily budget below your target CPA means you can't expect even one purchase a day, so results will be very slow to read.
How fast should you increase your budget?
Grow in steps, once cost per purchase has held at or below target for several days:
- Steady steps: raise the budget about 20 to 25% every two to three days while CPA holds.
- Hold when it slips: if CPA rises above target for a few days, stop raising and let it settle.
- Add, don't only raise: once one ad set is large, add a new ad set or campaign with fresh creative rather than pushing one ad set ever higher.
Meta's help page on significant edits notes that small budget changes are unlikely to reset learning, while large jumps may. Our guide to scaling Meta ads without killing ROAS goes further.
How do you know if you are spending too much or too little?
Check two numbers every week:
- Cost per purchase against target, from Ads Manager, by campaign.
- Blended efficiency, total store revenue ÷ total ad spend, your MER, from your store.
| What you see | What it suggests |
|---|---|
| CPA below target, MER steady or rising | Room to spend more. Raise in steps. |
| CPA at target, MER steady | About right. Keep feeding new creative. |
| CPA above target for a week or more | Spending past what current creative supports. Hold or cut, and test new angles. |
| Ads Manager CPA fine, MER falling | Ads may be taking credit for sales that would happen anyway. Check spend on existing customers. |
Common budgeting mistakes
- Choosing a round number first. Start from margin, not from what feels comfortable.
- Ignoring variable costs. Shipping, fees and returns quietly shrink your break-even CPA.
- Spreading a small budget thin. Too many ad sets, none of which learns.
- No separate testing money. Without it you never find the next winner.
- Judging daily. Daily results swing. Judge weekly, on spend thresholds.
- Doubling overnight. Big jumps can reset learning and spike costs.
Let rules hold the line
Once your target CPA and caps are set, the hard part is sticking to them day after day. Fly ROAS Autopilot works inside the monthly cap and daily limit you set, gives every test a 48-hour verdict and, by default, stops tests that spend 2× your target CPA without a sale (adjustable from 1.5× to 3×). You can also require approval before anything launches.
Budget from the bottom up
The right Facebook ads budget is the one your margin can support and your data can learn from. Find your break-even CPA, set a target below it, fund up to 50 purchases a week per ad set, keep a separate testing budget of two to three times target CPA per concept and grow in steps while cost per purchase holds. Write the numbers down before you start, and let them, not daily swings, decide what you do next.
Frequently asked questions
How much should a small business spend on Facebook ads?
Start from your margin. Work out the most you can pay for an order, set a target below it, then multiply by 50 and divide by 7 for an ideal daily budget per ad set. If that is too much, spend what you can afford in one ad set and judge results over two weeks rather than a few days.
Is $20 a day enough for Facebook ads?
It can be, if your target CPA is low and you keep everything in one ad set. At a $20 target CPA it buys about one purchase a day, which is enough to find buyers slowly but not to exit the learning phase. With a higher CPA, $20 a day makes results very slow to read.
How much should I spend to test a Facebook ad?
Plan on two to three times your target CPA per concept before judging it. At a $25 target CPA, that is $50 to $75 per concept. An ad that has spent that much with no purchase is unlikely to become a winner, while one with sales at or below target deserves more budget.
Can Meta spend more than my daily budget?
Yes, on some days. Meta's help page on daily budgets says it may spend up to 75% more than your daily budget on a day with better opportunities, but won't spend more than seven times your daily budget across a calendar week. Plan around the weekly total.
How do I calculate a monthly Facebook ads budget?
Either add your daily scaling and testing budgets and multiply by 30, or divide your monthly revenue goal from ads by your target ROAS. Target ROAS is your average order value divided by your target CPA. The Meta ads budget calculator does both.
