How Much Should You Budget to Test Facebook and Instagram Ads?
No single amount is right for everyone. Work out your test budget from the cost per result you can accept and Meta's learning phase.
· Strategy
A Facebook and Instagram ad test budget is calculated, not guessed: multiply the cost per result you can accept by about 50 results, the volume Meta says an ad set needs in a week to exit the learning phase, then divide by seven to get a daily budget. If that number is more than you can spend, change the result you optimize for instead of spreading a small budget thin.
This article walks through the method, shows how to estimate a cost per result without relying on an average found online, and gives a worked example with clearly stated assumptions. If you have never run a campaign, start with our guide Meta ads for beginners: your first campaign, then come back here to size the budget.
Why is there no "right" amount for everyone?
Because the cost of a result depends on what you ask Meta to find, your audience, your offer, your creative and the competition at the time your ads run. A link click, a message, a lead and a purchase do not cost the same. The averages published by agencies and tools mix different industries, regions and objectives: they do not tell you what your result will cost. Your budget should start from your own numbers, even rough ones, and be corrected after the first test.
What does Meta say about the learning phase?
Meta explains that an ad set goes through a learning phase, during which delivery is less stable. It usually exits after about 50 optimization events within the week following the last significant edit. If the budget cannot reach that volume, the ad set stays learning limited.
Two practical consequences for a test:
- One well-funded ad set beats several ad sets sharing a small budget: none of them would get enough data.
- Do not touch anything during the test. A significant edit restarts learning and makes the first days useless for comparison.
The "result" here is the one tied to the objective you choose. Meta offers six objectives, described in its help page on choosing an objective, and Ads Manager then shows a cost per result. That number is the base of the calculation.
How do you estimate your cost per result without a made-up average?
Three sources, from most to least reliable:
- Your own past campaigns. Even an old boosted post shows a cost per result in Ads Manager. It is the best base you have, because it reflects your audience and your creative.
- The highest cost you can accept. Start from your unit economics: for a sale, take the profit an order brings in and decide how much of it you are willing to spend to acquire the customer. That ceiling does not tell you what ads will cost; it tells you when they stop paying off.
- A one-week exploratory test, with a budget you are prepared to lose, just to get a first real cost per result. It will not give you a solid conclusion, but it replaces a guess with a measurement.
To read these metrics correctly (CPA, cost per result, ROAS), see CPM, CPC, CTR, CPA and ROAS explained.
The calculation, step by step
- Pick the result you want (purchase, lead, message, link click).
- Estimate its cost using one of the three sources above.
- Multiply by 50 to get a weekly budget that can, in theory, exit learning.
- Divide by 7 to get the ad set's daily budget. Meta notes in its help page on budgets that daily spend can vary from one day to the next but evens out over the week.
- Plan for at least one to two weeks of delivery without significant edits.
- Check that the amount stays above the minimum shown in Ads Manager, which depends on your currency and optimization.
Worked example, with stated assumptions
Lantern & Wick, a fictional candle brand that sells online, wants to test ads. The numbers below are example assumptions, not market averages.
| Item | Example assumption | Where it comes from |
|---|---|---|
| Profit per order | $22 | The brand's own books |
| Highest acceptable cost per purchase | $16 | The brand's decision (it counts on repeat orders) |
| Budget for 50 purchases | 50 × $16 = $800 a week | Calculation |
| Matching daily budget | about $114 a day | Calculation |
$114 a day is more than Lantern & Wick wants to commit to a first test. It has two honest options:
- Optimize for a more frequent result, such as add-to-carts or product page views. If its exploratory test shows $3 per add-to-cart, 50 add-to-carts cost $150 a week, or about $21 a day. It learns which creative and messages pull people in before moving to purchase optimization.
- Accept a test that will not exit learning, knowingly. It will show a trend, not a conclusion. That is fine for a very first try, as long as you do not turn it into a hard rule.
How should you set up the test?
- One campaign, one ad set, with two or three different creatives inside it.
- Automatic placements, with creative designed for Feed and for vertical formats. Our guidelines are in ad creatives: formats and safe zones.
- Your best organic posts as a starting point. A post that already got your followers to react is a good candidate. The difference between boosting that post and building a real ad is covered in boosting a post vs. creating a real ad.
- An end date set in advance, with a decision planned for that date.
Should you split Facebook and Instagram in the test?
Not for a first test. With automatic placements, which Meta recommends in its page on Advantage+ placements, delivery goes wherever the system finds the best results. Splitting Facebook and Instagram into two ad sets would cut your budget in half, and neither ad set would get enough results to exit learning. Once the test is over, you can check the breakdown by platform in your reports and then decide whether a placement deserves its own ad set.
What should you decide when the test ends?
| What you see | Decision |
|---|---|
| Cost per result below your ceiling, ad set out of learning | Keep going, then raise the budget in moderate steps |
| Cost per result close to your ceiling | Test new creative or a new offer before raising the budget |
| Cost per result well above your ceiling | Stop, then rework the offer, the landing page or the result you optimize for |
| Too few results to conclude | Optimize for a more frequent result or extend the test, without changing anything else |
Frequently asked questions
What is the minimum budget for a Facebook or Instagram ad?
Meta applies minimums that depend on your currency and optimization, shown in Ads Manager. For a useful test, the right budget is one that can get about 50 results a week, the volume Meta says an ad set needs to exit the learning phase.
How long should an ad test run?
At least one to two weeks without significant edits, since each significant edit restarts the learning phase. Set the end date and the decision you will make before you launch.
How can I know my cost per result before launching?
Look at past campaigns or boosted posts in Ads Manager, work out the highest cost your margin allows, or run a short exploratory test to get a first real number.
What if the calculated budget is too high?
Optimize for a more frequent result, such as add-to-carts or page views, or accept a test that will not exit learning and read it as a trend rather than a conclusion.
Prepare your tests with Oposto
Oposto does not buy ads for you: you stay in control of your Meta campaigns. It helps before and after. It measures the real performance of your posts so you can spot the ones worth putting budget behind, and the Pro plan includes ad advice and analysis. It also prepares visuals and short videos from your brand's content DNA. Start on the free plan, no credit card required, at oposto.app.