Set up a creative test in Meta Ads Manager: the budget, the limits, and what the results mean

A creative test in Meta Ads Manager copies one of your ads into 2 to 7 test ads inside a campaign that is already running, and spends a share of that campaign's budget on them. Meta suggests "no more than 20% of your existing budget", and says it will "aim to spend that amount per day on test ads". It only works on campaigns using the Highest volume bid strategy. At the end you get a ranking of the test ads against a metric you chose, with no confidence level, and nothing is switched off for you.

All four of those sentences come from Meta's help page. People confuse the tool with two other things: an A/B test, and simply putting more ads in one ad set.

Set up a creative test in Meta Ads Manager: Meta's steps

Meta's page, Set up a creative test in Meta Ads Manager, opens with what the tool is for:

In a creative test, you can compare up to 7 creative variants with delivery provided to new test ads. The test is set up in an existing campaign so that high-performing ads can continue to run after the test with delivery system learnings retained.

The steps, condensed but in Meta's order:

1. Go to Ads Manager and select an existing campaign. "Creative testing is supported in campaigns using the Highest volume bid strategy only. To run a creative test, switch your bid strategy to Highest volume."
2. Go to the ad level and "scroll down to the Creative test section. Click Set up test."
3. Choose how many copies. "You can create 2 to 7 copies." Then: "You can update the creative for each test ad after you confirm your test setup."
4. Choose the budget share of the existing campaign or ad set budget to spend across the test ads.
5. Set the test duration and the comparison metric, then click Confirm.
6. Publish. Before you do, Edit test settings lets you change them.

Step 3 explains a phrase that confuses people. The test starts from copies of an ad you already have, but the copies are placeholders: you swap in the new creative afterwards. So the tool tests new images, videos or copy, not seven identical duplicates.

Step 1 has a cost Meta lists on a different page. If your campaign runs any other bid strategy (cost per result goal, ROAS goal, bid cap or highest value), you have to switch it to Highest volume, and Meta's significant edits page lists "Changing bid strategy" as a significant edit, one that sends the ad set back into the learning phase. Meta's bid strategies page describes Highest volume as a spend-based strategy that aims to "maximize delivery and conversions you can get from your budget".

How much budget does a Meta creative test use?

There is no separate test budget. You carve a share out of the one the campaign or ad set already has. The full sentence on Meta's page:

Choose the amount of your existing campaign or ad set budget you'd like to spend across test ads. We suggest using no more than 20% of your existing budget to help reduce the potential of impacting your overall campaign performance during the test. We'll aim to spend that amount per day on test ads. If your campaign or ad set doesn't include any other active ads aside from test ads, they may receive more of the daily budget during the test.

Three things follow from it:

• The 20% is a suggestion, and it protects the campaign, not the test. Meta's stated reason is to limit the impact on "overall campaign performance".
• It is a daily target. "Aim to spend" is an intention with no guaranteed floor.
• If the test ads are the only active ads, the cap loosens. They "may receive more of the daily budget".

This 20% is not the same number as the testing budget in Meta's creative diversification guidance, which recommends "dedicating 20-30% of your overall ad budget towards testing new ads". That one is advice about how much of your total spend to put toward new creative over time; the creative test's 20% is a per-day share inside one campaign. What Meta publishes about creative volume covers the diversification page.

How many ads can a creative test compare?

Two to seven test ads, made by copying an existing ad. The introduction says "up to 7 creative variants"; the steps say "2 to 7 copies". Your existing ads keep running beside them on the rest of the budget.

How long should a creative test run?

Meta's creative test page gives no minimum and no maximum. You set the duration yourself, and the page adds a warning about what that duration controls:

Your test ads will start running on their ad set schedule when you publish them, even if you adjust the test schedule. The test schedule only determines when results are measured.

So the test schedule is a measurement window. It does not start or stop the ads.

The duration numbers you will see quoted, a seven-day minimum and a 30-day maximum, come from Meta's Best practices for A/B Testing, which says "we recommend a minimum of 7-day tests" and that an A/B test schedule must be "between 1 and 30 days". That page is about A/B tests. Using seven days as a floor for a creative test is a reasonable borrowing, but it is a borrowing: Meta has not said it about this tool.

What do creative test results mean?

When the test ends, Meta says "you'll receive an email with results identifying high performers if any were found". You can also hover over the beaker icon next to the campaign, ad set or ads in the test and open the charts view, or visit Experiments.

What the result is, and is not:

The results identify top performing test ads against your comparison metric as well as additional metrics. A confidence level is not included.

And what happens next:

After the test, the test ads will continue to run in your existing campaign alongside the existing ads using the original campaign or ad set budget. The test does not make any automatic changes based on the results, so they will run until you turn them off or they reach the end date of the ad set, if applicable.

Read together, the result is a ranking, not a verdict. "If any were found" means a test can end with no high performer. With no confidence level, a small gap between two test ads may be noise. And the losers keep spending until you pause them.

How is a creative test different from an A/B test?

Meta's About A/B testing page defines the other tool: "We show each version to a segment of your audience and ensure nobody sees both, then determine which version performs best." The two answer different questions.

Sources for the table: Meta's creative test page, About Experiments, About A/B testing, the winner page, the A/B best practices page, the significant edits page and About managing ad volume.

Two lines from that table carry the decision. The A/B test splits the audience so nobody sees both versions, and comes with a confidence percentage that Meta calculates by simulating outcomes "tens of thousands of times". The creative test does neither. What it offers instead is that the winners stay in the campaign that is already working, with its learnings, and there is "no need to merge them into another campaign where the learnings would reset."

Use a creative test when the question is "which of these new creatives should join this campaign?" Use an A/B test when the question is "which of these strategies is better?" and you need an answer you could repeat.

Why not just add the new ads to the ad set?

Because Meta's About managing ad volume page says that when an advertiser runs too many ads at once, "each ad delivers less often". Meta's delivery pages suggest the new ads lose most of that. They compete with the proven one for the same budget, and Meta's delivery system sends spend to the ad it predicts will do best, which is usually the one it already knows. Why new Meta ads get no spend quotes the pages behind that loop.

Meta's About managing ad volume page goes further. It says "too many ads can result in worse performance" and advises "Decrease ads per ad set, but maintain diverse creative assets per ad set." For testing several creative variants, its suggestion is dynamic creative, which "lets our delivery system find the best creative to deliver for a given user – all while using just one ad." (Dynamic creative and the flexible format covers what is still available.)

The creative test is the exception to that advice: it is the one setup Meta's help pages describe that gives new ads "delivery provided" inside a running campaign.

Meta lists "Adding a new ad to your ad set" as a significant edit, and its creative test page doesn't say whether publishing test ads counts as one. Check the ad set's delivery status after you publish rather than assuming either way.

Before you run a creative test

1. Check the bid strategy first. Highest volume only. Switching to it is a significant edit, so do it before a quiet week, not mid-push.
2. Start at Meta's 20% of the existing budget, and remember it is a daily aim, not a guarantee.
3. Swap real creative into every copy. The tool makes copies of one ad; the test is only as good as how different you make them.
4. Pick the metric before you publish, and set a measurement window long enough for that metric to collect results. Meta gives no minimum here; its A/B guidance uses seven days.
5. Read the ranking, then act. No confidence level, no automatic changes. Pause the losers yourself, or they keep spending.
6. Need a statistically separated answer? Run an A/B test instead.

Sources

• Set up a creative test in Meta Ads Manager — Meta Business Help Center (accessed October 24, 2026)
• About A/B testing — Meta Business Help Center (accessed October 24, 2026)
• About Experiments — Meta Business Help Center (accessed October 24, 2026)
• How winning campaigns are determined in A/B tests without a holdout — Meta Business Help Center (accessed October 24, 2026)
• Best practices for A/B Testing — Meta Business Help Center (accessed October 24, 2026)
• Significant edits and learning phase — Meta Business Help Center (accessed October 24, 2026)
• About bid strategies — Meta Business Help Center (accessed October 24, 2026)
• About managing ad volume — Meta Business Help Center (accessed October 24, 2026)
• Three steps to optimize your performance with creative diversification — Meta for Business, 19 November 2024 (accessed September 19, 2026)

---

The creative test page states no duration limits, so none are given here as Meta's. Meta revises these pages without notice, and tool availability varies by account, so check the source before planning a test around it.

Terrads makes the different creatives a test needs, each one a new concept rather than a recrop of the last.