For brands
How many UGC videos should you test per ad set?
Updated · 6 min read · By Yap Factory
The short answer
Start with three to five clearly different videos in one ad set. TikTok’s own guidance is “between 3-5 different creatives per ad group”, and Google’s for Demand Gen is at least three images or videos in each shape. Meta gives no number, but warns that with too many ads each one delivers less often and fewer leave the learning phase.
For a first test that means three creators, one video each, with 3 openings on every video. On Yap Factory that is about £63 ($75) with three months of ad rights.
Buying UGC for the first time? See how a deal works, and when you pay
What the ad platforms say
| Platform | Its guidance | The hard limit |
|---|---|---|
| TikTok | Its guidance“We suggest between 3-5 different creatives per ad group and 3-5 diversified ad groups per campaign.” | The hard limit50 ads in an ad group |
| Meta | Its guidanceNo recommended number. “When an advertiser runs too many ads at once, each ad delivers less often.” | The hard limit50 ads in an ad set |
| Google (Demand Gen) | Its guidance“Follow the rule of three”: at least 3 images or videos in each aspect ratio in every ad group | The hard limitNot stated |
TikTok says the same thing in its account recommendations: “Your goal should be to have 3-5 unique creatives per ad group.” So three to five is the published answer on one platform, the floor on another, and consistent with the third. It is a sensible place to start on all of them.
Why not one, and why not twenty
One video tells you nothing. If it fails you can’t tell whether the idea, the creator or the first line was wrong. And most ads do fail: AppsFlyer’s 2025 study of 1.1 million video creatives found that the top 2% took 43% of ad spend in non-gaming apps and 53% in games. You are looking for the rare one that works, so you need several chances.
Twenty videos starve each other. Ad platforms need results before they can deliver an ad well. Meta says an ad set usually leaves its learning phase “after about 50 results in the week after the ad set’s last significant edit”. TikTok says performance settles “after about 25 campaign results or 7 days”. Every ad you add shares the same budget, so each one gathers results more slowly. Meta puts it plainly: “too many ads can result in worse performance.”
A quick check before you order: take what one result costs you (an install, a sign-up, a purchase) and multiply by 50. That is roughly what one Meta ad set needs to spend in a week to learn. If your weekly budget is about that, run one ad set with three to five videos, not three ad sets.
What counts as a different video
TikTok’s advice is that “it is always better to use creatives with big differences, especially when testing.” Five versions of the same video with the logo moved aren’t five tests. There are three levels of difference, from biggest to smallest.
| What changes | Example | What it tells you |
|---|---|---|
| The idea | ExampleA review, a demo and a day-in-the-life | What it tells youWhich angle on your product people respond to |
| The creator | ExampleThe same brief filmed by three people | What it tells youWhich face and voice your buyers trust |
| The opening | ExampleThree first lines on the same video | What it tells youWhich hook stops the scroll |
Openings are the cheapest test and often the most useful. Meta found with Nielsen that up to 47% of a video campaign’s value was delivered in the first three seconds. One video with three openings gives you three ads that differ exactly where it matters, for a few extra takes.
A first test, and what it costs
Three creators, one video each, three openings on every video. That is three ideas or voices and nine first lines. Upload the best opening from each creator as your three to five ads, and hold the rest for the next round.
| Test | Creators | Videos | Openings | On Yap Factory, with 3 months of ad rights |
|---|---|---|---|---|
| Smallest useful | Creators3 | Videos3 | Openings9 | On Yap Factory, with 3 months of ad rightsAbout £63 ($75) |
| TikTok’s upper suggestion | Creators5 | Videos5 | Openings15 | On Yap Factory, with 3 months of ad rightsAbout £105 ($125) |
| One creator, to try the process | Creators1 | Videos1 | Openings3 | On Yap Factory, with 3 months of ad rightsAbout £21 ($25) |
You need ad rights to run a creator’s video as a paid ad, so budget for them from the start; posting on your own channels is included. For what other platforms and agencies charge, see how much UGC costs. A one-page brief keeps the three videos comparable: same product, same must-says, different people.
Reading the results
- Give it a week. Meta’s advice for A/B tests is to “run tests for at least 7 days”, because shorter tests “may produce inconclusive results”.
- Don’t edit mid-test. On Meta, any change to an ad’s creative, or adding a new ad to the ad set, counts as a significant edit and restarts learning.
- Change one thing at a time. If two ads differ in creator and opening and offer, you won’t know which made the difference.
- Expect one winner. One clear winner out of three to five is a normal, good result. So is none: then the idea needs to change, not the opening.
- Judge on what you pay for. Cost per install or per purchase, not likes or views.
What to order next
A winning ad wears out. Meta defines creative fatigue as what happens “when an audience has seen the same creative too many times”, and flags an ad once its cost per result reaches twice what your past ads cost. TikTok’s advice is to refresh creatives “when delivery results exhibit a consistently declining trend”, and to add the new ones to the existing ad group, not start a new one.
So the rhythm is small and steady. Keep three to five ads live. When one fades, replace it: new openings on the winning idea first, then the winning idea with a new creator, then a new idea. A small budget needs a few new videos at a time, not dozens, which is why paying per video suits testing better than a large one-off batch.
On Yap Factory you can go back to the creator whose video won and start another deal with them, or publish the same campaign again and invite new people. How a deal works has the steps.
Questions
How many creatives should I have per ad set?
How many UGC videos do I need to start running ads?
Should I test different creators or different hooks?
How long should I run a creative test?
How often should I refresh UGC ads?
Do I need ad rights to test UGC as ads?
How Yap Factory works
Post a campaign
A one-page brief and a price range. It’s free and takes a few minutes.
Invite creators
A ranked shortlist of 5 to 8. Inviting is free and needs no card.
Pay when you start
We hold the payment until you approve the draft. 3 revision rounds are included.
Your first video from about £15 ($18), all in.
No subscription. You pay only for the creators you start, and only when you start.
Sources
Checked 18 November 2026. Prices change; follow the links for the latest.
- TikTok: creative best practices for performance ads
- TikTok: available recommendations
- TikTok: ad account limits
- TikTok: learning phase
- Meta: about the learning phase
- Meta: about managing ad volume
- Meta: campaign, ad set and ad limits per ad account
- Meta: significant edits and the learning phase
- Meta: best practices for A/B tests
- Meta: about creative fatigue recommendations
- Meta: updated features for video ads
- Google Ads Help: creative excellence guide for Demand Gen
- AppsFlyer: AI, emotion and creative trends (2025)