UGC questions › Testing ads on Meta
Short answer
To optimize UGC for Meta retargeting ads, build each ad for what the warm viewer already knows. Skip the introduction they've already seen, open on the doubt that stopped them (price, proof, fit or timing), keep the video short, and give it a direct call to action. Split warm people by how close they came to buying, exclude buyers, and plan on refreshing creative sooner than you would for cold traffic, because retargeting audiences are small.
A retargeting ad talks to someone who already knows who you are, so it can skip most of what a cold ad has to do. A cold UGC ad spends its first seconds naming a problem and its middle explaining the product. A warm viewer has usually seen that already, either in your ad or on your site, and didn't buy. Showing them the same explanation again mostly adds frequency.
The job changes from "here's what this is" to "here's why the thing that stopped you isn't a problem." That plays out differently across the three parts of the ad:
Our guide to the hook, body and CTA of a UGC ad covers how to swap each part without rebuilding the ad. That modular structure is what makes retargeting versions cheap. One note from our own testing: when we cut a cold ad into a warm one, we rewrite the hook and the body and leave the CTA alone. A warm viewer needs a new angle and a new argument, not just a new opening line over the same footage, and the CTA is the part that rarely needs to change. The exception runs the other way: when a specific hook and body pulled in exactly the type of person you want more of, keep both and refilm them rather than rewriting.
Match the ad to how far the person got. Meta lets you build custom audiences from your website (through the Meta Pixel), your app, your customer list, and engagement on Facebook and Instagram, including video views, lead forms and your Instagram account (Meta Business Help Center). Each of those groups needs something different.
| Warm audience | What they already know | What's usually stopping them | UGC that fits |
|---|---|---|---|
| Watched at least 15 seconds of a cold UGC ad | The problem and the product, from one creator | Not sure it works, or not ready yet | A second creator with a different angle, a demo, or a FAQ-style video |
| Engaged with your Instagram or Facebook | Your brand, loosely | No reason to act now | A testimonial with a specific result, or a product-in-use clip |
| Visited your site or a product page | Price and details | Price, fit or trust | An objection ad: "I thought it was expensive too" or "here's how sizing runs" |
| Added to cart or started checkout | Nearly everything | Shipping, returns, a last doubt | A short reassurance clip or a static with the offer |
| Opened a lead form but didn't submit | The offer | What happens after they hand over details | A "what the call is actually like" video, often from the founder |
| Existing customers | Your product, first-hand | Usually nothing, so exclude them or cross-sell | A "how I use it now" or "what to try next" video |
Meta's own retargeting page suggests similar plays: dynamic ads that show people the items they're most interested in, and showcasing new products to your best customers (Meta for Business). For service businesses and coaches, the lead-form and booking stages are where a founder video often does more than a creator, which founder-led ads vs UGC covers.
You don't need every row. Most accounts under a few thousand dollars a month in spend get enough from two warm groups: people who engaged (video viewers plus social engagers) and people who visited the site without converting.
Shorter, faster to the point, and with one objection per ad. Here's what we change when we cut a cold UGC ad into a warm one:
FAQ-style ads are worth making for this stage in particular. In our study of 8,107 Meta ads, 94% of video ads that opened with an education or FAQ-style hook reached 90 days, against 59% for contrarian or warning hooks. A warm viewer with a question is exactly who that format is built for.
It depends on what the cold ad did. If the first creator built real trust (people commented, asked questions, watched to the end), a follow-up from the same person can feel like the next part of a conversation. If the cold ad got a quick view and no response, a different face with a different angle gives the viewer a second, independent reason to believe you.
A simple rule: use the same creator when the follow-up is a direct sequel ("a few of you asked about..."), and use a new one when the follow-up is more proof. Two different people saying the product worked is stronger evidence than one person saying it twice.
Running the retargeting ad from the creator's own handle is another option. Meta's partnership ads can run from a creator's account, and account-level partnership permissions let you include or exclude the creator's own custom audience. How Spark Ads, partnership ads and whitelisting work explains the setup and when it's worth it.
Use video engagement custom audiences, which turn the people who watched your cold UGC ads into a warm pool. In Ads Manager, go to Audiences, create a custom audience, choose Video, then pick an engagement level such as "People who viewed at least 10 seconds of your video." You can select up to 200 videos from a Facebook Page, an Instagram professional account, a campaign or by video ID (Meta Business Help Center, checked November 2, 2026).
In practice we set the threshold at 15 seconds or more, which matches Meta's ThruPlay definition and filters out people who barely registered the ad. We also build engagement audiences that combine the five interaction types Meta offers (profile visits, ad or post engagement, messages, saves and follows) separately for Facebook and Instagram, and we refresh video-view audiences about once a month so recent viewers stay in the pool.
One detail catches people out. Meta says that if you uploaded the video file when creating the ad, the audience size reflects the ad's reach. If you built the ad from an existing post, the audience may instead reflect the views on that post. Meta's example is an audience showing 1,000 views for a campaign that recorded more than 200,000. Check the audience size before you build a campaign on it, and if it's far smaller than the ad's reach, that's the likely reason.
There's a mechanical version of "don't show the same video twice," too. Meta lets you build a custom audience of people who viewed at least 3 seconds of an individual video. Build one for each video in your retargeting ad set and add it as an exclusion on the ad set carrying that video. As soon as someone watches three seconds of a video, they stop seeing that specific video and roll onto the next one in the rotation.
The other windows to know:
Shorter windows mean warmer, smaller audiences. A 7-day add-to-cart audience is about as warm as it gets but may only hold a few hundred people. A 90-day video viewer audience is cooler and much bigger. Pick the window that gives you an audience big enough to spend your retargeting budget without hitting the same people every day.
Often you don't, and this is where most retargeting guides are out of date. In sales, app and leads campaigns, Meta now turns on Advantage+ audience by default, and a custom audience you add is treated as a suggestion. Meta says it will show ads to people matching your suggestion "and other audiences when it's likely to improve performance" (Meta Business Help Center). In practice, a broad campaign already finds a lot of your warm people, because they're the ones most likely to buy.
Check your own account before relying on that default. In the accounts we manage, Advantage+ audience is switched off at the ad set level, which means any custom audience added there is a hard control rather than a suggestion. If yours is on, the suggestion behavior described above applies, and the decision rule below is how you tell which situation you're in.
You can see how much it's doing. Set up audience segments in your ad account settings (engaged audience and existing customers), then use the Audience segments breakdown in Ads Manager to split results into new audience, engaged audience and existing customers (Meta Business Help Center). Meta notes the feature may not be available to every account yet.
A decision rule:
For how to structure the campaigns around this, see CBO vs ABO vs Advantage+.
Faster than in cold campaigns, because the pool of people is small and fixed for the length of your window. Say 6,000 people visited your site in the last 30 days and you spend $30 a day at a $25 CPM. That buys 1,200 impressions a day, or 8,400 a week, which works out to 1.4 per person per week if Meta reached every one of them. It won't reach all of them, so the people it does reach will see your ads more often than that.
Meta's analysts found that by the fourth exposure to the same creative, the likelihood of a click or conversion was about 45% lower, and that direct-response ads don't "wear in" first (Analytics at Meta, May 2023). Meta's help center also says it counts exposures of the same image or video across your other campaigns (Meta Business Help Center), so a warm viewer who saw your cold ad five times is already tired of that video before your retargeting starts.
One caveat on frequency itself: if the campaign is profitable, a high frequency number alone isn't the problem. The 45% decline is about repeated exposure to the same creative, which is why rotating creatives beats shrinking the window.
Three habits help:
Judge it on cost per purchase, booking or qualified lead, and be suspicious of numbers that look too good. Retargeting audiences are full of people who were already likely to buy, so a retargeting ad often gets credit for sales that would have happened anyway. Low cost per result in retargeting is normal and doesn't prove the ads caused anything.
The cleanest check is a holdout test, where part of the warm audience sees no retargeting ads and you compare how many from each group buy. Meta offers this as a lift test in the Experiments section of Ads Manager. We haven't run a lift test on our own account yet, so the version we actually use is the rough one: pause retargeting for two weeks and see whether total sales move. If they don't, that budget is probably better spent on new cold creative that fills the warm pool in the first place.
Within the campaign, compare ads against each other on cost per result and on how many people watch past the first few seconds. A warm ad that people skip is answering a question they don't have.
For a lot of small accounts, retargeting is the least important part of the plan, and spending heavily on retargeting creative is the wrong call. If your site gets a few hundred visitors a month, the warm audience is too small to spend much on, and Meta's broad campaigns will already be reaching many of those people. Most of what makes retargeting work happens upstream: good cold UGC creates more warm viewers, and a warm audience of 500 people can only produce so many sales.
Our own account is the live version of this. It spends roughly $30K a month on a single leads campaign, all broad targeting, no separate retargeting campaign anywhere, and monthly frequency sits under 3. Most of the spend lands on Instagram Reels and Stories. The broad campaign reaches warm people on its own because they're the cheapest conversions for it to find, which is exactly why a dedicated retargeting campaign would mostly be paying twice for the same person.
We'd put retargeting creative ahead of new cold creative only when:
If those don't apply, put the money into cold creative and add two or three warm-angle ads to the same campaign. Ad Variations (starting at $3K) gives you 20 ads from 5 hooks, 2 bodies and 2 CTAs to test.
On Google
Add us as a preferred source and Google shows our UGC ad research more often in Top Stories and AI answers.
Book a 30-minute call. We review your current ads and your offer and see if we're a fit. No retainer required.
Book a Call