Short answer
Making UGC ads that convert comes down to four decisions, taken in order: who goes on camera, what the script says in its hook, body and CTA, which format and length prove your claim fastest, and how the batch is built so a winner teaches you something. The script and the hook move results the most. Format and length matter less than most people expect, and any claim the presenter makes about using the product has to be true.
A UGC ad is a paid ad made to look like something a regular person would post: someone talking to their phone, filmed vertically, with captions and a clear ask at the end. Most star paid creators rather than customers, which is normal and legal as long as the ad doesn't pretend otherwise. Our explainer on what UGC ads are and whether the people in them are real covers the definition.
The legal line shapes every decision after this one. Under the FTC's Endorsement Guides, a person who says they use the product "must have been a bona fide user of it at the time the endorsement was given." The working rule is that a creator who has used the product can speak from experience, and one who hasn't should be scripted as someone explaining the offer.
Choose the presenter by asking whether the buyer is choosing a person as much as a product. In our study of 8,107 Meta ads, every long-running ad that featured a person among the 24 local service, clinic, legal and mortgage advertisers used the owner, technician or doctor. Ecommerce looked different, with creator selfie testimonials and story hooks making up 30% of the long-running DTC ads. Founder-led ads vs UGC creators has the decision table and a way to combine both.
If creators are the answer, cast for believability with your buyer rather than followers, which barely matter when the ad runs from your own page. The useful vetting test is a real script read. How UGC agencies find and pick creators covers casting older, male and Spanish-speaking creators and when to ask for exclusivity.
AI avatars are the third option, and they suit a narrow set of jobs. Creatify's Pro plan works out to roughly 70 cents to a dollar a clip, against a median of about $175 for a raw creator video, which makes AI useful for testing angles, app demos and B-roll nobody can film. Of the 142 long-running video creatives we coded, 6 used AI visuals and none had an AI avatar presenter, and an avatar can't lawfully pose as a customer. AI UGC vs real creators has the labeling rules by platform and state.
The script decides whether the ad sells, so settle who writes it first. For organic posts, a creator riffing from bullet points is fine; for paid ads you plan to test, one writer should script every version so only the tested part changes. Who writes UGC scripts and how to brief them gives the nine things a brief needs and a worked 40-second example in six beats: hook, problem, turn, proof, offer and CTA.
Then split the ad into its three parts and judge each one separately. The hook stops the scroll in the first three seconds, the body makes the case, and the CTA asks for one action. In our study, hooks that answered a question reached 90 days 94% of the time and offer-led hooks 83%, against 22% for ads with no spoken hook. On our own account, better hooks took click-through from about 1% to over 3% on the same broad audience (what our own account taught us). The hook, body and CTA of a UGC ad has the full hook table and the common CTA mistakes.
Start from what your buyer doubts, then pick the format that proves it fastest. If the doubt is whether it works, show it working with a demo; if it's trust, put a person on camera; if it's "is this for me?", call the viewer out in the first line. In our study of long-running Meta ads, 46% of the long-running video creatives were a person talking to camera, while product demos were the steadiest after six months, with 67% still running at 180 days. The ten main UGC ad formats compares survival rates by format and shows where to find examples.
Length follows how much the buyer has to believe before acting. Use 15 to 30 seconds for a simple offer, 30 to 60 for a demo or story, and 60 seconds and up for expensive or unfamiliar offers that must answer objections. The median ad that ran 90+ days was 38 seconds, against 33 for ads switched off within 30 days, and by vertical the median ran from 24 seconds for local services to 70 for coaching and info. How long a UGC ad should be has each platform's own guidance.
Decide early whether every ad needs to be video. Statics are the cheapest way to test which angle earns clicks, and for DTC brands 64% of the ads running 180+ days in our study were statics. Video wins on Reels, where Meta's split tests found 9:16 video with sound cut cost per result by 34.5% against image ads, and for services that need trust. Static vs video ads lays out a test-angles-with-statics sequence.
Build a batch you can test one part at a time. A 5 hooks x 2 bodies x 2 CTAs batch is material for successive rounds rather than 20 ads judged at once: the 5 hooks first with one body and one CTA behind them, then the winning hook against the 2 bodies, then the 2 CTAs. That suits an offer that already converts; if nothing has worked yet, test different angles, presenters and formats instead, because twenty hook swaps on two bodies are still only two arguments. We build both kinds: Ad Variations (5 x 2 x 2, starting at $3K) and Fully Produced (20 unique angles, starting at $4K).
Mixing formats in one batch also gives Meta ads it can tell apart. In our own account, costs stabilized when we moved from a handful of ads to batches of 20 in different formats. Once the batch is ready, our guide to testing and scaling UGC ads on Meta covers the budget, the two test structures and when to cut an ad. You can also watch finished examples by industry in our ad samples.
New creative is the wrong fix when the offer is the problem,. In our study, the gap between the best and worst formats at 90 days was 72% against 43%, while the gap between the best and worst hooks was 94% against 22%. If people watch but don't click, look at the body and the offer; if they click but don't buy, look at the CTA and the landing page before you film anything new.
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