AI UGC ads
Do AI UGC ads work?
Almost as well as real creators. At the top level, we’ve seen AI ads perform about 90% as well. The difference is how many you can test: with AI we can test 100 concepts a week if you need it, so you find your winners faster and for a lot less.
Do they look fake?
Bad ones do. Everyone in these is AI. Tap one to watch it with sound.
Where real creators still win
When everything’s perfect, a real creator still wins. But that costs a lot and takes a long time, so you only get a few tries a month.
That’s why some of our clients find their winners with us, then have creators reshoot them.
Why the number of ads matters on Meta
Meta picks which ad each person sees. It shows them “the ad that’s most likely to achieve the lowest cost per optimization event,” so your ads have a big say in what you pay.
Meta says it plainly: “More diverse creative options means more opportunities for our AI to explore and capture opportunities for advertisers.” It also warns that “too many ads can result in worse performance,” because each one gets shown less. So what works is a steady supply of new ads that are actually different.
That’s where AI wins. Every creator video means a shoot and a wait. We can make you 50 a week.
How to test them without wasting money
Put our ads in their own campaign with a cost per result goal, like $150 a lead. Meta says it aims “to keep your weekly average cost near your goal,” and it “may spend less than your full budget” to stay there. So if our ads can’t hit your number, they don’t get the money.
If our ads don’t perform, Meta won’t spend on them, and we don’t get paid. How we get paid.
Why not just use an AI tool yourself?
You can, it’s what we use too. But great ads take a lot of time and testing, and we’ve already spent thousands of dollars figuring it out. You just get the finished ads. What each option costs.
Meta’s pages quoted here: About ad delivery · Demystifying creative diversification · About managing ad volume · About cost per result goal.