“We paid $300 a video and nine out of ten never beat our control.”
UGC ads are not converting for you because most ads do not convert for anyone, and a flat fee made you pay full price for videos that exist rather than videos that sell. One operator who runs thousands of creator ads a month says Meta’s own number is that one ad in 25 becomes a winner. The fix has two halves: test a batch big enough that a winner is likely, and move the pay from the video to the result, a share of the ad spend behind it or of the sales it brings. That turns nine losers into nine samples and one winner into a creator you are glad to pay.
Why are my UGC ads not converting?
Three mechanics, in the order they usually bite.
The batch is too small. When a brand says it “tested UGC” it usually means five videos. Lukas Pakter, Trybe’s founder, tells podcast hosts that “One out of 25 ads on average becomes a winner” and that at his own brand Haus you make the first 300 creatives to learn which styles work (Foundr #719, 22 October 2026; his numbers about his own brand are his to claim). A Sideshift reply to a brand that tested five creators put it in four words: “5 creators is a coin flip” (Squilly on X). Five videos tell you nothing about the channel. They tell you about five videos.
The pay model priced every video the same. A flat fee is set before anyone knows what will sell, so it is priced for the average: enough to get a creator to film, whatever happens next. Pakter calls paying “$500+ for a single video that likely won’t produce results” the single biggest waste in creator marketing (X, 13 August 2026). A brand founder did the arithmetic in public after six weeks on commission: 696 videos, and at a flat rate “If I paid your model of $166 a video, it would be $115K.” (kasramoradi on X). The median advertised per-video offer across 670 briefs was $100, with half between $50 and $200 (Pitchlo, December 2025 to November 2026). At $100 a video, 25 videos is $2,500 before you know if the channel works.
The winner wore out and nothing replaced it. When an ad’s cost per result climbs to twice what the account’s earlier ads achieved, Meta labels it creative fatigue in the Delivery column (Meta). A brand with one winner and no pipeline watches the winner fade and concludes UGC stopped working. The format did not stop; the supply did. That problem has its own page: creative fatigue on Meta ads.
Is it the video or the offer?
Before you change the pay model, rule out the offer. Run the check that the Wispr Flow growth lead and every media buyer runs: same product, same offer, same audience, creator video against your control. If the control wins on cost per result, the video lost. If both lose, the offer lost, and no creator will fix that.
Three things to read on a losing video, in order: the hook rate (did anyone stay past three seconds), the click rate (did the promise land), and the conversion rate on the landing page (did the page keep the promise). A video with a good hook and bad conversion is a page problem. A video nobody watched past three seconds is a video problem, and a cheap one to fix with the next batch.
The fix: pay for the result, not the file
The method any brand can run, with or without a platform.
- Set the batch before the price. Decide how many videos you need for a fair test. Twenty-five is the working number from the one-in-25 quote above. Then back into a unit cost you can afford 25 times.
- Move most of the pay to the result. Pay creators a share of the Meta ad spend behind their video, so the losers cost you samples and the winner costs you a slice of media you were buying anyway. Across $500 million in creator payouts, Lumanu found brands typically reimburse creators “between 1% - 7% of the amount of ad spend” they put through creator accounts, and brands that paired a flat fee with an ad spend share paid about 20 % less for content than flat fee alone (Lumanu, 2024).
- Keep a small base only if you need applicants. Commission with no base gets fewer takers. If your brief goes unanswered, add a per-video base below the $100 median and keep the share. The full argument is in creators won’t film for commission only.
- Cap the share per video. The cap is the most you would ever scale one video to, times the share. It lets you say yes to a 10 % share without fear.
- Kill in a week, scale the winner, refill the batch. A video that has not beaten the control after its first budget is done. The money goes to the winner and to the next batch.
Example. A skincare brand needs 25 videos for a fair test.
| Pay model | Before any result | After one winner scaled to $5,000 of spend | Total |
|---|---|---|---|
| Flat $300 a video | $7,500 | $0 more | $7,500 |
| $60 base plus 8 % of ad spend | $1,500 | $400 to the winner’s creator | $1,900 |
| No base, 10 % of ad spend | $0 plus samples | $500 to the winner’s creator | $500 plus samples |
Two mid-performers that each get $800 of spend add $64 each on the hybrid line and $80 each on the commission line. The 22 videos that never run cost $60 each on the hybrid line and a sample on the commission line. The flat line paid $300 for every one of them. Example numbers, not a benchmark.
This is why the creator ad that has run longest in Trybe’s own Meta library is a creator telling brands “don’t think you should pay me just because I made you a video” (Meta Ad Library, launched 26 August 2026). The incentive runs the other way too: when the creator is paid while the ad spends, the creator works to make the ad that spends.
How HourlyUGC does it
On HourlyUGC, a campaign can carry no base pay at all. The creator earns a share of the Meta ad spend behind their video, from 1 to 30 %, or a share of the sales their video brings, from 1 to 50 % of each order or a fixed $1 to $500 per order. You can run both on the same video, because they count different results.
- The sale is counted before it is paid. Attribution is by Meta, or by HourlyUGC through a connected Shopify store or the Hourly Pixel, last touch within 7 days, with creator links and codes. You choose which count pays. The method is in which creator drove sales.
- A base is optional, not required. If applicants are thin, add a per-video base (you pay for each approved video) or an hourly base (you book hours and each session ends in videos), and keep the commission on top, in the same campaign.
- The winner goes live from the same place. Approved videos launch as Meta Partnership Ads from inside HourlyUGC, so the spend that pays the creator is the spend on the ad you built from their video. Details in Meta partnership ads for UGC.
- Samples are a line in the campaign: you ship the product to the creator.
With commission, the creator gets paid when you run their video as an ad and it spends or sells. You only pay for what works, and creators work hard to make the video that wins.
Proof
Pup Labs, Growth Team: “It broke our previous performance ceiling. Same product, same offer, same audience. The only thing that changed was creative volume.” Kitsch runs its creator program on HourlyUGC; the workflow is in the Kitsch case study.
Next step
Start a campaign on HourlyUGC with a share of ad spend and a cap per video. Add a small per-video base if the first week brings too few applicants. The long version of the model is on the pillar, commission-based UGC.