Reflectly · B2C · AI journaling app
Pay small creators in each country to read a script, then run it as Facebook ads
Drove ~$3-4M ARR, but ad spend matched revenue
What they did
What happened
- Stage
- scale
In their words
Ryan Thorpe - Meet The Guy Dominating The App Store
Ryan ThorpeYes. So, Reflectly was the original app that was built back in 2017. And then at the time it was the AI journaling app before AI was a thing. It was kind of just a combination of machine learning. And then from 2017 to 2020, uh we scaled that ride- relatively large, um mainly we had just a really good arbitrage with Facebook Ads.
Ryan ThorpeUm we were doing micro-influencers at the time. So, finding UGC before it was a hot topic uh is one way of doing it. Um so we had influencers, UGC content creators in local countries, speaking like the local dialect and getting them to convert really well on Facebook Ads.
Joseph ChoiInteresting. So you'd find micro-influencers, but you wouldn't have them actually post their content organically. You would just pay them to read a script and then you'd use that as an ad?
Joseph ChoiA- And you said it got to like 2, 3 million ARR, and then it was break-even, but were you profitable leading up to that, like before you got to that like bigger revenue number?
Ryan ThorpeS- Yeah, so we were profitable um before we started fully scaling. And then through 2018 to 2020, Darius, who's the CMO at the time, he really uh drove the I guess the the UGC playbook in the really early days when Facebook was kind of like TikTok where there was such an arbitrage of, no one's making these odd-looking, not so great UGC videos. Um, but they were converting like crazy, because all the other big companies and advertisers were doing polished graphic ads at the time.
Ryan ThorpeThen that became the strategy that kind of really rose Reflectly to the fame that it that it got to uh in in the industry.
Get tactics like this every Monday
New tactics from the week's founder interviews, each linked to where it was said.
More from this episode
- Buy under-monetized apps and fix their paywalls instead of spending more on adsFirst deal paid back in 12-18 months; 30-35 apps bought over 5 years
- Run your own creator team: train them, brief them, and automate hiring through payment~250 regular creators, thousands of videos; drove tens of millions of downloads
- Post every video on your own TikTok first; put ad money behind the ones that take offCheaper testing off-peak; breakouts pushed four or five ways
- Test each video with $10-20, move winners to bigger campaigns, let the best 1-2 take 90%1-2 evergreen videos took 90% of spend; run by one media buyer
- Once a video holds a few hundred dollars a day of ads, have creators remake itSometimes compounded; the top viral video failed to copy across 20-40 creators
- Spend about 75% of the year's ad budget in January, when installs cost halfDone hit #1 free app in the US on Jan 2-3, two years running
- Test ads Dec 26-28, raise spend Dec 28-30, and 10x it on January's first SundayGuest's playbook from his ad accounts; no separate numbers
- Put a famous person's voice behind the video as the hookOne video hit 50-60M views, then a cease and desist; AI-voice copies partly worked
- Give creators strict briefs: 10-15 second videos, 25-27 cuts, POV phone shots, trending soundsSet a quality floor across ~250 creators; results still varied
- Recruit creators with 2K-20K followers by DM, send them to a creator page, then test them2,000+ creators on the books; ~10% became regulars, 1-2% of those went viral
- Pay creators $30-40 a video, raise to $100 for reliable ones, and add bonuses for hitsKept creators loyal long-term vs pure performance pay
- Hire agencies to produce creator contentDidn't work across many agencies; replaced with an in-house team
- Sell the app to employers, using stress tests to prove it lowers staff stress
Tags: meta, facebook, creative, micro-influencers, scripted-ugc, local-language, lo-fi