Scout

My app Pep AI has generated over $700,000 in revenue in less than 8 months. Halfway through that run, our #1 marketing channel basically got destroyed. Influencer marketing had driven a huge part of our growth, but creators in our niche started getting restricted on TikTok and Instagram, videos stopped getting distribution, and some creators were even losing their accounts. We had to find another way to keep scaling. That’s when we started taking Meta ads seriously. Here’s what I’ve learned and what can possibly work for your app too: 1. Start with proven winners If you already have influencer or organic videos that performed well, start there. Take your best performing videos and give them some initial spend. Don’t immediately pay for 100 brand new ads when you already have content that has proven people will stop and watch. Also make sure your funnel actually works before you scale. Meta won’t fix a bad onboarding flow, paywall, or offer. If people already aren’t converting, paid traffic can just make you lose money faster. 2. Optimize for the right event Optimize for the deepest conversion event you have enough data for. If you have a free trial, that may mean trial starts. If you have a hard paywall and enough purchase volume, optimize for purchases. I generally wouldn’t optimize only for installs if your actual goal is paying users. Meta will optimize toward people it predicts are more likely to complete the event you give it. If you ask for installs, it’ll find installers. If you give it a stronger downstream signal, it can optimize toward better customers. 3. Give Meta time, but don’t blindly burn money Meta takes time to learn. You probably won’t launch a campaign and instantly get an amazing cost per trial or purchase. Sometimes campaigns improve over several days or weeks as Meta gets more conversion data. But that doesn’t mean you should let terrible ads run forever hoping they magically recover. I learned this the expensive way. In August, Pep AI did around $170,000 in revenue and still wasn’t profitable. Yes, many of those subscribers may generate additional LTV through future renewals, but when you’re bootstrapping, you can’t rely on future renewals to fix bad cash flow today. We tightened things up, cut what wasn’t working, scaled our winners, and kept testing new creative. The next month we still did over six figures in revenue, but our ads were profitable. 4. Separate scaling from testing We keep proven ads separate from a lot of our experimental creative so the existing winners don’t immediately take all the spend. This gives new concepts enough delivery to actually learn something from them. At the same time, don’t overcomplicate your account with dozens of tiny campaigns and ad sets. Meta still needs enough conversion data to optimize. The goal is simple: give your winners room to scale while constantly searching for the next winner. 5. Track everything We use AppsFlyer and it has worked extremely well for us. You need to know which campaigns and creatives are driving installs, trials, purchases, and revenue. Before spending heavily, test your tracking end-to-end. Make a real trial or purchase and make sure the event is showing up correctly. Spending thousands while your conversion tracking is broken is an unbelievably expensive mistake. 6. How we find new ad concepts Study the companies that are better than you. Go through Meta Ad Library and look at apps in your niche that appear to be scaling. You can’t see their ROAS, so don’t assume every ad is profitable, but if a company keeps running the same concept or repeatedly makes variations of it, that can be a useful signal. Then look outside your niche. Find consumer apps that are growing quickly and study their ads too. Don’t copy the exact creative. Copy the psychology and framework behind it. 7. Make ads feel like content Some of the best-performing ads don’t immediately feel like ads. Don’t open with your logo and spend 30 seconds listing features. Teach something. Compare something. Tell a story. Show a problem. Then introduce your product as the solution. You’re trying to earn the viewer’s attention before asking them to care about your app. 8. Test concepts AND hooks There’s a big difference between a new concept and a new variation. A concept is the actual idea behind the ad. A variation might just be the same video with a different hook, creator, caption, or first few seconds. You need both. We often take one core concept and create 5 different hooks for the beginning of the video. So one testing batch might be: 5 concepts × 5 hooks = 25 creatives. The hook alone can completely change performance. But don’t only make variations of your winners. Keep testing completely different angles too. One person might respond to a comparison. Another might respond to a creator talking directly to camera. Another might respond to a screen recording or story. You want multiple ways to sell the same product. 9. Where we get the content made We currently use Trybe and SideShift. For faster experimental concepts, we may pay around $40 per video with multiple hook variations. For concepts we have more confidence in and want higher production quality on, we may pay around $100 per video with multiple hooks. Those are just the rates that have worked for us. You can also source creators yourself by messaging people in your niche on TikTok and Instagram. You can make the videos yourself or pay friends. The important part is building a constant pipeline of new creative. Finding one winning ad does not mean you’re done. Creatives eventually fatigue, so you need a system where new concepts and variations are constantly being produced, tested, killed, or scaled. You also need to have a place to keep all content stored. We use frame for this and have everything organized in a way where we know exactly where every ad concept is. This makes it very easy when dealing with hundreds of ad concepts. One more thing: reputation matters on creator platforms. When you’re starting, you may need to begin with simpler campaigns, build good reviews for your brand account, and prove that you’re easy to work with. Or have a professional Instagram that has lots of followers and is verified. Then it gets much easier to attract better creators. 10. Final Thoughts This entire system is how we continued scaling Pep AI after our #1 marketing channel basically got destroyed. If influencer marketing still works affordably in your niche, keep doing it. Influencer marketing + Meta ads together can be extremely powerful. But Meta has become one of the best tools we’ve found for consistently scaling the app. You do need a testing budget you’re comfortable losing. But every dollar should be buying information. If you’re spending money and you aren’t learning anything, you’re not testing. You’re just losing money. Once you start figuring out the right offer, creative angles, hooks, and conversion signals, Meta can completely change the scale of your app. It helped us continue doing over $100,000 a month after the marketing channel we relied on most stopped working. Hope this helps. Next up for Pep AI: We will be trying UGC and influencer marketing again. We hope to find a way that works for us and I will be documenting it. If you made it this far thank you for reading and I will be posting more content soon. I had to stop for awhile when we had to fix our marketing, but I will be posting a lot more so I appreciate the support. If you have any questions, feel free to comment, DM me, or book a call through the link in my bio.

150+ signed up for Scout after my post last week, so I wanted to share something even bigger we're working on solving for creator marketing: Campaign tracking Most brands still run creator campaigns out of a spreadsheet + Gmail inbox + a lot of 'wait did we already follow up?' SO... we made the agent do the work. And its literally my favorite thing I've ever used for running a creator campaign. Just start a new campaign, add your creators (either from AI search, or upload an existing roster) and Scout > tracks all the emails > detects quotes and counters > updates statuses for you > all real-time in the dashboard. v1 is now in beta, DM for early access - going live to all users in the next week! 🤠