WEBVTT 1 00:00:00.000 --> 00:00:04.937 Can two people really run 60 creators every single day, with zero employees? 2 00:00:05.237 --> 00:00:12.395 One founder says yes, and it took his app from 25 thousand dollars a month to 150 thousand dollars a month. 3 00:00:12.695 --> 00:00:13.975 So what's the playbook? 4 00:00:18.275 --> 00:00:27.627 Welcome to Build Wiz AI Show. Big AI ideas, made clear for builders. I'm Bob, and today Alice takes us through a marketing playbook. 5 00:00:27.927 --> 00:00:42.398 Hi everyone. It's an X article by Mau Baron, called How we ran a UGC army to 150 thousand dollars a month with 0 employees. UGC means user-generated content: creators filming short videos about your app. 6 00:00:42.698 --> 00:00:44.997 And we have two guests to push back on it. 7 00:00:45.297 --> 00:00:48.798 Yes, Sam and Maya each bring an outside source. 8 00:00:49.098 --> 00:00:51.631 So what's the app, and how big did this get? 9 00:00:51.931 --> 00:01:04.993 The app is Prayer Lock. He says UGC took it from 25 thousand dollars a month to 150 thousand dollars a month, with more than 120 million views and more than 450 thousand downloads. 10 00:01:05.293 --> 00:01:07.435 And all of that with just two people? 11 00:01:07.735 --> 00:01:22.938 Him in Mexico and his cofounder Ernesto in the US, and they never had a single employee. His point is that everyone talks about finding and paying creators, but almost nobody explains what comes after: running 60 creators every day. 12 00:01:23.238 --> 00:01:24.701 Why is that the hard part? 13 00:01:25.001 --> 00:01:34.901 Because it grows fast. You go from 5 creators to 30 to 60 in a couple of months. Without a way to manage it, you burn out or start wasting a lot of money. 14 00:01:35.201 --> 00:01:36.429 Where does it start? 15 00:01:36.729 --> 00:01:48.641 With the app itself. He says creators can't save an app that has nothing worth showing on camera. You need two things: a feature that makes people stop scrolling, and a funnel that makes them pay once they download. 16 00:01:48.941 --> 00:01:51.187 What counts as a stop-scrolling feature? 17 00:01:51.487 --> 00:02:03.216 One you can explain in one sentence and show in one screenshot. Theirs was block your phone until you pray, paired with their mascot. Think of it like a movie poster: one line, one image, and you get it. 18 00:02:03.516 --> 00:02:05.240 And if I don't have that sentence? 19 00:02:05.540 --> 00:02:15.598 Look at what your competitors' creators say in their videos, not at their app, and make your own version. He calls copying what already works just smart product thinking. 20 00:02:15.898 --> 00:02:18.196 Bold. And the funnel side? 21 00:02:18.496 --> 00:02:32.315 UGC traffic is cold traffic, so a weak funnel shows fast. They were at a 10 percent trial start rate and 60 percent day 30 retention when it took off. But he says 3 percent to 5 percent download to paid is enough to begin. 22 00:02:32.615 --> 00:02:35.802 Okay, hiring. How much money do I need? 23 00:02:36.102 --> 00:02:47.256 He says have around 15 thousand dollars ready. Not because creators are expensive, but because you spend a month of posts before you know what converts, and you need to scale the same week something hits. 24 00:02:47.556 --> 00:02:49.254 And how many creators to start? 25 00:02:49.554 --> 00:02:57.548 Three. They found their format with 3 creators first, and he'd do it again: 3 people you talk to every day beat 20 you don't. 26 00:02:57.848 --> 00:02:59.154 What's the hiring bar? 27 00:02:59.454 --> 00:03:14.762 You're hiring actors. Nobody gets in who hasn't gone viral at least once on their own, meaning 500k views on a single video. Then they want niche experience, a natural hook, good editing, fast replies, and people who actually belong to the niche. 28 00:03:15.062 --> 00:03:16.237 Belong how? 29 00:03:16.537 --> 00:03:25.784 Every one of their creators was Christian, and he says that mattered more than any editing skill. It's like casting: the audience can tell who really lives the part. 30 00:03:26.084 --> 00:03:26.842 And the pay? 31 00:03:27.142 --> 00:03:41.091 Base plus bonus. That's 500 dollars to 1,000 dollars a month for 60 posts, plus 20 dollars to 100 dollars per 100k views. The bonus is capped at 1M views and counted over the first week. 32 00:03:41.391 --> 00:03:42.933 Should I squeeze that down? 33 00:03:43.233 --> 00:03:51.879 Negotiate, but not so low the creator stops caring. A cheap creator who doesn't try costs you more than a fair one who wants you to win. 34 00:03:52.179 --> 00:03:53.459 And where do they post? 35 00:03:53.759 --> 00:04:01.021 Everywhere: TikTok, Instagram and YouTube Shorts, from fresh accounts warmed up in your niche for a couple of days first. 36 00:04:01.321 --> 00:04:04.299 What about shadow bans? Everyone blames those. 37 00:04:04.599 --> 00:04:17.242 He calls them a myth. If a video gets 200 views, it's not the algorithm, it's the hook. They saw one creator go from hundreds of views to hundreds of thousands just by changing the first sentence. 38 00:04:17.542 --> 00:04:19.423 Wow, one sentence! 39 00:04:19.723 --> 00:04:23.407 Now the part he says nobody explains: managing the army. 40 00:04:23.707 --> 00:04:36.402 Three things matter: tracking your numbers, keeping a content edge, and culture. Tracking first. Every single day you need to know how many views you got, which creator got them, and whether they turned into downloads. 41 00:04:36.702 --> 00:04:38.165 How do you read those numbers? 42 00:04:38.465 --> 00:04:51.030 Like a doctor reading symptoms. Under 1k views on a creator who has gone viral before is a hook problem, and it's your fault. 1k to 10k views is a retention problem, so look at the script and the delivery. 43 00:04:51.330 --> 00:04:53.341 And big views with no downloads? 44 00:04:53.641 --> 00:05:04.378 100k views or more with no downloads is a conversion problem: check where the app shows up in the video and what the comments look like. He says every bad week falls into one of those three. 45 00:05:04.678 --> 00:05:07.812 What did the daily grind look like at 60 creators? 46 00:05:08.112 --> 00:05:21.774 Every creator needed a script, a hook and an example video, every day, and someone had to chase whoever didn't post. Every Sunday night, for months, Ernesto matched every video to each creator's deal to build the payout list. 47 00:05:22.074 --> 00:05:23.772 Ouch, every Sunday. 48 00:05:24.072 --> 00:05:36.219 And their hook log was a Google Sheet always three weeks behind, so they kept re-testing hooks that had already stopped working. At 10 creators you can do it by hand. At 60, he says, it's a full-time job. 49 00:05:36.519 --> 00:05:38.478 So vanity views are a trap too? 50 00:05:38.778 --> 00:05:49.071 Yes. A 2M view day with a flat download line isn't a win, it's a conversion problem. You only catch it by looking at views and downloads side by side, every day. 51 00:05:49.371 --> 00:05:51.565 What does a format that converts look like? 52 00:05:51.865 --> 00:06:06.180 The app in the first 15 seconds, a demo or screenshot in the video, the app name in the caption, and comments full of people asking what the app is. Their best formats were reaction plus demo, talking head, and skits where the app is the joke. 53 00:06:06.480 --> 00:06:10.268 Next, the content edge. Do winning formats die? 54 00:06:10.568 --> 00:06:19.031 He says your winning format stops working at some point. After two or three months of millions of views a day, people have seen it and conversions drop. 55 00:06:19.331 --> 00:06:20.638 So what's the fix? 56 00:06:20.938 --> 00:06:37.290 An 80 20 rule: 80 percent of creators post what's proven and 20 percent test what's next, every week. Picture a restaurant that keeps its best sellers but always runs a few specials. The best ideas come from your own creators, sending what catches their eye while they scroll. 57 00:06:37.590 --> 00:06:40.438 And should I scale to a hundred creators once it works? 58 00:06:40.738 --> 00:06:50.873 He warns against it. They did that, and the return per creator dropped fast after 30. For most apps, the sweet spot is 10 to 30 creators you actually know. 59 00:06:51.173 --> 00:06:52.035 And culture? 60 00:06:52.335 --> 00:07:04.665 The one thing you can't delegate. Most apps treat creators like a vending machine, and creators know it. They hired people who used the app, shared every win, praised good work in the group chat, and explained every cut. 61 00:07:04.965 --> 00:07:06.062 Did that pay off? 62 00:07:06.362 --> 00:07:12.240 He says their creators turned down better offers from competitors to stay, and it had nothing to do with the pay. 63 00:07:12.540 --> 00:07:14.420 Last step: ads. 64 00:07:14.720 --> 00:07:28.304 With 10 to 30 creators you get thousands of videos a month, more creatives than any paid ads team could produce. He pushes all of them as TikTok Spark Ads, not just the best ones, because the algorithm picks better than you. 65 00:07:28.604 --> 00:07:30.302 All of them? Really? 66 00:07:30.602 --> 00:07:45.440 One of their best ads had 300 organic views when they uploaded it, and today it has millions. Spark Ads run the creator's own post as the ad, so you keep the comments and social proof. It's like a recommendation from a friend, not a billboard. 67 00:07:45.740 --> 00:07:47.020 Any setup gotcha? 68 00:07:47.320 --> 00:08:01.870 Get each creator's authorization code agreed on day one of the deal, not the day you launch. Then target the US broad, start at 100 dollars a day, and scale 15 percent to 20 percent a week if the 7 day numbers hold. 69 00:08:02.170 --> 00:08:03.528 And when do I cut an ad? 70 00:08:03.828 --> 00:08:12.240 Turn off anything under 1.5 times return on ad spend if you have a hard paywall, and use a mobile measurement partner for attribution. 71 00:08:12.540 --> 00:08:14.420 So why does the whole loop work? 72 00:08:14.720 --> 00:08:27.520 UGC feeds itself: organic views and downloads on one side, endless creatives to test as ads on the other. So creative fatigue, which normally ends paid campaigns, mostly stops being your problem. 73 00:08:27.820 --> 00:08:32.392 Time for pushback. Sam, you're skeptical of influencer marketing. 74 00:08:32.692 --> 00:08:49.097 I am. Andrew Chen, in his newsletter post Every marketing channel sucks right now, says influencer marketing brings a big spike of traffic that doesn't convert. The spike goes away after a few days, and there's lots of babysitting. 75 00:08:49.397 --> 00:08:52.375 Babysitting sounds exactly like managing an army. 76 00:08:52.675 --> 00:09:05.318 And his bigger point is the Law of Shitty Clickthroughs: over time, all marketing strategies result in shitty clickthrough rates. When a channel works, everyone jumps in and it decays. 77 00:09:05.618 --> 00:09:15.388 To be fair, the article sees that too: it's why they keep 20 percent of creators testing new hooks. But Andrew's warning is about the whole channel, not one format. 78 00:09:15.688 --> 00:09:18.666 Maya, you looked at sponsored posts themselves. 79 00:09:18.966 --> 00:09:34.691 Yes. Thales Bertaglia and colleagues, in an arXiv study of over a million Instagram posts by 400 creators in four countries, found sponsored posts get lower engagement on average. But properly disclosing ads doesn't reduce engagement further. 80 00:09:34.991 --> 00:09:37.656 So paid creators should just label their posts? 81 00:09:37.956 --> 00:09:44.983 That's what the data suggests: disclosure didn't cost them engagement, and the authors stress the importance of disclosure compliance. 82 00:09:45.283 --> 00:09:52.911 My take: these are one company's numbers, in one niche, over about a year. Treat them as a starting point, not a forecast. 83 00:09:53.211 --> 00:09:57.312 Time for Builder's Corner. Alice, what can I do this week? 84 00:09:57.612 --> 00:10:08.740 First, write your app's one-sentence feature, the kind you can show in one screenshot. Second, check your funnel: he says 3 percent to 5 percent download to paid is enough to begin. 85 00:10:09.040 --> 00:10:10.216 And before hiring? 86 00:10:10.516 --> 00:10:22.140 Start with 3 creators who have gone viral at least once, and set up one daily view of views next to downloads, per creator. Then read every bad week as a hook, retention or conversion problem. 87 00:10:22.440 --> 00:10:23.537 Let's sum it up. 88 00:10:23.837 --> 00:10:32.823 The app has to be worth filming, and the funnel has to convert cold traffic. Hire few creators, hire for fit, and fix the hook before blaming the algorithm. 89 00:10:33.123 --> 00:10:38.478 Track views and downloads together, every day, and keep some creators testing what's next. 90 00:10:38.778 --> 00:10:44.081 And treat creators like a team, not a vending machine. Then run all the videos as ads. 91 00:10:44.381 --> 00:10:45.504 What's still open? 92 00:10:45.804 --> 00:10:58.787 Whether it carries over beyond a niche as close-knit as theirs, where every creator shared the audience's faith. And how long a creator channel lasts before, in Andrew Chen's words, everyone jumps in and it decays. 93 00:10:59.087 --> 00:11:04.730 That's the episode. Our main source was Mau Baron's X article on running a UGC army. 94 00:11:05.030 --> 00:11:13.232 Our guests brought Andrew Chen's Every marketing channel sucks right now, and the Instagram disclosure study by Thales Bertaglia and colleagues. 95 00:11:13.532 --> 00:11:18.652 That's all for today on Build Wiz AI Show. Thanks for listening, and see you next time.