Matt Gittleson - Our app makes $750k/mo thanks to this UGC strategy

Almost ignore stats (no Modash filtering). Scroll a creator's TikToks in any niche (beauty, lifestyle, pets) and ask: how much do I want to watch and rewatch this, and why? If you can't say why, that's a good sign. Checklist: (1) emotive, not monotone, especially with the eyes; (2) pause on the first second: is the frame aesthetic (bright light, good positioning, interesting background, well-proportioned on-screen text)? This proxies for natural 'viral sense'; (3) extroverted 'yapper' who can tell a story. Hidden gems often have one exceptional skill; the rest can be taught. College students are a good pool. These creators come cheaper than influencers. Creators cross-post, but most results come from TikTok.

Build rapport and a real relationship with each creator and commit to improving their content rather than being transactional. Don't expect a million views in week one. Keep the creator's natural charisma and coach the learnable parts (lighting, framing, text size, scripts). Pay includes performance bonuses. Treat creators as team members, not disposable numbers, because competition for good creators is fierce.

No shortcut on sourcing: hours of scrolling to find hidden gems (DMs can be sent by a tool or VA, but not the picking). Outreach DMs to college-age creators are written all in lowercase, casual and personable, with as much specific social proof as possible (Jenni has 5-6M users, results of existing creators, links to creator accounts). Avoid vague, long pitches. Set an aggressive target (80 creators in the first month). Expect lower reply rates than influencer outreach: these creators rarely check DMs and are wary of scammy jewelry-brand offers. Then move to a call, where most close.

Format: a 5-second clip of the creator making a surprised face, with a long text overlay that takes ~15 seconds to read, so viewers watch it about three times. Matt writes the overlay word for word. Example: 'Just seen a girl in the library writing an essay, no notes, no open tabs, no ChatGPT, no Jenni AI, no textbooks, no music, just her and her own thoughts against the world, like a psychopath.' The aim is 'engagement bait', not pure rage bait: the claim should split opinion so some viewers attack it and others defend it (here, 'no tabs and no music is what makes this crazy'). The product name sits inside the text. Coming up with the angle takes emotional intelligence about what people are talking about now, adapted to the niche.

After the library video plugged Jenni too weakly, later scripts made the app the answer to the story. Example overlay: 'Our professor just said, if I even think ChatGPT wrote your essay, you get a zero... Chat, I am begging you to please use a humanizer for the grammar and Jenni for the fake citations. Do not risk your whole degree for ChatGPT. It doesn't love you back like that.' Rule: a viral study TikTok that doesn't mention Jenni is useless; adjust each video depending on whether it leaned too far toward virality or toward the plug.

No creators: the founder made all the content himself, sometimes with clips of friends and family. His first real TikTok (after five meme posts) was him talking into his phone, adapting an existing format: 'Made a website called I have an essay due at midnight and ChatGPT won't stop giving me fake citations .com...' He named the pain, explained the tool (upload ChatGPT citations, see which are fake, get real replacements, helps with AI detectors like Turnitin) and said 'I'll put a shorter link in the comments.' Advice: early on, don't invent formats; imitate and adapt. Filming yourself beats carousels, and it teaches you what to tell creators later.

Prescribed 'doom scrolling' to build viral sense (judging whether a video will go viral): half an hour a day on content in your niche (for him, studying) and half an hour on content outside it, then adapt what works elsewhere. Don't rely only on outlier-finder tools, because you also need to see what fails. Judge principles, not just view counts, since outcomes are noisy.

Understand the principle behind a winning video, then ask 'what's the one thing I would change to make it better?' and make many such variations. Views vary a lot by luck (near-identical videos differed 2,000x), so the goal is more chances at luck, not guessing the cause of one hit. At Jenni's scale they test patterns such as eye expressiveness across many videos.

Another brand had its creators copy Jenni creators' videos as closely as possible, buying near-identical outfits and filming in front of near-identical backgrounds.

An experiment over the next few months: Matt spotted a strong format in a completely different niche that hasn't been done in study, came up with a product or feature that would fit it, asked engineering if it could be built, and it is being built. Product is designed from the format outward.

Citesure solved a narrow slice (fake citations) of the problem Jenni AI solves, so Jenni understood it instantly. Jenni's account DMed the Citesure Instagram (which showed clips of a friend) to recruit the 'creator'; Matt replied with an eyes emoji offering 'one better' if they wanted an acquisition, saying Jenni had been an inspiration. The CEO offered a call right away; they spoke the next day and he made a first offer on that call. Closing took 4-5 months of lawyers and UK tax work. Advice: pick the acquirer on the founder and culture you'd learn from, not the product.

Rather than hiring a pure creative, acqui-hire a young founder who had to grow their own app with scrappy content. They understand the whole funnel through to revenue and are used to sink-or-swim pressure, so they write for conversion, not just views.