Everyone will tell you that you can't scale a consumer app without a UGC army and a marketing budget. You need 30 creators. You need to fire 20 of them. You need $10k to even find out if there’s any signal to continue.
We had none of that. No UGC program, no capital for ad spend, and I spent half that month traveling around NY and SF. Yet we still did 12,000+ downloads in our first month of marketing organically. How?
Here's exactly what we did, step by step, and how you could take this way further than we did.
(for context, it's a screen-time app. My cofounder runs the technical side, I run marketing and growth.)
The one idea this whole thing runs on
Find one winning video, because if it goes viral once, it can (and will) go viral again.
That's it. That's the whole strategy. Everything below is just me taking one video that worked and squeezing every download out of it that I possibly could.
Most programs go viral once, celebrate, and then waste additional resources trying to recreate the outcome.
(to preface, I still believe that UGC is the best distribution play in consumer right now. It's the top of every marketing thread for a reason, but it's a machine. You're hiring creators, testing formats, and eating higher CPMs until you find your format. I didn't have the time or the capital to build that machine this month, so instead I found a loophole that we're now building the whole thing around.)
(same video, one post apart)
Step 1: Create a script that you know can go viral (and convert)
This is quite literally the only real requirement to run this whole system, to be able to manufacture virality.
It isn't luck, and it isn't the algorithm blessing you. Take TurboAI for example, millions of users in under a year off UGC. That isn’t luck a hundred times in a row, that is people who understand consumer psychology and going viral. Maybe not every script they run hits, but enough do that their business model works.
Everything else in this guide is just mechanics. This is the one thing you actually have to be able to do. Find my entire guide to virality here: https://www.atjasonwang.com/guide
Step 2: Find your hook variations
Now that you have your script, you need to maximize the odds it hits.
You don’t need to reinvent the wheel to go viral, 80% of going viral is simply taking what already works and putting your own spin on it. Unless you’re already leading over your competitors, don’t focus on that other 20%.
Find hooks that are already working on your target audience. You don’t need to be innovative at this point, nor do you need to spend hours scrolling every day, saving videos, trying to reverse-engineer why they hit. That is a full-time job on it’s own and is the exact thing we’re trying to avoid here.
When it comes to actually finding viral hooks, I usually use one of two methods.
- Dedicating an iMessage groupchat with my cofounder to sending and storing viral hooks in.
- Using lightreel (shoutout @Jibran_05 for the intro) to setting up alerts for the screen time space and adjacent niches.
Now I receive a text whenever a new hook/format starts picking up traction, which helps us stay ahead of the curve by hopping on trends before they become oversaturated.
(make sure your script is polished before putting it into lightreel, especially if you’re asking for hooks that fit the voice of your script. Also don’t just blindly use each hook lightreel generates, see if it actually fits. I typically go back and forth a few iterations, but it’s been surprisingly good at pulling real hooks instead of just synthesizing them and generating ones that sound like AI. I usually end up using 2/3 of the hooks i’m recommended, and then just create 2 iterations of each one.)
Step 3: Record it once. Actually, six times.
Record the same video six times in one session, with a different hook each time.
Now you have 6 slightly different videos ready to be A/B tested and “frankensteined” for virality.
Step 4: The loophole (this is the part that actually matters)
This whole system works because of one difference between how TikTok and Instagram handle near-duplicate content.
TikTok is notorious for shadowbanning accounts that post duplicates. So most people cross-post each video once and move on.
Instagram doesn't do this. As long as there's even the slightest difference between two videos, it won't flag them. Different background music, a word in the caption, trimming a couple frames near the end, it’s all enough for Instagram to read it as brand new.
Once you realize this, you stop thinking in videos and start thinking in parts.
I had 6 hooks, 5 body clips (6 variations each), and 6 CTAs.
Frankenstein those together and it's not 6 videos anymore. It's over a thousand. Every one of them new to Instagram.
And it doesn't stop there. Analyze retention graphs, take the variations that are actually working, change the tiniest detail again, and keep going. Your winners just spawn more winners. Again, I’ve also set up an alert on lightreel so it watches all our videos and texts me whenever something starts working.
One recording session. Basically unlimited content.
Step 5: Turning one video into millions of views
Once I found my winners, I had Codex automatically create iterations of them, tiny differences each time, new titles, trimmed frames, etc.
The results?
- one hit 920k views
- another did 340k
- multiple cleared 100k
- and a long tail sitting in the 10k–50k range
Somewhere around 2 million views total, all off of one original 40k view video.
It wasn't a one-hit wonder, It wasn't one lucky video with a slow trail-off of views, the wins were spread across the iterations, with the highest performing one near the end.
12,000 downloads. Zero dollars in ad spend. No creators on payroll.
Now imagine if I'd actually had time and money
I did this solo on the marketing side, with no budget and no team. I had one mediocre video and squeezed ~2M views out of it by myself.
Now picture running this with actual resources:
- a UGC program with 50+ creators posting 5x a week
- VA-run accounts mass-posting iterations of every creator's biggest hits
You wouldn't have one viral video getting recycled. You'd have a machine that takes every winner anyone on your team ever produces and reprints it dozens of times.
I did the small version of this by accident, because I was broke and busy. The big version is just sitting there for anyone willing to build it.
All you need? Codex, lightreel, and the ability to create just one viral script.
I'm going to keep sharing the exact playbooks we're running (and the ones we find next). Follow if you want them.
