There is a $4 app on your phone right now that makes more money than most B2B SaaS companies I interview.
It doesn't run ads and it doesn't have a sales team. It grows through TikTok accounts you scrolled past this week and never clocked as marketing.
Pat Walls and Gus from Starter Story spent an episode pulling this apart, and it is the best breakdown of TikTok app marketing I have seen. Three consumer apps, real Sensor Tower data, and the exact video format all of them run.
Here is the episode, and below it, the full playbook with the numbers.
The Tweet That Started It
Gus found the thread the way most good research starts: scrolling Twitter.
A guy named Gam posted that we have entered a "TikTok multi-account hyper-growth era for apps", then listed app after app with their revenue and view counts. TurboLearn. Cal AI. Names Gus recognized from Starter Story interviews, and a bunch he had never heard of.
The pattern behind every single one: dozens of TikTok accounts, all posting the same video format, all pointing at the same app.
None of these are official brand accounts grinding out content. They are fifteen or twenty accounts that look like regular people, shilling the app in a format so repeatable you can spot it from the first frame.
Gus went down the rabbit hole on three of them. The numbers are stupid.
Locket: $300K a Month From a $4 Widget
You probably know Locket even if you don't know you know it. It is the widget that puts live photos from your best friends directly on your home screen.
The backstory is famous. Matt Moss built it as a gift for his long-distance girlfriend and shipped it on New Year's Day 2022 with one TikTok. Two weeks later it had over 2 million users and hit #1 on the App Store in 30 countries. By August 2022 it had raised $12.5M.
That launch spike could have died like most viral apps do.
It didn't, and the episode shows why. At the time of recording (April 2025), the tweet Gus pulled up had Locket at roughly 150 million TikTok views a month, about 1 million downloads a month, and around $300K in monthly revenue. It was sitting at #13 in social networking on the App Store.
The price? $3.99 a month, or $36 a year.
Gus opened the TikTok accounts driving it. Every video is the same:
→ One line of text on screen. "My best friend did this."
→ A hand-over-face reaction shot.
→ Then the app, on screen, in use.
One of those videos had 8 million views. The accounts look like regular college kids. Same faces show up across multiple accounts. Whether the "I just moved away from college" story is even true doesn't matter to the algorithm.
It is an ad. It just doesn't look like one.
The Math: 150 Million Views Into $63K of New MRR
This is my favorite part of the episode, because Pat refuses to hand-wave. He opens a calculator and backs into the revenue.
Follow the funnel:
→ 150,000,000 views a month.
→ About 1,000,000 downloads. That is a 0.7% view-to-download rate. Less than 1 in 100 viewers.
→ At a conservative 1.5% download-to-paid conversion (industry range is 1 to 5%), that is roughly 15,000 new paying users a month.
→ At $4 a month, you just added about $63K in new MRR. This month. Again next month.
Then he layers in 6% monthly churn and lets the table run.
If those numbers held, the business crosses $500K a month by month 12. By month 30 it is approaching $1M a month. From a widget that shows you photos of your friends.
The assumptions are the whole game, obviously. Attention like that does not stay constant forever. Which is exactly why these teams sprint when a format works instead of slowly "testing" it.
But notice what the funnel teaches you. Every individual rate is terrible. 0.7% download rate. 1.5% paid conversion. The volume is what makes it print, and the volume comes from one repeatable video format, not from creativity.
CoupleJoy and Einstein: Same Playbook, Different Niche
One example could be a fluke. Gus brought two more.
CoupleJoy is a relationship app for long-distance couples. Widgets, shared streaks, messages on your partner's lock screen. Sensor Tower showed about 500,000 downloads in a month and #19 in lifestyle, with pricing from $12.99 up to around $40 a month.
The TikToks? Identical structure to Locket's. "I've been in a long-distance relationship for two years and just found out this app lets me send messages to his lock screen." Hand over face. Show the app. 3 million views. 2 million views. 1 million views.
The one upgrade: tracked links like couplejoy.app/nessie, so the team knows exactly which creator drove which install.
Einstein AI Homework Helper is a study app at $12.99 a month. Around 18 million views a month, 70,000 downloads, roughly $90K in monthly revenue per the Sensor Tower data in the episode, ranked #93 in education.
Its creators run fake street-interview clips. "What did you get on the exam? What did you use to study?" Gus even caught the same girl answering on two different accounts.
Pat ran the calculator again: at a 4% paid conversion (students before an exam convert hard), Einstein is adding roughly $37K in new revenue a month. Hold that for a year and you are at $346K a month.
All of that from a homework app, run by TikTok accounts pretending to interview students on the street.
The 4 Rules of the Format
Strip the three apps down and the playbook is four rules.
1. One hook, then the app. Always the app.
Pat calls this intent. Every video with millions of views shows the actual product being used. The moment you go viral about your breakfast instead of your product, the views are worthless. People do this constantly and wonder why 500K views drove 12 installs.
2. Run many accounts, not one.
Each app had 15 to 20 accounts running the same format. Part of it is surface area. Part of it, according to David Park from Jenny AI, is a hedge: accounts get banned, and if one dies you have nineteen more. I broke down how David rode this exact strategy to eight figures in the Jenny AI story.
3. Copy a proven format until it stops working.
Hand over face. One line of text. Screen recording. These creators are not artists, and that is the point. SNL has run the same format for 50 years. Find the format that converts for your app and repeat it until the numbers say stop.
4. Ride a platform shift.
Locket and CoupleJoy both exist because Apple let apps put widgets on the home screen. A boring OS update, and suddenly there is real estate 100 phone-checks a day land on. Pat's challenge to founders: what is the equivalent for your market? (His idea: your MRR from Stripe, live on your home screen. Someone should build that.)
Should You Copy This?
Pat and Gus wrestle with this on camera, and I respect that they don't pretend it is clean.
Is a 26-year-old posting "moving away from college" content she scripted for an app... real? No. Is it that different from any TV commercial you have ever yelled at? Also no. Gus lands on "these are just ads", and honestly, that is where I land too.
The disclosure question will catch up with this playbook eventually. The platforms will force labeling, the same way they did with creator ads.
But the underlying lesson is not "run fake college girl accounts". The lesson is that attention solves every other problem in your business. It validates the idea, funds the roadmap, and tells you what to fix. The founders winning right now, like the Cal AI teenagers who did this with fitness creators, treat distribution as the product.
If you want the tame version of this playbook for your own app, I wrote up how PlayKit drove 12 million downloads with UGC. Same physics, less cosplay.
And if you sell to businesses instead of college kids, don't dismiss this. The format changes but the math doesn't. You are better off posting one repeatable format 50 times than one clever video once.
FAQ
What is TikTok app marketing?
TikTok app marketing is growing app downloads with short-form TikTok videos instead of paid ads. The version exposed in this episode uses networks of 15 to 20 creator-style accounts posting one proven video format (a one-line hook, a reaction, then the app on screen) with millions of monthly views funneling to the App Store.
How many TikTok views does it take to make money with an app?
In the episode's Sensor Tower examples, view-to-download ran under 1% (Locket: 150M views to about 1M downloads) and download-to-paid ran 1 to 5%. So a $4/month app needs millions of views to move the needle, while a $13/month app like Einstein added roughly $37K in new monthly revenue from 18M views.
Do these TikTok accounts disclose that they are ads?
Mostly no, and that is the uncomfortable part of the playbook. Many accounts pose as regular users, and Gus found the same creator running multiple accounts. Expect platforms to force ad labeling on this eventually, so build the skill (repeatable formats showing your product), not the loophole.
Does this playbook work for B2B SaaS?
The literal format won't, but the mechanics do: pick one channel, find one repeatable content format that shows your product solving the problem, and post it at volume with tracked links. Founders on the podcast have applied the same math to LinkedIn, X, and YouTube instead of TikTok.
Want the playbooks before they get exposed?
Every week I sit down with bootstrapped founders doing $100K to $10M a year and pull apart exactly how they grew. No fluff, just the numbers and the moves.
Listen to the Profitable Founder Podcast →