How an AI app went from 10,000 to 2 million users at $0.02 per user

ZuAI went from about 10,000 users to 2 million at a blended customer acquisition cost of $0.02. No viral trick. We tested 150 or more creatives a month, ran TikTok UGC and Reddit side by side, tracked blended CAC every week, and killed anything that did not hold its number. Most of the work was boring.

Where it started

ZuAI is a consumer AI app. When I started on growth, it had roughly 10,000 users. That is the number everything after gets measured against.

At 10,000 users my job is not to grow. My job is to find one channel that produces users at a price that repeats. Growth comes later, and only from channels that have already proven they can repeat.

I say this on every founder call. A channel that produces ten customers is an experiment. One that produces them repeatedly inside a predictable CAC range is a growth engine. At 10,000 users, ZuAI had experiments. The work was turning one into an engine, then a second.

Today the account runs about $300,000 a month in ad spend. The numbers are on the ZuAI case study page. This post is about how the work was actually done.

Most of the work was boring

The Tribune ran the story in September 2026 under the headline An AI App Reached 2 Million Users at $0.02. Its AI App Growth Marketer Says Most of the Work Was Boring. Business Standard carried it too. The headline is accurate.

Founders ask me for the trick. There was no trick. There was a loop, and it ran every week:

1. Research what users say about the problem the app solves.
2. Turn that language into creatives.
3. Launch the creatives with a small, fixed budget.
4. Read the numbers at a fixed spend threshold.
5. Kill the losers. Scale the winners.
6. Repeat.

None of those steps is clever. What matters is that they happen every week without exception. The compounding comes from repetition, not from any single week.

150 creatives a month

The number people quote back to me is 150 creatives a month. Here is why the volume matters.

In my campaigns, most creatives lose. Not some. Most. A creative that looks strong in the edit dies in the feed. A hook I would never have picked beats the one I loved. I cannot predict which is which, and I stopped trying. Volume replaced prediction.

What counts as a creative here:

• A distinct hook in the first 2 seconds.
• A distinct script or on-screen text.
• A distinct format: talking head, screen recording, text on screen, reaction, duet.
• A distinct creator or voice.

Change one of those and it is a new creative. Change the caption only and it is not.

Production is a pipeline, not a studio. Briefs go out on a schedule. Raw footage comes back from creators. Cuts are fast and rough. Polished creatives do not beat raw ones in my tests, so polish is not where the hours go. The hours go to launching, reading and killing.

How TikTok UGC and Reddit fed each other

Two channels did most of the work: TikTok UGC and Reddit. They look unrelated. They are not.

Reddit gave us the words. People on Reddit describe their problem in full sentences, with the frustration intact. I read threads where users asked questions the app could answer and copied the exact phrasing. Those phrases became TikTok hooks. A hook in the user’s own words beats a hook written by a marketer.

TikTok gave us the volume. Creators shot short videos using the app on the problems people described on Reddit. The videos ran as paid UGC ads and as organic posts. The comments told us the objections. Those objections went back into Reddit answers and into the next round of scripts.

I have run Reddit tests across 40 or more accounts. Reddit rewards answers and punishes pitches. A comment that solves the problem, with the app mentioned once as one option, survives and keeps sending installs for months. A comment that reads like an ad is gone within the hour.

The loop between the two:

• A Reddit thread reveals a question.
• A TikTok creative answers it in 15 seconds.
• TikTok comments reveal the next objection.
• A Reddit reply handles the objection in plain text.
• Both feed next week’s creatives.

How blended CAC was tracked

The $0.02 figure is blended CAC. All acquisition spend, paid and organic, divided by all new users, paid and organic. One number for the whole account. I do not report platform CPI to founders. Platform CPI hides organic and hides creator fees.

In the numerator:

• Ad spend on every platform.
• Creator fees for UGC.
• Tools for tracking and creative production.

In the denominator:

• New users in the period, from every source, attributed or not.

I read it weekly. Weekly is slow enough to have real numbers and fast enough to kill something before it burns a month.

Under blended CAC sit two more numbers. Activation rate, meaning users who reached the first real moment of value in the app. And day 7 retention by weekly cohort. A cheap install that never activates is not cheap. It is waste with a nice CPI.

The growth calculators I use with founders are free if you want the arithmetic.

What got killed

Killing is the half of the job nobody writes about. Some of what died:

• Creatives that won on click-through and lost on activation. The hook promised something the app did not do in the first session.
• Subreddits that produced upvotes and no installs. Attention is not acquisition.
• Formats that held CAC at a small budget and drifted the moment I scaled them. If it cannot hold at 3 times the budget, it is not a winner yet.
• Any campaign I could not explain in one sentence. If I cannot say why it works, I cannot repeat it.

The rule was simple. Every creative and every channel had a number to hold. Miss it two weeks in a row and it was off. No favorites.

What a founder can copy

You do not need my budget to run this loop. You need the loop.

1. Pick one number. Blended CAC, all spend over all new users. Set a ceiling you can afford given your revenue per user.
2. Pick one paid channel and one organic channel. For most consumer AI apps I would start with TikTok UGC and Reddit.
3. Make creatives in volume. Start with 20 a month if 150 is out of reach. Change the hook, not the caption.
4. Set a fixed spend threshold per creative. Read the numbers only when it is hit.
5. Kill on schedule. Same day every week.
6. Scale only what has held its number two weeks in a row.
7. Write down why the winner won. That sentence is your next brief.

If you want to run this alongside other founders doing the same, The Goal Digger Club is where I am building that. It is early and there is a waitlist.

If you want someone to run it for you, that is the AI app growth marketer work I do. The user acquisition page covers the scope, or you can book a call.

FAQ

Is $0.02 CAC realistic for my app?

Probably not as a target, and it should not be. ZuAI’s CAC comes from a consumer product with wide appeal, a heavy organic share, and a long run of creative learning. Your number depends on your revenue per user. Set a ceiling you can afford, then work it down through testing. The method transfers. The exact number does not.

How many creatives should I test each month?

As many as you can launch and read properly. I run 150 or more a month on ZuAI because the budget supports it. A founder starting out can run 20 with a phone and two creators. What matters is that each creative changes the hook or the format, gets a fixed spend, and gets a clear keep or kill decision.

Did ZuAI grow from paid ads or organic?

Both, and I refuse to separate them in the headline number. Paid TikTok UGC bought volume. Reddit and organic TikTok posts brought users at close to zero marginal cost. Blended CAC counts every dollar and every user together. That is the only honest number for an app where paid and organic feed each other every week.