When Should You Start Running Paid Ads for Your App? Not Before 10% of Downloads Start a Trial

The short answer: wait for a 10% download-to-trial rate and a few organic winners

Do not start running paid ads for your app until two things are true. Your onboarding already turns at least 10% of downloads into trial starts, and you already have around 5 pieces of organic content that worked without any spend behind them. Hit both and paid ads pour fuel on a fire that is already lit. Miss either and you are paying to fill a leaky funnel faster, which is the most common way a new founder burns a first budget.

The logic is simple. Paid ads do not create demand and they do not fix a bad funnel. They scale whatever your funnel already does. So the readiness question is never can I afford ads, it is does my funnel already convert the free traffic I have. Get the funnel right first, then let paid buy more of the users it already knows how to convert.

Paid ads scale whatever your funnel already does, good or bad

Per-user revenue sets the ceiling on what you can afford to pay for a user. Apps that earn more per user can always outbid you in the ad auction, and the only way to out-spend them is to out-earn them first. That is why you fix the funnel before you open the ad account. A better funnel raises LTV, a higher LTV raises the CAC you can afford, and a higher affordable CAC unlocks more ad inventory. This is stated as core economic logic in the tasu brain, an assertion rather than a dataset.

Run the math the other way and the early-ads mistake is obvious. If your onboarding converts 3% of downloads to trials, every paid install is priced against that 3%. Double your traffic with spend and you have doubled a 3% leak, not built a business. The hard paywall versus freemium data shows how far conversion swings on funnel structure alone, so the same ad budget can be profitable or ruinous depending purely on the funnel it feeds.

There is one more reason to wait. The ad platform returns whatever user type you optimize for. Optimize for installs and you get window shoppers. Optimize for trial starts or purchases and the delivery algorithm goes and finds payers instead. That lever only works once your funnel can actually convert the payers it delivers.

Gate one: an onboarding that converts at least 10% of downloads to trials

The first hard number is 10%. Iterate your onboarding until at least 10% of downloads start a trial, and treat that as the threshold that unlocks paid. It is a founder heuristic from Mau Baron, who scaled Prayer Lock to $25,000 a month, not a benchmark from a dataset, but it is the line he draws before spending. Below 10%, the fix is the onboarding, never the ad.

This is download-to-trial, the top of the funnel, not the trial-to-paid stage further down. Both matter, and they are different levers. Once trials are flowing, the trial-to-paid conversion benchmark is where you tune the next stage. To reach the 10% trial-start rate, the work is onboarding: get the user to the value moment fast, and do not gate the aha behind a wall, because a user who never feels the product almost never comes back to pay (most trial cancels happen on Day 0).

Gate two: five organic winners, turned into Organic Ads

The second gate is a content system that has already produced winners. You do not need many. Around 5 high-quality pieces a week is enough to start testing ads. High-quality here does not mean expensively produced. It means content that already worked organically, that earned watch time and saves and shares on its own before you put a cent behind it.

Then run those winners as what Mau Baron calls Organic Ads: a post with a genuinely organic hook that feels like a real organic video, but carries one clear, direct CTA. The failure mode is a vague post that only converts through the comment section, which dies the moment it becomes an ad. The platform data backs the format. On TikTok, Spark Ads that boost real organic posts beat dark posts by 30 to 142% on completion and 37 to 69% on conversion rate, and UGC-style creative outperforms polished creative by about 55% ROI (practitioner-reported figures, no dataset size given).

You earn those organic winners before you pay for reach. The cheapest way in is your own content, plus creator deals like the no-budget micro-influencer playbook that cost nothing upfront. Paid ads come after, once you know which hooks already convert.

You need 300 installs a day before a test even means anything

Even with the funnel ready, there is a volume floor. Below 300 installs a day, a 2-week A/B test returns noise, not signal, so you cannot tell whether an ad or a paywall change actually worked. At 300 or more installs a day, tests reach significance in 5 to 7 days. Two independent founders, the Refeel founder ($30K MRR) and Mau Baron, land on the same 300-a-day number as the line where optimization starts to pay off.

The platform has its own volume rule. A Meta ad set needs roughly 50 optimization events in a 7-day window to exit the learning phase, and below that, CPAs run 20 to 50% above the post-learning average (Meta platform mechanics, no dataset size given). The practical formula is minimum daily budget equals target CPA times 50, divided by 7. Most subscription apps spending $300 to $2,000 a day should optimize for trial starts, not purchases, because that is the deepest event their volume can still feed 50 times a week.

TikTok runs cheaper but burns faster. CPMs there run $2.60 to $6.60 versus Meta's $9 to $15, but creative fatigue hits in 2 to 7 days versus Meta's 2 to 4 weeks, so the bottleneck is how much creative you can make, not how much you can spend. Budget a real test: about $5,000 over 14 days is the minimum to evaluate the channel (practitioner-reported).

Market first or optimize first? The one real disagreement

There are two schools on sequencing, and they only look like they conflict. One says optimize first: test 5 to 10 paywall variations, then onboarding, then marketing, because scaling spend into an unoptimized funnel just scales a leaky one (Steven Cravotta). The other says distribution first: for a brand-new app, get organic traffic running before you optimize anything, and do not touch the app until you are near 300 downloads a day (Mau Baron).

Here is the part that resolves it. Both camps gate paid spend on an already-converting funnel. Mau's own 10% download-to-trial rule before ads says exactly what Cravotta says, which is do not scale paid into a leak. The real disagreement is narrow. For a new app, do you optimize the paywall first, or get organic distribution running first. On paid ads specifically, there is no disagreement at all.

So the deciding variable is organic versus paid, not marketing versus product. Organic content and creator outreach can and should run early on a new app, before the funnel is dialed in, because that is how you find the winning hooks and reach test-worthy volume. Paid spend waits for the 10% funnel. Get organic distribution to 300 a day, tighten onboarding to a 10% trial-start rate, then turn on ads.

How to know you're ready to turn on paid ads

Wait for the two gates and paid ads become an accelerant instead of a leak. Prayer Lock scaled from $2,500 to $25,000 a month at 50% profit margins on paid ads, run only after organic content and onboarding were already working (the full Prayer Lock teardown). Turn them on before that and you are just buying expensive data about a funnel you could have fixed for free.

The whole readiness list is a funnel-quality question, and that is the part most founders get wrong before they spend. tasu is the MCP your AI coding agent connects to for sourced conversion benchmarks while it builds your onboarding and paywall, so the funnel is ready before you pay for a single install. It is at /mcp.

• Onboarding converts at least 10% of downloads to trial starts. If it is under 10%, fix the onboarding, not the ad
• You have around 5 pieces of content that already worked organically, with no spend behind them
• You are at or near 300 installs a day, so a 2-week test returns signal instead of noise
• You run winners as Organic Ads: an organic hook with one clear CTA, never a vague post that only converts in the comments
• You optimize campaigns for trial starts, not installs, and you can feed that event about 50 times a week
• You budget a real test: roughly $5,000 over 14 days on TikTok, or target-CPA times 50 divided by 7 as a daily Meta floor