How Much Content Should You Post to Promote an App? About 21 Posts a Day, Before You Optimize Anything

The short answer: about 21 posts a day, 7 originals across 3 platforms

How much content to post to promote an app has a concrete operator answer: roughly 21 posts a day, built from 7 original pieces cross-posted to TikTok, Instagram Reels and YouTube Shorts. The same file goes to X at no extra production cost, so the real number you produce is 7, not 21. The volume rule underneath it is blunt. A few posts a day is a real account. One a day is a hobby.

That's a founder playbook from Mau Baron, who took Prayer Lock to $25K a month solo, not a controlled study. Treat it as a target, not a law. The point of the number is that it forces a production system instead of a posting habit, and a production system is what generates the data you need before you touch your funnel.

Post first because you cannot test a funnel nobody enters: the 300-install gate

Here's the measurement reason volume comes first. Below 300 installs a day, a 2-week A/B test on your onboarding or paywall produces noise, not signal. At 300 or more a day, tests reach significance in 5 to 7 days. Two unrelated operators converge on the same number: the Refeel founder at $30K MRR, who says don't touch your paywall or onboarding until you have 300 installs a day, and Mau Baron, who says 300 daily downloads is when he stops obsessing over distribution and goes back to the product.

So the sequence is distribution, then volume, then data, then conversion work. Optimizing before that isn't rigor. It's guessing with a dashboard open. If you want to know what you'll be optimizing later, the screen-count question and the aha moment are where the work goes once the traffic exists.

Volume matters on the testing side too. Apps running 50+ paywall experiments a year line up with 18.7x more revenue than apps running one, in the same category and at the same price point, and top performers average around 15 experiments a year (dataset size not disclosed). As Steven Cravotta puts it, it's not about time, it's about volume. Content volume is the upstream version of the same rule: more shots on goal, faster feedback, sooner truth.

Produce at volume with permutations, not with 21 separate ideas

Nobody films 21 original videos a day. You build a permutation system. Split UGC into two swappable components: reaction clips, recorded 50 in one session, and body or demo clips, recorded in 3 variations. Then assemble permutations by independently swapping the reaction, the body and the text hook. One afternoon of production yields weeks of content (practitioner-reported, @ErnestoSOFTWARE).

Keep the demo under 8 seconds. The hook is the reaction clip plus the overlay text, and the demo is only proof. Mau Baron calls this Reaction plus Demo the format that never fails and the best one to start with. It took his app to $2.5K a month on its own.

The other volume formats have a clear trade: slideshows convert well and double as ad creative, with some apps reportedly making $20K+ a month from slideshows alone, but they need one cohesive brand identity rather than random aesthetic images. Scraped video hook plus CTA converts worse and automates best, to the point where a script stitched 700 clips in 30 minutes. Decode before you produce: collect 5 to 10 overperforming posts in your niche until the pattern is obvious, because one example is a copy and ten is a pattern.

Messy is fine. Expensive is not the bar. In paid data, UGC-style creative outperforms polished creative by about 55% ROI, and Mau Baron's definition of a high-quality piece is simply one that worked organically, not one that cost money to make.

Cross-post the same piece everywhere, but don't literally re-upload on a cold account

Cross-posting one piece to TikTok, Reels, Shorts and X is the whole point of the 7-to-21 math, and it's safe. The nuance sits inside one platform, not across them, and the brain carries a real tension here. TikTok's 72-hour trust score on new accounts fails on duplicate images, and duplicate content sits on the reach-decay suppression list alongside banned hashtags and hard CTAs in more than 30% of posts. Meanwhile the Permutation Engine deliberately ships dozens of near-identical posts and reports no penalty as long as engagement signals are there.

The deciding variable is account trust and how literal the duplication is. A brand-new account still inside its 72-hour window, or a frame-identical post, reads as automation and gets suppressed. An established account posting structurally similar but not frame-identical variants, each with real engagement, reads as normal creative iteration. Warm the account first, then vary at least one element per post.

The warmup sequence is specific. Day 1 and 2 are engagement only: scroll 30+ minutes, like 15 to 20 videos, leave 5 to 7 comments, follow 8 to 10 accounts, no posting and no bio link. Day 3 is one non-promotional post with trending audio. Over 1,000 views is a pass. Under 200 means start the account over.

Once you're running, expect decay and know which kind you have. A sudden 70 to 90% drop is suppression, and the fix is to pause 48 to 72 hours rather than post through it. A gradual 4 to 8 week slide with completion rate under 40% is format fatigue, and the fix is rotating format every 4 to 6 weeks. Healthy traffic is 50 to 80% from the For You feed. Under 5% means you're suppressed (practitioner-reported figures, no dataset size given).

Burn in subtitles, because the written hook is the real hook

Quick tip with an outsized return on a volume workflow: put burned-in subtitles on every video before you cross-post it. A tool like autosubbed does the transcription and styling in bulk, which matters when you're shipping 7 pieces a day and cannot afford a manual caption pass on each one. Being straight about evidence weight here: the brain has no dataset on sound-off viewing share, so treat this as a production tip, not a benchmark.

The mechanism behind it is sourced, though. Feeds are scored on completion, not on intent, and format fatigue is diagnosed at completion under 40%. A viewer with the volume off and no text on screen has nothing to complete. Written text is also what carries the hook archetypes that actually work: Contradiction, Curiosity Gap, Specificity with numbers and unusual detail, named Authority, and Story Entry dropped mid-narrative (taxonomy validated against Captain Hook AI, February 2026, and an OpusClip analysis of 34,635 clips).

Two hard constraints for the written layer. Slide 1 text has a ceiling around 21 words, so the first frame is a headline and not a paragraph. And in a slideshow, every slide needs its own mini-hook in under 12 words, because a carousel that only hooks on slide 1 bleeds viewers at every swipe. Length punishes you the same way: a 10-slide carousel might show its CTA slide to only 20% of viewers, while a 3-slide carousel shows it to nearly 100% (single skincare-app case study, @adriamatz).

Volume finds winners you could not have predicted, and winners become ads

The strongest argument for mass content isn't reach. It's discovery. A big account going viral is audience-dependent and tells you nothing you can copy. A 1K to 10K micro-account going viral is format-dependent, because there's no audience advantage to explain it away, which makes the format itself the thing worth reverse-engineering. Confirm it's repeatable before you commit: 2 to 3 viral posts is a pattern, one is luck. The same test works when hiring creators, where follower count is ignored and 2 to 3 viral videos is the filter.

Then the winners graduate. You only need around 5 high-quality organic pieces a week to start testing ads, where high-quality means it already worked for free. Boosting a real organic post beats making a new ad from scratch: on TikTok, Spark Ads beat dark posts by 30 to 142% on completion and 37 to 69% on conversion rate, and UGC-style creative outperforms polished by about 55% ROI (practitioner-reported).

Volume stays the constraint even after you start paying. TikTok CPMs run $2.60 to $6.60 against Meta's $9 to $15, but creative fatigue hits in 2 to 7 days there versus 2 to 4 weeks on Meta. The bottleneck is how much creative you can produce, not how much you can spend, which is exactly why the organic content machine is the asset. One gate before you turn spend on: your onboarding should already convert at least 10% of downloads into trials, otherwise you're paying to fill a leak. The full readiness checklist is in when to start paid ads.

More views does not mean more downloads, so measure installs, not vanity

This is the part that saves you months. Views are not installs, and a viral post that converts nobody is information, not success. Mau Baron launched three faith apps before Prayer Lock and got zero conversions on launch day for every one of them, including one that went viral on TikTok. He killed them. Prayer Lock got 5 conversions on day one and became the $25K a month app.

His thresholds are usable as a keep-or-kill rule on a new app: 1 to 5 trials or a direct conversion on launch day means keep going, and 100K views with zero conversions means the idea is dead, move on (single-operator pattern, no dataset). For a live app in steady state, the daily dispatch is different and simpler. Compare today's downloads to the trailing 7-day average. Below it, the constraint is attention, so go make content. At or above it, the constraint is conversion, so go fix the funnel.

Measuring the link from content to install is genuinely hard, and you should know that going in. True view-through attribution for organic social does not exist, because MMP infrastructure is built for paid. Every method is a workaround: a UTM-tagged bio link captures bio clicks only, deferred deep links via Branch or AppsFlyer are reliable on Android but capture only 30 to 40% on iOS thanks to ATT opt-out, and spike correlation against App Store Connect's daily install delta is the crude fallback. The TikTok Pixel does not track app installs at all. It only fires on a website, which is one more reason the web-to-app funnel is worth a look.

When to stop posting and start optimizing: the honest disagreement

Two camps disagree about a new app's first move, and the disagreement is narrower than it looks. Steven Cravotta says optimize first: paywall, then onboarding, then marketing, because scaling into an unoptimized funnel scales a leak. Mau Baron says distribution first, and don't touch the app at all until the content engine is running near 300 daily downloads. The Refeel founder's 300-install gate backs the distribution side on the measurement point.

Where they agree is paid spend. Mau's own rule gates ads on a 10% download-to-trial onboarding, which is the same thing Cravotta is saying. So the deciding variable is organic versus paid, plus app phase. Organic content can and should run before the funnel is tuned, because that's how you find hooks and reach test-worthy volume. Paid waits for the converting funnel.

The practical resolution: a pre-optimization new app posts at volume until it reaches roughly 300 installs a day, then optimizes onboarding, then scales paid. A live app in steady state dispatches by the daily number against its 7-day average. Never scale paid into a funnel converting under 10% of downloads to trials.

How to run the volume phase

The volume phase has one job, which is to buy you enough traffic and enough winning hooks that your funnel numbers finally mean something. After that the leverage moves inside the app, and cheap reach without a converting funnel is just an expensive way to be ignored (the trial-to-paid benchmark is where that shows up first). If you want the funnel ready for the day the traffic lands, tasu is the MCP your AI coding agent connects to for sourced onboarding and paywall benchmarks while it builds. It's at /mcp.

• Target 7 original pieces a day, cross-posted to TikTok, Reels and Shorts, which lands you near 21 posts. A few a day is a real account, one a day is a hobby
• Build permutations instead of ideas: 50 reaction clips in one session, 3 demo variations, demo under 8 seconds, swap reaction plus body plus text hook
• Warm a new TikTok account for 72 hours before posting, and vary at least one element per post rather than re-uploading an identical file
• Burn subtitles into every video and treat the on-screen text as the hook, with slide 1 under about 21 words
• Cap carousels near 3 slides so nearly everyone reaches the CTA, and give every slide its own hook in under 12 words
• Rotate format every 4 to 6 weeks, and if reach drops 70 to 90% overnight, pause 48 to 72 hours instead of posting through it
• Judge posts on installs and trials, never on views. 100K views with zero conversions is a pivot signal on a new app
• At roughly 300 installs a day, switch modes: now your onboarding and paywall tests will return signal instead of noise