Posting habits that make app videos work on TikTok and Reels

Here’s the frustrating bit nobody says out loud: a single video tells you almost nothing. Short-video feeds show every new post to a small group first and push it further only if that group watches. Which video travels depends on a lot of things you can’t see. So one post does well because it caught a lucky moment, and the next one, which was honestly better, dies at 150 views.

If you only post five times, that luck decides what you think you “learned”. You’ll copy the lucky one, it flops, and you’ll conclude short video doesn’t work for your app. With thirty posts the luck starts to wash out, and you can see which lines keep landing near the top and which keep landing at the bottom.

That’s the real reason to post daily. Not because the algorithm rewards hustle, but because you need enough tries to see a pattern. The four habits below are what keep you going long enough, and cleanly enough, to see it.

Nobody can promise views, us included. What you control is how often you post, what you test, and what you keep.

Most people who give up after a week give up because every single post felt like starting from scratch: what do I record, what do I write, what time is it even. Spend one hour up front and that feeling goes away.

1. Set up the account properly. Name it after the app, use the app icon as the picture, and write a bio with one formula: what it does, for whom, where to get it. For example: “Splits shared bills in seconds. For flatmates and trips. Free on iOS and Android ↓”.
2. Record three moments of the app, 5 to 15 seconds each. The moment a new user goes “oh, nice”; a satisfying result (a finished plan, a streak, a chart filling up); and a fast flow from tap to result. Start on the action, never on a splash or sign-in screen.
3. Clean the recordings. Hide notifications, tidy the status bar, and fill the app with demo data that looks real but belongs to nobody. Someone’s real name or bank balance in a video isn’t just ugly, it’s a privacy problem.
4. Write down raw material for lines: three kinds of people the app is for, three problems it fixes, and three phrases real users have said about it (from reviews, support emails, comments).
5. Decide where the numbers go. Getuser, or a plain sheet with the date, the line, the posting time and the views. Pick now, not on day ten.

Pick one or two times a day you can keep for a month and stick to them. Ten videos in one excited week and then silence feels productive, but it teaches you very little: you get one week’s worth of results and nothing to compare them with. One video a day for thirty days is less exciting and far more useful.

Choose times from your users’ day, not yours. When does the person your app is for have a few idle minutes with a phone? The commute, the lunch break, the sofa after dinner. A student and a parent have very different days. Treat your first pick as a guess to test, not a rule.

Use their time zone. If you’re in Jakarta and your users are in the US, 19:00 for you is early morning on the US east coast. Schedule for where they are, not where you are.

Compare time slots fairly. Keep each slot for two weeks, then compare the middle value of views at each one, not the single best video. If one slot is clearly weaker, move it two or three hours and give the new time another two weeks.

Keep a buffer. Have a week of videos ready before day one. A sick day, a launch, a bad week at work: none of them should break the rhythm, because a broken rhythm also breaks your comparison.

In Getuser: set your posting times once in Schedule. Every batch fills them, and Getuser posts the videos you confirm to TikTok, or Instagram from the Mid plan, at those times. You can still download any video and post it yourself.

The line on screen is the hook. In the first second it decides whether someone keeps watching or flicks past. Of everything you can change cheaply, it moves the result the most. So if you only have energy to test one thing, test this.

A good line has three parts: who it’s for, a tension they recognise, and one specific detail. “POV: you’re a freelancer emailing about the same invoice for the third time” has all three. “The best invoicing app” has none, and it sounds like an ad, because it is one.

Short enough to read before the second second ends. Rule of thumb: if it wraps past two lines on a phone, cut it. Put the words that carry the meaning first; drop “so”, “honestly”, “guys”.

Don’t write the same kind of line every time. Each pattern does a different job, and rotating them shows you which one your audience actually responds to:

Steal your users’ words. Read your app’s reviews, your competitors’ reviews, your support inbox and the comments under your own videos. The phrases people use there (“I always forget who paid”, “my streak broke and I gave up”) beat anything you’ll write at a desk at midnight.

Write in your audience’s language. Want Indonesian users? Write Indonesian lines, the way people actually type them, not the way a textbook would.

Keep it honest. When a video opens with an AI-generated person, the line never makes them claim they used your app, got results, or that their friends use it. And don’t promise outcomes you can’t prove, like kilos lost or money earned.

In Getuser: the built-in patterns are filled in from your answers about the app. Save your own lines to Your lines once and they go into every batch. You can edit or drop any line before a video renders.

This is where most people quietly lie to themselves. You look at yesterday’s video, it’s doing worse than last week’s, you panic and change everything. Three rules stop that.

Compare videos at the same age. A video posted this morning will always look worse than one from last week; it just hasn’t had the time. Note each video’s views about two days after posting and compare those.

Use the middle value, not the best. If a line has been in five videos, sort their views and take the middle one (the median). One lucky video won’t crown a line, and one unlucky one won’t bury it.

Change one thing at a time. Change the line, the recording and the time all at once and you’ll get a different result with no idea why. Keep two fixed while you test the third.

Once a week, sort your lines by that middle value and act on it:

Views are not installs. This one hurts. A video can hit 20,000 views and bring zero downloads. Views tell you the hook works; they don’t tell you the right people came. Put your download link wherever the platform allows, use a link that counts clicks (a store campaign link or a short link), and watch your store’s daily downloads over the same weeks. A line with fewer views but more of the right viewers can be the better line.

Read the comments like research. A question in the comments (“does it work offline?”) is your next video’s line, and the answer is its recording.

In Getuser: type the views from TikTok or Instagram into Getuser. The lines that work rise to the top, so your next batch starts from them.

Post from one account that’s obviously your app’s. Someone who likes a video should know in one tap what they’re looking at and how to get it. A personal account mixing holiday photos and app videos makes that harder. Spreading the same videos over five accounts is worse: it splits your numbers so none of them tells you anything.

Then switch on the settings the platforms ask for, every time:

Treat the caption as a second hook. Name the app, say in a few words what it does and where to get it. A few hashtags that describe the app and its category beat a wall of trending tags that have nothing to do with you.

Reply from the account. Answer questions plainly and fast. It’s the cheapest support you’ll ever do, and the questions feed your next lines.

Skip the tricks, however tempting they look when you’re stuck at 300 views: bought followers, a dozen fresh accounts posting the same video, fake comments, or passing an AI-generated person off as a customer. They break the platforms’ rules against inauthentic behaviour, put the account at risk, and poison the very numbers you’re trying to learn from.

When Getuser posts a video with an AI-generated person for you, it switches the platform’s AI label on and keeps it on. More in How Getuser uses AI.

Habits survive when they’re small and scheduled, not when they rely on motivation. Here’s the whole week:

1. Day 1: prepare as above. In Getuser, answer the five questions about your app and upload your recordings.
2. Day 1: make your first batch, watch every video, and pick one or two posting times in your audience’s time zone.
3. Week 1: post every day at those times. Don’t judge anything yet, even if it’s quiet. You’re collecting the first data, that’s all.
4. Week 2: type in the views at two days old. Save the best lines to Your lines and write five of your own from reviews or comments.
5. Week 3: test one change only: a new recording, a different pattern, a moved time slot, or a slideshow.
6. Week 4: sort lines by their middle value, make the next batch from the top quarter, and keep the time slots that did best. Compare your store’s downloads across the month.

What it costs in credits: posting once a day is about 30 videos a month. A video is 25 credits, so that’s 750 credits, plus a little for captions written with AI, which show their price before you use them. The Free plan’s 100 credits a month make about 3 videos with captions, enough to see how it works. Compare the plans on Pricing.

• Stopping after a week because it was quiet. One week is too few tries to tell a weak line from bad luck. Quiet in week one is normal.
• Judging the whole app, or the whole format, from one video. One result is noise; a month of results is a pattern.
• Changing the line, the recording and the time at once. You get a different result and no clue why.
• Generic lines that could be about any app. If the line fits your competitor just as well, it’s not doing its job.
• Hiding the app. Viewers should see it working in the first seconds. The app is the proof.
• Passing an AI-generated person off as a user or a review. It misleads viewers and can break consumer-protection rules.
• Forgetting the AI label or the brand disclosure. Risking the account over a two-second setting is a bad trade.
• Posting the same video from several accounts. It splits your numbers and looks like spam to the platform.

How often should I post? Once a day if you can keep it up. If you can’t, post less often but at fixed times. The rhythm matters more than the count, because it keeps your comparisons fair.

What’s the best time to post? There isn’t one for every app, whatever the charts online say. Start from when your users are idle with a phone, in their time zone, then let your own numbers decide after two weeks.

Same video on TikTok and Instagram? Sure. Each platform has its own audience, so it’s a separate test on each. Just keep the numbers separate when you compare.

Should I delete videos that flop? No need. A quiet video costs you nothing, and deleting posts makes your own history harder to read later.

How long until it works? Nobody can tell you, and anyone promising views is guessing. Give it a month of daily posts before you decide.

Do I have to show my face? No. See How to market your app on TikTok without showing your face.