You made the account, posted your first video, and it stopped at 200 views. Three more that week, all landing in the same place. Nothing was taken down and there is no warning in the inbox, so it looks less like a penalty and more like nobody is home.
The standard fix passed around is warming the account: spend time in the app before you post seriously so the system has something to work with. It is worth doing, and the people telling you to do it are mostly right. It is also the part everyone obsesses over while ignoring the thing that actually decides whether a video travels.
What happens to a new video
Every video, from every account, goes out to a small audience first. If those people watch it through, rewatch, or share it, it goes wider. If they swipe away, it stops there. That is why a post can quietly die at a few hundred views with nothing wrong with your account.
Two things follow. Watch-through is the metric that matters, not likes. And your first job is not pleasing an algorithm, it is holding a stranger for three seconds.
Does your own activity help? Probably, and here is the honest version
The argument for warming is that a brand-new video has no history, so the system has to guess who wants it, and an account that has spent two weeks watching one category is a more useful clue than one created an hour ago.
Experienced operators will tell you this is exactly how it behaves, and they are working from thousands of repetitions, which is real evidence. What does not exist is a published confirmation from the platform. TikTok's own explainer covers interaction signals and says follower count is not a direct factor, but it does not address whether your viewing seeds your own distribution.
You do not need that settled. If the mechanism is real, two weeks in your category helps place your first videos. If it is not, you spent two weeks learning what performs in your niche, which is the thing that decides whether the video survives its test audience. Same action either way. Do it, and stop building rituals around a mechanism nobody has published.
What operators actually do in the first two weeks
Twenty minutes a day, on your phone, in the real app. Two jobs at once: teaching the system what this account is about, and teaching yourself what works here.
Watch your niche properly. Search the terms your customers would search and watch results all the way through, not a swipe past. A full watch is a strong signal, a two-second bounce is a weak one. Within days your feed becomes your competitors, which makes it a permanent research tool.
Comment in your niche, usefully. This is the most underrated part of a warm-up and the one that produces something the same day. Leave genuine, actually helpful replies on videos in your category. Not "great video", not a pitch. An answer to the question in the comments that you happen to know cold because you built a product in this space. Operators report profile visits converting to followers at a few percent from this alone, and it costs nothing. It is also the one warm-up activity that is indistinguishable from being a useful person, because it is being a useful person.
Save what held you. Not the videos with the biggest numbers, the ones you personally watched to the end. Write down why: the first line, the first cut, whether something visibly happened on screen inside two seconds. Study shape, not topic. Where the hook ends, when the payoff lands, whether it works on mute, how it closes. Shape transfers to your product. Topic does not.
Finish your profile before the first post. Picture, one plain line saying what the product does, link in place. Otherwise everyone who liked a video taps through to an empty page and leaves.
What actually makes the video work
Everything above is preparation. This is the part that decides the outcome, and it is where most founders spend the least time.
The first three seconds
Completion, rewatches, shares and comments all cascade from one moment, and it is the opening. This is where the real work is, and it is where most founders spend the least time.
The openings that work tell the viewer the payoff before they have time to leave, and they let the right viewer self-select. A problem-first opening does both: name a specific pain and the people who have it stay while everyone else goes, which is exactly what you want, because a small well-matched audience finishing the video beats a big mismatched one bouncing.
Some shapes worth stealing:
- Problem-first. "If you are losing three hours a week to [specific thing], here is why." The specific thing does the filtering.
- Result-first. Open on the finished state, then explain how. The payoff is the hook rather than the reward for staying.
- Warning. "Do not post this yet." Cheap, and it works because it implies a mistake the viewer might be making.
- Test result. "I tried three versions of this. One worked." People stay for the answer.
Write the opening as a separate job from making the video. It is not the intro to the content, it is the reason anyone sees the content.
One recording, many openings
The highest-leverage habit in this whole playbook: you do not need five times the content, you need five times the angles.
Take one clip of your product doing something and cut it four ways, each with a different opening and caption aimed at a different group. Same footage. One version speaks to people drowning in admin, another to people who just want the thing to be faster, another to people who have tried three competitors. Each variant gets tested against a different pocket of the audience, and you find out which framing lands without shooting anything new.
This is also why the raw material matters more than the schedule. Four openings over good footage is four real shots on goal. Four openings over a shaky, unreadable capture is four wasted ones, which is the case for getting the recording right before any of this.
Founder on camera beats brand polish
The consistent pattern across app accounts that grow is that the person who built the thing is in it, using it, on an ordinary phone. Polish reads as an ad and gets filtered on sight.
The clearest version of this is TripBFF, a solo-travel app built by Isabella Comellini and Ethan Brimhall. Comellini films herself actually travelling and actually using the app: where she goes, who she meets, what happens. Ordinary phone footage, no production, and the app reached hundreds of thousands of downloads inside a year with no ad spend. The strategy is not a trick. She is the target user, so filming her real use of it is automatically the content.
Most founders can do a version of this. You built the product because you had the problem. Show yourself having the problem and solving it.
Measure the right number
Founders quit at day 30 because they are watching follower count, which is the number least connected to whether this is working.
Watch-through tells you if the video works. Signups tell you if the account works. A video with 12,000 views and 80 signups is a better outcome than one with 400,000 views and 10, and you will only notice that if you are tracking the second number.
The same logic says do not queue a month of content at a new account. Thirty scheduled posts commit you to a format no viewer has tested. If posts one through five fail for one fixable reason, a full queue means you learn it on post thirty. Post a few, give each a couple of days, read watch-through, then batch and schedule once something works. Four to six times a week, sustained, beats thirty in week one and silence in week three.
The one warm-up that backfires
Following fifty accounts in an afternoon, liking a hundred videos in a burst, running the routine from a script, or paying for engagement to clear the first test is a different activity from everything above. It is against the community guidelines, and burst behaviour on a fresh account is the exact pattern automated detection looks for. A real founder with one real account can get restricted for doing an impression of a bot, which is a stupid way to lose an account. The same goes for running content through a stable of pre-warmed accounts, which is why we do not build that and do not intend to.
Warm the account, then win the first three seconds
Two weeks in your category, commenting usefully, profile finished, and you have done the warm-up worth doing. It is cheap and it probably helps. It is not the lever.
The lever is the opening, and the fact that something real visibly happens on screen. Get footage of your product doing something, cut four openings over it, post, read watch-through, keep the one that held people.
That last loop is the repetitive part, and it is the part worth handing off. ClipMyApp turns one recording into a batch of openings and formats you approve, scheduled to your own account, and it will never pretend to be you anywhere else. The plans are here if that is the split you want.