Your Growth OS for Apps. You build. Layers gets you users.
Everywhere.
Joined April 2008
- Tweets326
- Following51
- Followers3.1K
- Likes426
Pull up your top 3 posts from the last few months. Odds are they're the same one or two ideas repeated.
Take those ideas and make five more, each with a different hook, clip, or example. This is the kind of grunt work an agent can do for you.
Either way, they'll beat a brand new idea almost every time.
When a post does well, most people screenshot the view count and move on.
Something should be logging why it worked. One line like "short demo, no talking, result shown in the first 3 seconds."
Whether that's you or an agent, that line is the asset. The post was just one example of it.
The exact words in your best post matter less than you think.
What worked was the angle underneath it. The promise, the setup, the reason someone stopped scrolling.
Change the words, keep that, and it works again.
"This post doubled my views."
From 40 to 80.
Numbers that small are mostly luck. You can't learn much from them. Post consistently and wait until the numbers are big enough to mean something.
Comparing yourself to creators with 50x your audience makes you feel bad and teaches you nothing.
Watch your own baseline instead. Beat your normal, then beat it again.
When you're small, that's the whole job.
If you post on TikTok: 21 rules everyone repeats, every one tested against 23 million videos.
Most are myths.
When you post is noise. What you post is signal.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok, these numbers come from 23 million TikToks and 13.3 trillion views.
Count each view as a single second: 421,000 years of watch time. About 5,770 human lifetimes, end to end. And one second per view is the most conservative math possible.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok, here's the whole report in four numbers:
Best hour instead of worst hour: about 0
1080p instead of sub-720p: +7% views
No hard sell in the caption: +12%
9:16 instead of square: +14%
Spend your energy on the bottom three.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok, here's what we couldn't test honestly and won't pretend we did: hashtag counts. Trending vs. original sounds. Shadowbans. Delete-and-repost. First-hour windows.
Those need data only the platform has. Anyone selling you certainty on them is guessing.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
You already have months of data on what your audience responds to. It's sitting in your own post history. Almost nobody reads it back. Nearly every week turns into another guess at a question you already answered.
If you post on TikTok, three upload settings measurably change your views:
1080p instead of sub-720p: +7%. (4K adds nothing.)
9:16 instead of square or landscape: +14%.
No hard sell in the caption. A link or "DM me" costs you 12%.
Sell in the video. Use the caption for the hook.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok and a video got huge views with almost no likes, you're not about to be punished for it.
One big hit in ten is hollow: massive reach, under a 2% like rate. Folk theory says you pay for it next time. You don't. That next post does slightly better than usual.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok, saves are the button that predicts your winners.
Your most-saved videos outrank your least-saved by about 18 places out of 100, nearly double the views. Saves pull twice as hard as shares. Likes barely register, and comment-heavy videos rank lower.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
Every post you make starts from zero.
You work out what your audience likes, forget it, then work it out again next week.
You'd never run the rest of your business that way.
If you post on TikTok, "delete it if it doesn't pop" is advice almost nobody actually follows.
We tracked a million deleted videos. Fewer than 2 deletions in 100 happen in the first week. Creators clear out year-old posts, and the ones they clear out are their average performers.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok and you're watching your likes, you're watching the wrong button.
Per viewer, a creator's biggest hits collect the same number of likes as their average post. What separates them is shares: 17 for every 10 an ordinary post gets.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
You made a post that popped once.
Could you make another one on purpose?
Most people can't. They never worked out why the first one landed, so they can't do it again. Your best post was mostly luck.
If you're starting a TikTok account from zero: you don't need followers to get views.
A typical video from an account under 1,000 followers reaches more people than the account even has. At a million followers it's one viewer per 70. Reach tracks what the feed thinks people want to watch.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok, "200-view jail" isn't real. Nothing is holding your video anywhere.
View counts of low performers form one smooth curve peaking around 100-150 views, with no cliff at 200. Flops cluster there because that's simply where the curve peaks.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…
If you post on TikTok and you're afraid going off-topic will confuse the algorithm: it won't.
Across 7,540 creators who post about more than one thing, videos outside their main lane land within half a place of videos inside it. The penalty isn't in the data.
xitter.cf/layers/status/20900820…
We analyzed 23M TikTok videos and 13.3T views.
Frankly, a lot of social media advice is bullshit.
Best time to post: basically noise.
Don’t post too much: wrong.
Keep it short: wrong.
Stay in your niche: barely matters.
Take a break and the algo punishes you: nope.
21 rules tested.
When you post is noise. What you post is signal.
layers.com/research/?utm_sou…