How does the TikTok For You algorithm work?

TikTok says it ranks videos using your interactions, video information such as captions and sounds, and device settings. It then delivers a feed unique to each person.

Does TikTok video length affect the algorithm?

TikTok's recommendation post gives no ideal length. It does name finishing a longer video as an example of a strong signal, so any exact seconds figure you read elsewhere is not coming from that post.

Does follower count matter on TikTok?

TikTok says neither follower count nor previous high-performing videos are direct factors in recommendations. Bigger accounts still tend to get more views because more people see their videos.

Is the YouTube Shorts algorithm different from TikTok's?

Both personalise feeds, but YouTube lists watch history, search history and satisfaction surveys, and says relative watch time is broadly more important for short videos. TikTok lists interactions, video information and device and account settings.

Will clips from someone else's video be recommended on TikTok?

TikTok says content with unoriginal or reused material without anything new is ineligible for the For You feed. Adding your own edits, commentary or context matters.

Why is my TikTok video not getting views?

TikTok says low views don't necessarily mean a rule was broken, and a video can still be found through search or your profile. TikTok's analytics can show whether any videos were made ineligible for recommendation.

Guide

TikTok algorithm explained: what TikTok says, and what it means for clips

By Rakshit · Updated October 11, 2026

TikTok says its For You feed ranks videos using user interactions, video information such as captions, sounds and hashtags, and device and account settings, which get lower weight. It says follower count and past hits are not direct factors, and that finishing a longer video is a strong signal. Content reused without anything new is ineligible for the feed.

What TikTok itself says about For You

Most TikTok algorithm advice is secondhand. This section sticks to what TikTok published in its newsroom post on how it recommends videos, plus its Community Guidelines. TikTok describes the For You feed as a stream ranked for each person, with no single feed shared by everyone.

Three groups of factors: user interactions (videos liked or shared, accounts followed, comments posted, content created), video information (captions, sounds, hashtags), and device and account settings (language, country, device type).
Device and account settings receive lower weight than the other data, because users don't actively express them as preferences.
Strong signals count for more than weak ones. TikTok's example of a strong indicator is whether someone “finishes watching a longer video from beginning to end”; its weak example is viewer and creator sharing a country.
TikTok says “neither follower count nor whether the account has had previous high-performing videos are direct factors.” A bigger account still tends to get more views simply because more people see it.
The feed generally won't show two videos in a row with the same sound or by the same creator, and it does not recommend duplicated content or content already seen.
Videos just uploaded or under review, and spam that tries to artificially increase traffic, may not be eligible for recommendation.

TikTok's For You eligibility page adds that a video that doesn't reach the feed can still be found through search or the creator's profile, and that low views don't necessarily mean a rule was broken.

TikTok's statements versus claims you'll see elsewhere

Here is how some common claims line up with what TikTok's pages I read actually say. “Not stated” means I did not find it, not that it is false.

Checked against TikTok's own pages on 2026-10-11
Claim you'll see	On TikTok's pages?
The For You feed uses watch behavior, likes, shares, follows and comments	Yes. TikTok lists interactions such as likes, shares, follows, comments and content you create, and names finishing a longer video as a strong signal.
Captions, sounds and hashtags help the system understand a video	Yes, as “video information”.
Follower count controls reach	TikTok says it is not a direct factor, though larger accounts tend to get more views.
Reposted or reused material is penalised	Yes. Content with reused material and nothing new is ineligible for the For You feed, and the originality policy lists content “largely repurposed from another source without adding any creative edits.”
There is an ideal video length in seconds	Not stated in the recommendation post. The length figure TikTok does publish is for Creator Rewards, which asks for original videos over 1 minute. That is a payout rule, not a ranking rule.
The first 3 seconds decide everything	Not stated for organic videos in the pages I read. TikTok's ad best-practice page advises a hook in the first 6 seconds and a content proposition in the first 3, which is ad advice.
There is a best time to post	Not stated in the pages I read.

What that means for the clips you cut

TikTok gives no clip length, and it names finishing a longer video as an example of a strong signal. Turn that into choices:

Cut for completion, not for a target duration. A clip viewers watch to the end matches TikTok's strong-signal example better than one they leave halfway, so length matters only as far as it helps people finish.
Start on the thought, not the lead-in, so there is less to sit through before the point.
Write captions and a description that say what the clip is about; TikTok reads them as video information.
Don't count on an account's history. TikTok says previous high-performing videos are not a direct factor.
Add something. A caption, a cut, a stronger start or commentary gives the clip a reason to exist beyond the source.
Keep the source's watermark and logos out of the picture. TikTok's originality policy counts someone else's visible watermark as unoriginal.

YouTube Shorts versus TikTok: the official differences

Both feeds are personalised, and both companies publish a short list of signals. The wording differs in useful ways.

TikTok newsroom and YouTube Help, read 2026-10-11
Topic	TikTok says	YouTube says
Main signals	Interactions, video information, device and account settings	Watch history, search history, subscriptions, likes, dislikes, Not interested, satisfaction surveys
Account size or history	Follower count and past hits are not direct factors	YouTube says its system finds videos for the audience rather than promoting them, and ranks videos by performance and relevance to that audience
Watch time and short videos	Finishing a longer video is a strong signal	Absolute and relative watch time are signals; broadly, relative watch time is more important for short videos and absolute watch time for longer ones
Tags and hashtags	Hashtags count as video information	Tags are described as not important, used mainly to correct common misspellings
View counts	Not defined on the pages read	A view counts when playback starts, since 24 August 2026 across formats

The practical difference for a clipper: YouTube names relative watch time as the one that matters more for short videos, while TikTok names finishing a longer video as a strong signal and lists captions, sounds and hashtags as video information. Neither page gives a target length. For the YouTube side in more detail, see how the YouTube Shorts algorithm works.

Where KeyMo fits

KeyMo does not predict TikTok performance and makes no claim about views. What it does is cut clips from a long video: each clip gets a score out of 10 and a one-line reason, and only clips scoring 7 or higher are kept. You can pick a clip length (Auto 15 to 90 seconds, Short 20 to 40, Medium 40 to 60, Long 60 to 90), 12 caption styles with word-by-word captions, and an automatic face-tracking 9:16 reframe. Output is 1080x1920 with no watermark. See pricing; there is no free plan.