The TikTok algorithm, in TikTok’s own words
What the For You feed ranks on, what makes a video ineligible, and what changed in 2025 and 2026.
First brand read free. No filming.
On this page
The short answer
TikTok’s For You feed ranks videos on three factors: User interactions (what you like, share, comment on, watch in full or skip), content information (sounds, hashtags, views, country) and user information. Interactions, including time spent watching, are generally weighted most. Follower count is not a direct factor, and unoriginal or reused videos are not eligible for the feed.
Updated
Work with it, step by step
- 01
Hook in 3 seconds
TikTok: “hook your viewers within the first 3 seconds”, with a question, a timer or a quick tease of the result.
- 02
Hold to the end
Change something on screen every few seconds. Finish rate and average watch time are what the feed weighs most.
- 03
Tell it who the video is for
Use the topics, hashtags and sounds of your community so the system can place it with the right people.
- 04
Post often enough to learn
“The more videos you post, the more opportunities you give the recommendation system to understand your content”, but quality beats a daily rush.
- 05
Make each video new
Reuse a format, not footage; add a real twist when you join a trend.
- 06
Read your analytics
Retention rate, average watch time and “watched full video” show which openings and lengths hold your audience.
What TikTok actually says
Three factors
User interactions, content information and user information. Device and account settings get “lower weight” because people do not choose them as preferences.
Watching weighs most
Interactions, “which may include the time spent watching a video, are generally weighted more heavily”. Finishing a longer video is a strong signal.
Followers are not a factor
“Neither follower count nor whether the account has had previous high-performing videos are direct factors”, though big accounts get views from their followers.
Performance earns reach
Videos go further “based on the recommendation system’s performance predictions and user engagement”.
Originals only
Content “is also ineligible for the FYF if it includes unoriginal or reused material without anything new”, and so is bait for likes, follows or gifts.
AI labels do not cost reach
Realistic AI content must be labelled; turning the label on “won’t affect the distribution of your video” if it follows the rules.
The sources
Only TikTok’s own pages, read on 2 October 2026 (EU versions where TikTok serves them). Where a source is old, its date is given. Percify is not affiliated with TikTok.
TikTok Help Center
How TikTok recommends content, plus Manage topics and AI-generated content.
Source: support.tiktok.com; quote: AI-generated content, read Invalid Date
- What it says
- Three factors; user interactions including watch time “generally weighted more heavily”
- On eligibility
- Realistic AI content must be labelled; the label does not affect distribution
- Date
- Help Center, read 2 October 2026
TikTok Newsroom
How TikTok recommends videos #ForYou, and the 2025 AI content update.
Source: newsroom.tiktok.com; quote: AI content update, read Invalid Date
- What it says
- Finishing a longer video is a strong signal; follower count is not a direct factor
- On eligibility
- New uploads, videos under review and spam may be ineligible
- Date
- 18 June 2020; 19 November 2025
For You feed standards
Community Guidelines: For You feed eligibility and integrity and authenticity.
Source: tiktok.com; quote: Integrity and authenticity, read Invalid Date
- What it says
- Ineligible videos can still be found in search and on the profile
- On eligibility
- Unoriginal or reused material and engagement bait are ineligible
- Date
- Effective 24 September 2026
TikTok Creator Academy
Video traffic and the For You feed, creator tips and analytics.
Source: tiktok.com; quote: Creator tips, read Invalid Date
- What it says
- Reach follows performance predictions and engagement; hook in the first 3 seconds
- On eligibility
- Join trends with “a unique take”
- Date
- Updated 18 August to 1 October 2026
How the For You feed behaves
It learns from people like you
Predictions “also consider interactions from others who appear to have similar interests”.
It mixes in the unexpected
You will see videos outside your interests on purpose, and usually not ones you have already seen.
Viewers can steer it
Not interested, Manage topics sliders, a feed refresh, and in the EU a non-personalised Popular feed. TikTok is testing a control for how much AI-generated content people see.
Ineligible is not invisible
A video kept off For You can still be found through search and the profile; “if a video isn’t getting many views, it also doesn’t necessarily mean it broke a rule.”
Repeat offenders drop out
Post unsuitable content repeatedly and your account and posts stop appearing in For You and get harder to find in search.
Original, labelled, daily
What the rules reward is a new video each time, made to be watched to the end. Percify makes one for you daily from formats that performed in your market, with your own face and voice, labelled as AI where TikTok asks.
Read next
Your questions, answered
Short answers to what people ask next.
It predicts how interesting each video is to you from your interactions (likes, shares, comments, full watches and skips), the video’s information (sounds, hashtags, views, country) and your user information. Interactions, including watch time, generally weigh most.
Not directly. TikTok says neither follower count nor past hit videos are direct ranking factors, though larger accounts get more views from their own followers.
Usually because viewers skip it early or do not finish it. It may also be ineligible for For You, for example reused material or engagement bait; ineligible videos can still be found in search and on your profile.
TikTok says turning on the AI-generated label does not affect distribution as long as the video follows its rules. Realistic AI content must be labelled, and viewers can choose how much AI content they see.
Yes: refresh your For You feed in settings to start again from popular content. It cannot be undone. You can also use Not interested and the Manage topics sliders.