How the TikTok For You algorithm works in 2026

TikTok explains its For You feed in more detail than most platforms. Its help center, its Safety Center and its newsroom each describe parts of the system, and they agree on the shape. This guide puts those pieces together as of October 2026: how a video gets picked, what TikTok predicts about you, what it weighs most, and what keeps a video out entirely. When TikTok hasn’t said something, we leave it out.

What does TikTok say the For You feed is built from?

TikTok’s Safety Center page on the recommendation system lists five steps.

The predictions are the heart of it. TikTok combines the scores for each predicted interaction into “an overall score for each video, with higher scores signifying a closer match to your preferences.”

Which interactions does TikTok predict?

The Safety Center appendix lists the likelihood that you will:

• like, share, comment on, or mark a video “Not Interested”
• follow the video’s author or interact with their profile
• finish, skip, or favorite a video
• spend a certain amount of time viewing it
• tap the video’s soundtrack

Every item is about what a viewer does after the video starts. None is “this video already has a lot of views.” That is the most important point for anyone thinking about paid numbers. If you buy TikTok views, the count on the video goes up, which can change how the next viewer reads it. The views themselves don’t finish, share or follow, and TikTok’s guidelines say it removes inflated signals when it detects inauthentic metrics. TikTok’s current Community Guidelines also say it doesn’t allow “the trade of services that artificially boost engagement or trick the recommendation system.” A video still has to hold real viewers.

What does TikTok weigh most heavily?

TikTok’s help center groups the factors into three types and is unusually direct about their weight.

Then it says: “For most users, user interactions, which may include the time spent watching a video, are generally weighted more heavily than others.”

TikTok’s 2020 newsroom post made the same point with an example. Whether a viewer finishes a longer video from beginning to end is a strong indicator of interest, so it gets more weight. Whether the viewer and the creator are in the same country is a weak one.

Notice “number of video views” sits in content information, the lighter group. Views are a piece of context. Watch behavior is the signal.

Does follower count matter on TikTok?

Not directly. TikTok’s 2020 explainer said: “While a video is likely to receive more views if posted by an account that has more followers, by virtue of that account having built up a larger follower base, neither follower count nor whether the account has had previous high-performing videos are direct factors in the recommendation system.”

This is why a new account can have a video take off and why a large account can post a flop. Each video is scored on how viewers respond to it.

Followers still matter in other places. The Following feed shows videos from accounts people follow. Account recommendations look at “number of followers and views an account has.” And features such as LIVE have follower minimums, covered in how to go live on TikTok.

What information about the video itself counts?

The Safety Center lists the video signals TikTok takes into account:

• the time the video was posted
• the region it was posted from
• the language setting of the author
• the soundtrack
• the length of the video
• the hashtags associated with it

These help TikTok understand what a video is and who might want it. They don’t decide whether it spreads. A well-tagged video that people skip will still stall.

Length is worth a note. TikTok’s “Why this post” panel can say a video was shown because “this video is longer and you seem to enjoy longer videos.” So length isn’t good or bad in itself. It matters relative to what each viewer tends to watch.

What keeps a video out of the For You feed?

This is where most “my views died” cases come from. TikTok separates what is allowed on the app from what it will recommend.

Moderation first. Every uploaded video goes through automated and human moderation. Videos that break the Community Guidelines are removed. Videos “not suitable for a broad audience” are made ineligible for recommendation. Only videos that pass go into the pool.

FYF eligibility standards. TikTok marks certain content as “FYF ineligible” throughout its Community Guidelines. It says the feed is meant for “everyone from teenagers to great grandparents.”

Age of the creator. “Any content created by an account aged under 16 is not eligible for recommendation to the FYF.”

Unoriginal content. “Content is also ineligible for the FYF if it includes unoriginal or reused material without anything new.”

Fake engagement. “If we detect accounts or content with inauthentic metrics, we’ll remove fake likes, followers, or other inflated signals. Content that tries to manipulate people into giving gifts or inflating likes or follows isn’t eligible for the FYF.”

New or under review. TikTok’s 2020 post said “videos that have just been uploaded or are under review” may be temporarily ineligible.

Deceptive accounts. TikTok says accounts that mislead or try to manipulate the platform can be restricted, including limits on appearing in top search results or in the FYF.

TikTok also adds a reassurance worth repeating: “If a video isn’t getting many views, it also doesn’t necessarily mean it broke a rule.” If a post is ineligible, TikTok shows the reason in the post’s analytics and lets you appeal. Our guide to TikTok views stuck at 200 walks through those checks one by one.

How do other viewers’ choices affect what you see?

TikTok doesn’t only learn from you. Its help center gives a simple example of what is usually called collaborative filtering: if User A likes videos 1, 2 and 3, and User B likes videos 1, 2, 3, 4 and 5, the system may predict that User A will also like videos 4 and 5.

For creators, that means your audience is defined by overlap. If the people who finish your videos also finish videos from a certain group of creators, TikTok will tend to show your next video to fans of that group. It explains why a niche account can find its audience quickly: a few strong overlaps tell the system a lot.

It also explains why a sudden change of topic can stall an account. The overlap that worked for cooking videos says little about who wants your gym videos, so the system has to learn again from scratch.

How does TikTok handle brand-new viewers?

A new user has no history, so TikTok starts somewhere else. According to its help center, new users may be invited to pick interest categories such as pets or travel. Those who skip that step see recent posts that are popular with others. The first set also leans on “popular content appropriate for a broad audience” and on the person’s location and language settings.

From there, every like, comment, replay and skip narrows the feed. This is one reason broad-audience eligibility matters so much. A video that is ineligible for a broad audience can’t appear in those early, general feeds, which are where many people first meet new creators.

Why can a video stall halfway through taking off?

Sometimes it’s a safety review. TikTok’s help center says its safety teams “may review certain content as it rises in popularity,” and that content can be put under review when it is uploaded, when it gains popularity, or when it is reported. A review ends one of three ways: the video stays eligible, it is made ineligible for the For You feed, or it is removed.

If a video slows sharply after a strong start, check its analytics for a notice before assuming viewers lost interest. If there is no notice, the more likely explanation is simpler: the wider audience it reached skipped it more often than the first viewers did.

How do the Following, Friends, LIVE and Search feeds differ?

The For You feed gets the attention, but TikTok runs several recommendation surfaces. Its help center describes what weighs most on each.

Search is the outlier. There, how well a video matches the query matters more than your history. That makes clear captions and on-screen text more valuable for videos people look up, such as tutorials, reviews and local recommendations.

How do viewers shape their own feed?

TikTok gives viewers several controls, and each one changes who your videos reach.

• Not interested. Long-press a video to see less like it.
• Refresh your feed. Viewers can reset For You; TikTok then shows popular videos while it relearns their taste.
• Manage Topics. Launched in June 2025, it lets people adjust how often more than 10 popular topics, such as Creative Arts, Travel, Nature and Sports, are recommended. TikTok says it won’t remove topics entirely.
• Word and hashtag filters. Viewers can filter words and hashtags out of For You and other feeds.
• Non-personalized feeds. In the EU, people can turn off personalized feeds, which turns For You into a Popular feed of videos from their region and beyond.

For creators, this argues for a clear topic. A viewer who turns Travel up is more likely to get travel videos that TikTok can confidently classify as travel.

Does TikTok try to keep feeds varied?

Yes, on purpose. The similarity check removes top-ranked videos that are too alike, for example two in a row with the same sound. TikTok’s 2020 post said the feed “generally won’t show two videos in a row made with the same sound or by the same creator,” and that it doesn’t recommend duplicated content, content you’ve already seen, or spam.

TikTok also says it interrupts repeated recommendations of some topics that are fine occasionally but could be harmful in bulk, such as extreme fitness or dieting.

There’s also a regional rule: TikTok says it ensures For You includes content from creators in your region.

Two creator takeaways. Jumping on a trending sound means competing with every other video using it for limited slots in each viewer’s feed. And a niche that TikTok limits for repetition, such as extreme diet content, will hit a ceiling no matter how well each video performs.

What can a creator actually test?

The published system points to a short list of things that move predictions.

Change one thing at a time and compare against your own recent median, not someone else’s viral post. More TikTok explainers are in our TikTok guides.