How Does the TikTok Algorithm Work?

TikTok's algorithm recommends each video to the people most likely to watch it, based on what they have watched, liked, shared, commented on and searched for, plus information about the video itself. Follower count is not a direct factor, which is why a video from a small account can still reach a large audience.

Everything below comes from what TikTok has published about its For You feed. TikTok does not share the exact formula, so treat anything more precise than this with caution.

What signals does the TikTok algorithm use?

TikTok groups the inputs to its recommendation system into three types:

1. User interactions. The videos someone likes or shares, the accounts they follow, the comments they post and the content they create. TikTok's Community Guidelines also mention searches and current trends.
2. Video information. Details such as captions, sounds and hashtags.
3. Device and account settings. Language preference, country setting and device type. TikTok says these get less weight, because people don't actively choose them as preferences.

New users are asked to pick categories of interest, such as pets or travel, to seed their first feed. People who skip that step start with a general feed of popular videos.

Which signals count the most?

TikTok says it weights each signal by how strongly it shows interest. Its own example: finishing a longer video is a strong signal, while viewer and creator being in the same country is a weak one.

That corrects a common myth, which this page used to repeat: that TikTok shows a new video to people nearby first. Location is one of the weakest signals TikTok lists. Interest is what drives distribution.

Does follower count matter?

Not directly. In TikTok's words, "neither follower count nor whether the account has had previous high-performing videos are direct factors" in the recommendation system.

Popular accounts can still get more views, because more people follow them. But a brand starting from zero is not locked out of the For You feed. Each video is judged on how the people who see it respond.

What keeps a video out of the For You feed?

TikTok does not recommend every video. According to its own explanations:

• New or under-review videos may not be eligible yet.
• Duplicated content, spam and content the viewer has already seen are not recommended.
• Content unsuitable for a broad audience can stay on the platform but be marked ineligible for the For You feed. TikTok's Community Guidelines list these categories in "FYF Ineligible" sections.
• Repetitive content that may lead to a negative experience, such as extreme dieting or sadness, is spread out rather than shown again and again.

TikTok also notes that a video with few views "doesn't necessarily mean it broke a rule".

Why does the feed feel varied?

TikTok says it builds in diversity on purpose, to avoid a filter bubble. The feed generally won't show two videos in a row with the same sound or from the same creator, and it sometimes adds videos outside a person's usual interests. For a brand, that means a video can reach people who have never engaged with your category before.

How can users change what they see?

Viewers have several controls, and they shape which audiences your videos reach:

• Not Interested on a video, and options to hide videos from a creator or with a sound.
• Keyword filters, including AI-powered Smart Keyword Filters that also catch similar words.
• Manage Topics, added in 2025, which lets people turn more than ten popular topics up or down.
• Refresh, which resets the For You feed.

In the EU, TikTok has offered a non-personalised For You feed since 2023 to meet the Digital Services Act. People who switch it on see videos that are popular in their region and language instead of a profile-based feed.

Is the algorithm different in the US?

It is changing. In January 2026 TikTok set up a US joint venture with Oracle, Silver Lake and MGX as managing investors to comply with US law. TikTok said the joint venture "will retrain, test, and update the content recommendation algorithm on U.S. user data", secured in Oracle's US cloud. How far US recommendations will drift from the rest of the world is not yet clear, so check your own results by market.

What does this mean for marketers?

TikTok's published signals point to a few practical rules for brand accounts:

• Earn the full watch. Completion is one of the strongest signals TikTok names. Put the point of the video in the first seconds and cut anything that makes people scroll.
• Make the topic obvious. Captions, sounds and hashtags help TikTok understand a video. Use words your buyers would search for, not generic tags.
• Post original work. Duplicated content is not recommended, so reposting the same clip across accounts works against you.
• Judge each video on its own. Because follower count is not a direct factor, one strong video can outperform months of average ones. Look at which videos hold attention and make more like those.
• Don't buy engagement. TikTok treats content that artificially inflates traffic as spam.

For campaign ideas, see effective TikTok marketing strategies and how to gain brand exposure with TikTok marketing. To compare performance across channels, read how to measure social media marketing success, and browse more social media tools.