The TikTok algorithm

TikTok does not show videos to your followers first. It shows every video to a small test audience, measures how that audience behaves, and expands or stops the video based on what it sees. Understanding that loop is the difference between guessing and diagnosing.

This topic breaks the recommendation system into its parts: the signals it reads, how a video moves between audience pools, what suppression really looks like, and how to measure any of it against your own niche rather than against anecdotes.

The TikTok algorithm is the recommendation system that powers the For You Page (FYP). It reads signals from every view to decide which videos to show to each individual user. Unlike social media algorithms that primarily surface content from people you follow, TikTok's algorithm is designed to surface content from anyone — making it possible for new creators to reach millions of viewers.

How the Algorithm Ranks Videos

TikTok has publicly shared that its algorithm considers three main categories of signals:

1. User Interactions (Strongest Signal)

  • Watch time and completion rate: The #1 factor. If viewers watch your video to the end (or rewatch it), TikTok interprets this as high-quality content.
  • Likes, comments, and shares: Each interaction signals engagement, but comments and shares carry more weight than likes.
  • Follows from the video: If someone watches your video and then follows you, that's one of the strongest positive signals.
  • "Not interested" signals: Scrolling past quickly, hiding a video, or selecting "Not Interested" tells the algorithm to reduce similar content.

2. Video Information

  • Captions and text on screen: TikTok reads and categorizes your caption text
  • Sounds and music: Using trending sounds can boost initial distribution
  • Hashtags: Help TikTok categorize content, though they're less influential than watch time
  • Effects and filters: Trending effects may receive a small distribution boost

3. Device and Account Settings

  • Language preference
  • Country setting
  • Device type

These are weaker signals — they influence content selection but don't determine virality.

The Batch Testing System

Every video on TikTok goes through a batch testing process:

  1. Initial batch: Your video is shown to a small group of viewers, often non-followers interested in the topic
  2. Performance evaluation: TikTok measures watch time, completion rate, likes, comments, shares, and follows
  3. Expansion or suppression: If the video performs above average, it's pushed to a larger batch (thousands). If not, distribution slows.
  4. Viral escalation: Videos that continue performing well get pushed to progressively larger audiences — potentially millions

This is why a video can go viral days or even weeks after posting. TikTok continuously re-evaluates content, and a video that gets a surge of engagement can re-enter the distribution cycle.

What the Algorithm Does NOT Consider

TikTok has said (in its 2020 explainer on how it recommends videos) that follower count and past high-performing videos are not direct factors. In practice:

  • Follower count: A 10-follower account has the same initial chance as a 10-million-follower account
  • Previous video performance: Each video is evaluated independently
  • Account verification status: Verified badges don't boost algorithmic reach
  • Using #fyp or #foryou hashtags: These have no special algorithmic effect

How to Work With the Algorithm

  • Optimize for watch time. Make the first second compelling. Use hooks, visual surprises, or on-screen text that creates curiosity.
  • Keep videos short enough to be rewatched. A 15-second video watched 3 times sends a stronger signal than a 60-second video watched once.
  • Post consistently. More videos give the algorithm more chances to find a winner; see how often to post on TikTok for the data.
  • Ride trends early. The algorithm favors content using trending sounds and formats. Tools like Virlo help you identify trends in the early stages, before they peak.
  • Encourage comments. Ask questions, make slightly controversial takes, or leave "mistakes" that viewers will comment on. Comments are a high-weight engagement signal.

More questions about the TikTok algorithm

Does posting time matter for the algorithm?

Yes, but less than you might think. Posting when your audience is most active helps your video perform well in its initial batch test. Check your TikTok Analytics to see when your followers are most active. However, a great video posted at a "bad" time can still go viral — the algorithm will find its audience eventually.

Does deleting and reposting work?

Sometimes. If a video gets stuck in the initial batch with poor performance, deleting and reposting can give it a fresh start with a new test audience. However, doing this frequently may trigger spam detection.

Why do some videos go viral weeks after posting?

TikTok continuously re-evaluates content. If a video suddenly gets engagement (someone shares it, or a similar trend takes off), the algorithm can re-enter it into the distribution cycle and push it to new audiences.

Key terms

Common questions

Does TikTok show my video to my followers first?
No. TikTok shows a new upload to a small test audience drawn from people likely to be interested in the topic, not from your follower list. Follower count influences the size of later audience pools, not the first one, which is why an account with few followers can still reach millions on a single video.
What signals does the TikTok algorithm weigh most?
Completion and re-watch rate first, then shares, then comments and likes. All of them are read relative to the video's length and to how similar videos in the same topic perform, so a 12-second video and a 90-second video are not judged against the same bar.
Is shadowbanning real?
Suppression is real; the word shadowban usually is not. Videos get limited distribution for concrete reasons — a community-guideline flag, unoriginal or reposted content, a sound with a rights problem, or simply a weak test-audience result. Account-wide invisible penalties are rare, and a sudden drop across every video almost always has a cause you can find.
How long does it take for a video to take off?
Most videos settle within the first few hours, but the loop can restart at any point: TikTok re-tests older uploads when a sound, hashtag or topic becomes active again. A video that did nothing for a week and then climbs is normal behaviour, not an anomaly.

Put it to work

Stop guessing what the algorithm wants

Virlo tracks what is actually breaking out across TikTok, Reels and Shorts, and tells you why. Start a 7-day free trial.