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:
- Initial batch: Your video is shown to a small group of viewers, often non-followers interested in the topic
- Performance evaluation: TikTok measures watch time, completion rate, likes, comments, shares, and follows
- Expansion or suppression: If the video performs above average, it's pushed to a larger batch (thousands). If not, distribution slows.
- 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.