TL;DR
- TikTok says follower count and past hit videos are not direct factors in For You recommendations; each video is judged by how its viewers react.
- Per TikTok's 2020 Newsroom post, finishing a longer video is a strong signal, while language, country and device settings carry lower weight.
- TikTok publishes no 200-view threshold. Just-uploaded or under-review videos may not be recommended yet, and ineligibility shows in TikTok Analytics.
- Reused content posted without creative edits, such as clips with another account's watermark, is ineligible for the For You feed under TikTok's Community Guidelines.
- TikTok for Business help says Spark Ads views, likes, comments, shares and follows are attributed to the original organic post.
The TikTok algorithm is the recommendation system that picks videos for each person's For You feed, based mainly on how that person behaves: how long they watch, whether they finish a video or swipe past, and whether they like, share, comment on or save it. It combines those signals with information about the video and the account's settings. TikTok says that follower count, and whether an account has had high-performing videos before, are not direct factors in recommendations, so each video is judged by the reactions of the people who actually see it.
This article covers ranking on the For You feed only, which is the channel where the system recommends videos without the user typing a search. Getting videos into TikTok search results is a separate subject. Every mechanism described here comes from TikTok's own explanations in its Help Center, Safety Center, Newsroom and Community Guidelines.
How does the TikTok algorithm work, and what does the For You feed pick videos from?
In its Safety Center guide to the For You feed, TikTok explains that the recommendation system works in stages. Once you know the stages, you can see where a brand's video can pass or drop out.
- Selecting videos: the system retrieves a large number of videos that have passed initial content moderation and are eligible for the For You feed.
- Prediction: it estimates the chance that the user will like, share, comment on or skip each video, based on signals such as the videos the user recently watched, finished, liked or skipped.
- Ranking: videos are sorted by those prediction scores, and the feed is built from the top-ranked ones.
- Similarity check: the system replaces videos that are too alike, for example ones that use the same sound, so the feed has variety.
- Recommendation rules: extra rules are applied at different points, such as making sure the feed includes videos from creators in the user's region.
Before a video can enter this sequence it goes through content moderation, which uses both automated systems and human moderators. TikTok says the recommendation system only picks from videos that have passed its safety checks. Videos that break the Community Guidelines are removed, and videos that break no rule but are not suitable for a broad audience are made ineligible for recommendation.
The signals the system uses, and how much each one weighs
In its TikTok Newsroom article on how it recommends videos on the For You feed (June 2020), TikTok groups the signals into three sets:
- User interactions: videos the user likes or shares, accounts they follow, comments they post and content they create.
- Video information: captions, sounds and hashtags.
- Device and account settings: language preference, country setting and device type. These weigh less because users do not actively express them as preferences.
TikTok explains that each signal is weighted by how strongly it indicates interest. Its own example: finishing a longer video is a strong signal, while the viewer and the creator being in the same country is a weak one. The newer Safety Center guide adds to the list of things the system learns from: how long a video was watched and whether the user finished or skipped it, who the user follows, and how other users with overlapping interactions engaged with content.
The same guide lists signals about the video itself: the time it was posted, the region it was posted from, the language setting of its author, its soundtrack, its length and its hashtags.
What is not a direct factor
TikTok states plainly that follower count, and whether an account has had previous high-performing videos, are not direct factors in the recommendation system. It acknowledges that an account with many followers may still get more views as a result of having built a larger audience, but the follower number is not fed directly into the calculation. For a brand with a new account, this means early videos can be recommended if the people who see them watch for a long time and react.
How new users start
When someone signs up, TikTok's Help Center says it may invite them to select categories of interest, such as pets or travel. Users who skip this step are first shown recent posts that are popular with others, and the initial set also reflects their location and language settings. From then on, their interactions act as signals, which is why TikTok says feeds remain unique to each person.
How much do watch time and completion rate matter?
Watch time, meaning how long viewers watch a video, sits inside the group TikTok weights most. Its Help Center says that for most users, user interactions, which may include the time spent watching a video, are generally weighted more heavily than content or user information. Finishing a longer video is the strong-signal example from TikTok's 2020 Newsroom post.
TikTok Studio's video analytics include metrics that check this directly:
- Average watch time: the average duration viewers spent watching the video.
- Watched full video: the share of viewers who watched to the end.
- Retention: a curve that shows the second at which viewers start to drop off.
- Traffic source: where viewers came from, so you can separate views from the For You feed from other sources.
The trap is comparing completion rate across videos of different lengths. A 10-second video and a 90-second video have very different bars for being watched to the end. A fairer comparison is between videos of similar length, or looking at the retention curve for unusually early drop-off.
What is the best time to post?
TikTok does not publish a single best posting time that works for every account. What TikTok does confirm is that the time a video was posted is one of the video signals the system takes into account, without saying how much weight it carries. The useful question is when a given account's viewers are active.
The tool that answers this most directly is the Viewers section of TikTok Studio analytics, which TikTok's Help Center says covers key metrics, demographics and activity times of an account's viewers.
A practical test: pick two of the busiest windows shown there, then alternate videos of a similar type and length between those two windows for two to three weeks. Compare Average watch time and the share of views coming from the For You feed, not just total views.
Why did my views drop, or why is my video stuck at 200 views?
TikTok has never published a fixed view threshold, such as 200 views, that a video must pass before it is recommended further. What TikTok does explain is why a video may not be recommended, or may be recommended less:
The video has not finished review, or was just uploaded
TikTok's 2020 Newsroom post says videos that were just uploaded or are under review may be ineligible for recommendation. A video whose views stay flat right after posting may still be at this stage. TikTok's Help Center adds that its safety teams may review certain content as it rises in popularity to help ensure it is suitable for everyone, on top of the moderation every video goes through.
The content is not eligible for the For You feed
TikTok's Community Guidelines include For You feed Eligibility Standards, which describe content that can stay on the platform but will not be recommended. The items most relevant to brands sit under unoriginal content. The current guidelines list two kinds as ineligible for the For You feed:
- reused or unoriginal content posted without creative edits, such as clips that show someone else's watermark or logo;
- low-quality or minimally edited content, such as short clips made only from GIFs.
Marketing teams that repost ad videos from other platforms with the watermark still on are directly at risk here.
The account posts ineligible content too often
TikTok says some accounts break no rules but still post a lot of content that is ineligible for the For You feed. In those cases TikTok may make the whole account and its content ineligible for the For You feed and harder to find. The impact is then not limited to one video.
Fake engagement and attempts to trick the system
TikTok does not allow the trade of services that artificially boost engagement or trick the recommendation system. Accounts that do this may be restricted, which can include limiting their ability to appear in top search results or in the For You feed. TikTok also prohibits automation tools or scripts designed to bypass its systems.
The system deliberately spreads variety
The feed generally will not show two videos in a row made with the same sound or by the same creator. TikTok typically does not recommend videos a user has already seen, and its 2020 Newsroom post says duplicated content is not recommended either. The system also tries not to recommend some kinds of content repeatedly even when they break no rule, such as extreme fitness or dieting content, or overgeneralised mental health advice. Brands in health and fitness should know this, because a video may be distributed less for variety reasons, not because it is a bad video.
Note that TikTok itself says a video with few views has not necessarily broken a rule. If a video was made ineligible for the For You feed or otherwise restricted, that information appears in the Analytics tool. The first step in fixing a views drop is therefore to open Analytics and check the video's status before concluding it has been suppressed.
How does the algorithm affect TikTok ads and Spark Ads?
The For You feed and the ad system connect at some points TikTok has explained, and at others it has not disclosed, so it helps to keep them apart:
- Spark Ads keep results on the original post: TikTok for Business help states that all views, comments, shares, likes and follows gained from boosting a video with Spark Ads are attributed to the original organic post, and stay there after the campaign ends. Spark Ads can use the account's own posts, or other creators' posts with their authorization.
- Commercial content must be disclosed: the Community Guidelines require anyone promoting a product, brand or business to use TikTok's content disclosure setting.
- LIVE sessions that send buyers off-platform: in markets where TikTok Shop is available, the guidelines state that TikTok reduces the visibility of LIVE content directing users to purchase products off-platform.
What TikTok has not disclosed is whether ad spend makes an account's other organic videos get recommended more. Plans should not count on that effect. A more grounded approach is to use For You data to pick the videos that already have strong watch time and retention and boost those as Spark Ads, since viewers have already shown they keep watching them, and the engagement from the ads accumulates on the same post.
Do hashtags still help?
Hashtags are still part of the video information the system uses, together with captions and sounds, as TikTok describes. But video information is only one of three signal sets, and viewer interactions indicate interest more strongly. Hashtags help the system understand what a video is about. They do not turn a video people swipe past quickly into one that gets recommended.
The other side, which brands often overlook, is that users can filter hashtags out of their feed. TikTok's Help Center says the Filter keywords feature lets users remove content containing specific words and hashtags, and in June 2025 TikTok Newsroom announced that users can also choose AI filtering that catches similar keywords, such as synonyms. Adding unrelated hashtags to ride a trend can therefore put a video into a group some viewers have already filtered out.
How can users shape their feed, and what should brands know?
Viewers can also steer their own feed. TikTok gives them several tools for it:
- Not interested: long-press a video and choose Not interested, and the system limits that kind of content.
- Filter keywords: filter words or hashtags the user does not want to see.
- Manage topics: see more or less of a topic. TikTok announced it in June 2025 covering more than 10 popular topics, and it does not remove a topic entirely.
- Refresh your For You feed: update what the For You feed recommends, then interact with popular content to help reshape it.
- Muted accounts: mute accounts the user wants to see less of, including their posts.
For brands, the practical meaning is that a video that sells too hard, or looks like an ad cut and pasted from elsewhere, risks being swiped past, and it can also lead a viewer to tap Not interested or mute the account, after which TikTok shows that viewer less content like it.
TikTok algorithm signals at a glance
| Signal or mechanism | What TikTok states | What a brand controls |
|---|---|---|
| Long watch time and finishing | Finishing a longer video is a strong signal | The opening, editing pace, a length that fits the content |
| Likes, shares, comments, saves | Interactions the system uses for prediction | Topics that invite conversation, a question at the end |
| Language, country, device | Weigh less because they are not expressed preferences | Language in the video and caption that matches the audience |
| Follower count | Not a direct factor | No need to wait for a follower base before posting quality videos |
| Originality | Reused or watermarked content is ineligible for For You | Re-edit for TikTok, never post watermarked files |
The TikTok algorithm in the Thai market
Some of these mechanisms affect brands marketing in Thailand directly. First, TikTok says the system considers the region where a video was posted, and has a rule that the feed includes videos from creators in the user's region. A brand that wants viewers in Thailand should post from an account and location that match the market, rather than relying on a single regional account posting from elsewhere.
Second, language. The device language setting is a low-weight signal, but TikTok also lists the language setting of the video's author among the video signals, and watch time depends on viewers understanding the video. Brands that market in two languages may need separate videos per language rather than one video for both audiences.
Third, selling. Thailand is a market where TikTok Shop is available, so the guideline that reduces visibility for LIVE content sending viewers to buy off-platform applies to live selling that sends people to an outside store. Brands that sell on LIVE should design the purchase path around this rule from the start, and use the commercial content disclosure setting whenever they promote a product.
FAQ about the TikTok algorithm
Does the TikTok algorithm look at follower count?
Not directly, because TikTok states that follower count and an account's history of hit videos are not direct factors in recommendations, even though accounts with many followers may get more views since more people see their videos at the start.
Does deleting a low-view video and reposting it help?
TikTok has not said that deleting and reposting helps a video get recommended more, and it states that duplicated content is not recommended. A better path is to open Analytics, check whether the video was made ineligible for the For You feed, and apply the lesson to a new video with a reworked opening.
Why does the same kind of video get different views on different days?
Because the recommendation system works viewer by viewer, results depend on how the first people who see the video react, and on when the account's viewers are active, which the Viewers section of TikTok Studio analytics shows.
Does the algorithm suppress business accounts?
TikTok has not said business accounts are shown less just for being business accounts, but commercial content must use the disclosure setting, and LIVE content that sends buyers off-platform in TikTok Shop markets gets reduced visibility.
If you want a brand's videos to work both on the For You feed and in ad campaigns, the Relevant Audience team can help plan everything from content to boosting. See TikTok advertising services, social media advertising, content marketing services, or influencer marketing for campaigns that run creators' videos as Spark Ads.
