How the TikTok FYP algorithm works

The TikTok FYP algorithm works by scoring each video separately for each viewer. In TikTok's own description, the system reads three groups of signals: user interactions (videos you like or share, accounts you follow, comments you post, content you create), video information (captions, sounds, and hashtags), and device and account settings (language preference, country setting, device type). Those signals produce a prediction score, an estimate of how likely a specific person is to interact with a specific video. The highest-scoring videos are served to that person's feed. Follower count is not one of the inputs.

The For You feed is the personalized video stream TikTok opens on, and most of what appears in it comes from accounts the viewer does not follow. Creators shorten it to FYP. No two people see the same feed, because it is assembled from one account's own interaction history rather than a single global ranking.

Most advice about this system is written without a source. This page keeps two categories apart: what TikTok states in its published explanation of the recommendation system, and what is a third-party estimate.

What signals does the For You feed actually use?

TikTok groups the inputs into three categories. The table adds two columns that matter to a creator: what is known about each group's weight, and how much of it you can change.

The pattern there matters: the only group a creator fully controls is the one with the least documented influence. Captions and hashtags are yours to write, but their job is to classify the video, not amplify it. For a number that reflects account health rather than reach, calculating your engagement rate is a better starting point.

Which signals carry the most weight?

TikTok is explicit that signals are not weighted equally. Its published example contrasts two directly: a viewer who finishes a longer video from beginning to end is a strong indicator of interest, while the viewer and the creator being in the same country is a weak indicator that receives less weight. The same document states that device and account settings receive lower weight relative to other data points. That single comparison reorders most TikTok growth advice: time spent watching outranks the metadata around the video, so a one-minute video people finish beats a fifteen-second video people abandon halfway.

Does follower count decide whether a video reaches the FYP?

No. TikTok states that neither follower count nor whether an account has had previous high-performing videos is a direct factor in the recommendation system. A video from an account with no followers enters the same scoring process as one from an established account. This is the clearest structural difference between TikTok and platforms where distribution starts with your existing audience.

Follower count still helps indirectly. A large audience produces real watch signals faster in the first hours after posting, so the advantage comes from speed of feedback rather than the follower number being rewarded. This matters when someone offers to sell you followers: purchased accounts do not watch your videos, so they cannot produce that signal. The wider effect is covered in what bot followers do to an account.

How does a new video get its first audience?

Every published video is first shown to a limited set of viewers, and how that set responds shapes how widely it is distributed next. TikTok describes this through prediction scores rather than fixed stages: it estimates each viewer's likelihood of interacting, serves the highest-scoring videos, and updates the estimate as real data arrives. Reach is a series of small approvals, not one door that opens.

One honest correction here. The specific figures that circulate online, such as a first pool of a fixed number of viewers or a threshold percentage needed to reach a second pool, do not appear in TikTok's documentation. TikTok does not publish audience pool sizes. The mechanism is real; the numbers attached to it are third-party estimates. Any guide that gives you an exact pool size should be asked for its source.

Why does the feed show videos that similar viewers liked?

TikTok's system also learns from other people's behavior. In its description, if another viewer engaged with two of the same videos you engaged with, the system may predict you are likely to engage with a third video that person interacted with. This is how a niche opens up for a creator with no change in follower count. TikTok has also built a viewer-facing version of this into the app: its Why this video feature names reasons such as content you watch, like, share, comment on or search for, accounts you follow, and content posted recently or popular in your region.

Why does the feed avoid similar videos back to back?

TikTok states that it works to intersperse different types of content and avoids showing two videos in a row made with the same sound or by the same creator. It also does not recommend duplicated content, content you have already seen, or content considered spam. Diversity is a design choice, not a side effect. The practical consequence is that posting one format repeatedly in a short window does not multiply reach, which is part of why Instagram Reels reach drops for accounts that repeat a single template.

What makes a video ineligible for the For You feed?

TikTok publishes a separate document inside its Community Guidelines called the For You feed Eligibility Standards. Content covered by those standards is not removed, it is simply not recommended: the video stays reachable through your profile, through search, and through a direct link, but it is not distributed in the feed. That distinction explains the most common creator complaint, a video that was never taken down and never got views.

• Unoriginal content: reuploads without new or creative edits, including content carrying someone else's visible watermark or logo.
• Low quality content: extremely short clips, or videos made only from looping images.
• Engagement manipulation: content that tricks viewers into giving likes, follows, comments, gifts, or other engagement.
• Themes unsuited to a broad audience: sexually suggestive material, extreme dieting and fitness content, expressions of hopelessness, overgeneralized mental health information.
• Age and review status: content created by people under sixteen, and new uploads still under review.

Three claims that are not in TikTok's documentation

Does the number of hashtags change your reach?

Documented: hashtags are part of the video information signal group. Not documented: an ideal number of them. TikTok does not recommend three, five, or ten, and nowhere states that hashtag count increases reach. Their function is classification. Popular but unrelated tags classify the video wrongly, send it to viewers who will not finish it, and lower watch time.

Does posting at a specific hour trigger the algorithm?

Posting time is not listed among the signals TikTok names. It has an indirect effect: publishing while your audience is awake produces earlier reactions, so the system gathers real data sooner. That is not the same as time being a ranking criterion. A fixed evening hour may be a sound observation about one audience, but it is not a TikTok rule.

Is the TikTok shadowban real?

TikTok does not use the term shadowban and defines no penalty by that name. Two documented mechanisms produce the effect people describe. Content falling under the For You feed eligibility standards is not recommended; and when TikTok finds deceptive account behavior it may restrict the account, which it states can limit posting, appearing in top search results, or appearing in the For You feed. The label does not exist; the mechanism does.

A diagnostic order for a video that gets no views

Work through these in order. Each step maps to something TikTok documents, and the earlier steps are most often the cause.

1. Check eligibility first. Watermarks, reuploads, very short clips, and engagement bait wording are the common reasons a video is quietly not recommended.
2. Read average watch time before likes. Completion is the documented strong signal; likes are not described as comparable.
3. Cut the video to the real length of the idea. Padding twenty seconds out to forty only lowers completion.
4. Rewrite the caption in the words people search. It is a documented signal and the only text field that also works in TikTok's own search.
5. Use hashtags as topic labels. A few accurate tags classify better than popular unrelated ones.
6. Vary sound and format between posts. The feed spaces out similar videos for the same viewer.
7. Compare against your own earlier videos, not a public benchmark. TikTok publishes no threshold values, so any target rate you have read is an estimate.

What a views or likes service can and cannot do

Start with the rule. TikTok's Integrity and Authenticity policy states that it does not allow the trade of services that artificially boost engagement or deceive its recommendation system. When TikTok becomes aware of accounts with inauthentically inflated metrics it removes the associated fake followers or likes, and deceptive account behavior can lead to restrictions on posting, on search visibility, and on appearing in the For You feed. That is the documented cost, and no seller can remove it.

Then the technical limit. The FYP score is computed per viewer from real watching behavior. A counter raised from outside does not create the information the system needs, which is whether that viewer watched to the end. The number moves; the measured signal does not. What a purchased number does to people who land on your profile is social proof, which is a different thing from distribution. The tradeoff is set out in organic growth versus paid engagement.

Where TakiPlus Mini fits

TakiPlus Mini is a coin-based app at tmini.takiplus.com. You collect coins by watching rewarded videos, completing daily tasks, keeping a streak, playing mini-games, and inviting friends, then spend them on TikTok, Instagram, and YouTube service packages. Packages are fixed rather than free-quantity: you pick one, paste a profile or post link, and the server sets the amount, with a limit of one order per service per hour. The full flow is in the order placement guide.

The honest description of what that buys is narrow. It moves a counter. It does not produce the watching behavior the recommendation system measures, and it does not restore a count that falls after delivery. If a provider cancels an order your coins are returned in full and a partial delivery is refunded in proportion, but no re-delivery is promised once an order completes.

What reaches the For You feed is not your account but the reaction your video produces in viewers. Almost every signal TikTok documents describes watching behavior, and none of it can be bought directly.