70% Off Today |00h00m00s

Jan 24, 2026•9 min read

TikTok Algorithm 2026: What Drives FYP Distribution

Post author & contributors
Outfame Team
Outfame TeamContent Writer

The TikTok algorithm ranks each video by a prediction score: an estimate of how likely a specific viewer is to watch, finish, like, share, comment on or skip it. Videos with the highest score for that viewer are served to their For You feed, after a similarity check and a set of recommendation rules. Follower count and past hits are not direct factors, so a new account can reach the For You feed (FYP) without an existing audience, although a bigger following still brings more first views. (TikTok Newsroom, TikTok Transparency Center)

Most "algorithm guides" quote exact weights, such as "completion rate is 70% of the score." TikTok has never published weights like that, so this guide sticks to what TikTok itself documents and then turns it into practical steps.

How the For You feed picks a video, step by step

TikTok's Transparency Center breaks the recommendation system into five stages. Each one gives you something to work with.

  1. Selection. The system pulls a large pool of videos that passed content moderation and are eligible for recommendation. A video that is still under review, or that breaks the Community Guidelines, never enters the pool.
  2. Prediction. Machine learning models estimate the probability that this viewer will like, share, comment on, finish, skip, favourite or mark the video "Not Interested," follow the author, or tap the sound. Those probabilities combine into one score per video.
  3. Ranking. Videos are sorted from highest to lowest score for that viewer.
  4. Similarity check. If top-ranked videos are too alike (for example, the same sound), some are swapped out for variety.
  5. Recommendation rules. Extra rules apply, such as including creators from the viewer's region, and keeping out content from accounts under 16.

Source: TikTok Transparency Center, "Introduction to the TikTok recommendation system".

The takeaway: there is no single "algorithm score" for your video. Every viewer gets their own prediction, so a video wins by being the right match for a specific group of people, not by being good in the abstract.

The signals TikTok says it uses

TikTok groups its inputs into three kinds. The list below is TikTok's own, not a third party's guess.

Signal typeExamples TikTok namesWhat TikTok says about weight
User interactionsVideos watched, finished, liked, shared or skipped; accounts followed; comments posted; content createdNo category weight published; finishing a longer video is its example of a strong signal
Video informationCaptions, sounds, hashtags, video length, time posted, region and author's language settingNo weight published
Device and account settingsLanguage preference, country setting, device type and operating systemLower weight, because users don't actively choose them

Sources: TikTok Newsroom, "How TikTok recommends videos #ForYou" and the appendix of the Transparency Center page.

TikTok also says a strong indicator of interest, such as watching a longer video from beginning to end, receives greater weight than a weak one, such as viewer and creator being in the same country. That is as close as TikTok comes to ranking its signals, and it is the reason retention work pays off more than tweaks to settings.

What this means for each part of your video

Watch behaviour: finishing and skipping

Finishing, skipping and time spent viewing are all on TikTok's list of predicted actions. You cannot see the formula, but you can control what the viewer experiences:

  • Make the first seconds answer "why stay?" Skipping is a predicted action, so a video that loses people immediately gets a lower prediction for similar viewers. State the payoff or show the most interesting frame at the start.
  • Match length to substance. A video the viewer finishes is a stronger signal than a longer one they abandon. Cut anything that does not earn its seconds.
  • Check retention in TikTok's own analytics for each video and look at where the audience drops off, then fix the first drop-off point in your next upload.

Interactions: likes, shares, comments, follows

TikTok lists liking, sharing, commenting and following the author as predicted actions. It does not publish how much each one counts, so treat claims such as "a share is worth 5 points" as unverified. What you can do is make the video worth reacting to: ask a real question, give a tip someone would send to a friend, or finish with a clear reason to follow for the next part.

Negative signals

"Not Interested" is a signal TikTok both predicts and collects, and viewers can also hide a creator, hide a sound or add keyword filters, all of which shape future recommendations. (Transparency Center, Newsroom) Misleading hooks that promise something the video does not deliver tend to produce exactly these reactions, so keep the promise and the content aligned.

Captions, hashtags and sounds

Video information (captions, sounds, hashtags, length) is a named input. Practical use:

  • Write the caption in the words your target viewer would type or say, so the system can tell who the video is for.
  • Use hashtags that describe the topic rather than generic ones. TikTok does not publish an ideal count, so use the few that fit.
  • Choose a sound because it suits the video. TikTok's similarity check swaps out top-ranked videos that are too alike, for example ones using the same sound, and the feed generally won't show two videos in a row made with the same sound.

Account basics that matter

  • Followers are not a ranking factor. TikTok states that neither follower count nor previous high-performing videos are direct factors, though a bigger audience naturally means more first views. (Newsroom)
  • Duplicated content is not recommended. TikTok does not recommend duplicated content, content a viewer has already seen, or spam such as videos trying to artificially increase traffic. (Newsroom)
  • Region and language matter. Recommendation rules include content from creators in the viewer's region, so post in the language your audience uses. (Transparency Center)
  • Age eligibility. Content created by accounts under 16 is not eligible for the FYP. (Transparency Center)

Why a video can reach people who didn't ask for that topic

TikTok says it deliberately mixes in videos outside a viewer's stated interests, so the system can learn what is popular across wider groups and so viewers discover new creators. The feed also avoids showing two videos in a row from the same creator. (Newsroom, Newsroom update, Dec 2021) For you, that means a small account has a real chance of a surprise reach spike, and it also means one viral video does not guarantee the next.

A repeatable weekly routine

  1. Pick one audience. Decide who the video is for, then write the caption and on-screen text in that audience's words.
  2. Open with the payoff. Rework the first seconds until the reason to stay is clear.
  3. Publish several times a week and compare retention and shares across videos, not just view counts.
  4. Read the drop-off curve in TikTok analytics and fix the earliest drop in the next upload.
  5. Keep the topic consistent for a few weeks. TikTok's models learn who responds to your account, and a clear topic gives them clean data. (This is a practical recommendation, not a documented TikTok rule.)

If you are starting from a small account, the 1,000 followers on TikTok playbook walks through the first month in more detail, and why TikTok views suddenly drop covers what to check when reach falls. To reset a feed that has drifted off-topic, see how to refresh your For You feed.

Where Outfame fits

Content quality drives what TikTok recommends. If you also want people from your niche to find your profile, Outfame's TikTok growth service follows, likes and views the profiles of TikTok users who already watch creators like you. It needs only your username, never your password, and the first month on the Go plan is $49 (pricing). You can see how the setup works on the how it works page.

FAQ

How does the TikTok algorithm decide what goes on the For You page?

It estimates, for each viewer, how likely they are to watch, finish, like, share, comment on or skip each eligible video, combines those estimates into a score, and serves the highest-scoring videos after a similarity check and some recommendation rules. Signals include the viewer's own interactions, video information like captions, sounds and hashtags, and device and account settings.

Does follower count affect the TikTok algorithm?

Not directly. TikTok states that neither follower count nor whether an account has had high-performing videos before is a direct factor in recommendations. A larger following can still mean more early views, which gives the system more data.

Is watch time or completion rate the most important signal?

TikTok does not publish a ranking of signals, so no one outside the company can give you an exact weight. It does say that strong indicators of interest, such as watching a longer video to the end, count for more than weak ones, and it lists finishing, skipping and time spent viewing among the actions it predicts.

Do hashtags and captions still matter in 2026?

Yes. Captions, sounds, hashtags and video length are all listed by TikTok as video information used in recommendations. Use hashtags and caption wording that describe the video's topic accurately, rather than generic tags.

Can a new account go viral on TikTok?

Yes. Because follower count is not a direct factor and TikTok intentionally includes videos from outside a viewer's usual interests, a new account's video can reach a large audience if viewers respond well to it. Nothing guarantees it, though, and each video is judged on its own response.

Why did my TikTok views suddenly drop?

Common causes are a video that lost viewers early, a change of topic that confuses who the video is for, content that is ineligible for recommendation, or simply the mix-in of other videos in viewers' feeds. See our guide on sudden view drops for a checklist.

Sources

  • TikTok Newsroom, "How TikTok recommends videos #ForYou": https://newsroom.tiktok.com/en-us/how-tiktok-recommends-videos-for-you
  • TikTok Transparency Center, "Introduction to the TikTok recommendation system": https://www.tiktok.com/transparency/en/recommendation-system
  • TikTok Newsroom, "An update on our work to safeguard and diversify recommendations": https://newsroom.tiktok.com/en-us/an-update-on-our-work-to-safeguard-and-diversify-recommendations
  • Outfame pricing: https://www.outfame.com/pricing
  • Outfame TikTok growth service: https://www.outfame.com/tiktok-growth-service

Thinking about a growth service? Read Outfame reviews from 2,140+ verified customers.

YOUR INSTAGRAM DESERVES TO GROW

*without spending $2,000 per month.

No bots – ever · Cancel anytime