Skip to main content

How each platform's algorithm actually rewards content

Plain-English breakdown of how LinkedIn, X (Twitter), Instagram, Facebook, Threads, YouTube, Pinterest, Bluesky and TikTok currently reward content. Updated when our canary flags a shift across our user base, or when our research team bumps a profile version.

By Tayyab, Founder

Platform profiles

bluesky

v1

Bluesky is chronological with custom feeds and no algorithmic boost to game, so timing and reposts drive discovery. Post during active hours, be genuinely useful, and put links in the body. They are not penalized like on X.

Bluesky is chronological plus user-composed custom feeds, with no central algorithmic boost:
- There is no reach multiplier to game. A post is seen by followers in order and travels further only through reposts and custom feeds.
- Because there is no boost, timing matters: post during active hours (weekday mornings and
Reviewed 8/19/2026The full Bluesky breakdown

facebook

v1

Facebook prioritizes reels and native video over images and link posts. Keep posts short (40-80 chars), put links in the first comment, and write for shares.

Facebook's feed in 2025 is mostly driven by reels + native video:
- Reels and native video dramatically outperform image posts and link posts.
- Posts with external links in the body lose ~50% reach. Put the link in the first comment.
- Sweet spot post length is short: 40-80 chars. Anything past ~480 chars is truncated
Reviewed 5/31/2026The full Facebook breakdown

instagram

v1

Instagram weights saves and shares above likes and comments. Reels reach further than carousels, which reach further than single posts. Short captions with a strong first line outperform long ones.

Instagram's public algorithm framework (Mosseri 2024-2025):
- Saves are the strongest signal of "valuable content." Optimize captions to drive saves: lists, how-tos, before/after.
- Shares (DMs and shares to story) are second strongest. End on a shareable line.
- Reels get the largest reach lift; carousels beat single 

linkedin

v1

LinkedIn's feed is driven by comments in the first 60 minutes and dwell time. Always put links in the first comment, never in the post body. Native video and PDF carousels reach further than text-only.

LinkedIn's feed runs on first-hour comment volume + dwell time:
- Comments inside the first 60 minutes drive feed exposure. Posts that get 3+ first-hour comments routinely 5-10x their reach.
- Native video > carousel (PDF) > image post > text-only > article. Articles get ~30% less feed reach than equivalent post text.
Reviewed 5/31/2026The full LinkedIn breakdown

pinterest

v1

Pinterest is a visual search engine. Win with 2:3 vertical images, keyword-rich titles and descriptions, and fresh pins. Saves and outbound clicks drive reach, and pins keep surfacing for months.

Pinterest is a visual search engine, not a social feed:
- Distribution comes from keyword relevance in the pin title, description, and board name. Write for search, not for a timeline.
- Saves (repins) and outbound clicks are the signals that expand a pin's reach. The outbound destination link is the whole point of a p

threads

v1

Threads rewards reply-driving content within the first 45 minutes. Always attach an image — text-only posts reach about 40% less. Aim for 100-280 chars; longer posts lose reach.

Threads ranks on reply volume + reposts:
- Replies are the #1 lever. Posts with 3+ replies in the first 45 minutes routinely 3-5x their reach.
- Image-attached posts reach ~40% further than text-only. Always include an image when possible.
- 100-280 char range is the sweet spot. The full 500-char limit exists but the a
Reviewed 5/31/2026The full Threads breakdown

tiktok

v1

TikTok rewards watch-time and completion above everything. Hook in the first 1-2 seconds, shoot 9:16 vertical, ride a trending sound, and write a short caption. Links go in your bio, not the caption.

TikTok's For You page is won on watch-time, not follower count:
- Completion rate and average watch-time are the dominant signals. The first 1-2 seconds must hook or viewers swipe and the video dies.
- Shares (sends to friends) and comment velocity in the first hour drive the wider For You push. Rewatches count heavily
Reviewed 8/19/2026The full TikTok breakdown

twitter

v1

X (Twitter) rewards posts that get replies fast. Threads beat single posts. Skip URLs in the body. Keep hashtags to 0-1.

X rewards content that triggers replies in the first 30 minutes. The 2024 algorithm leak confirmed:
- Replies count 27x more than a like for ranking inside the home timeline.
- Posts with URLs in the body lose ~40% reach. Put URLs in the first reply.
- Threads (3+ tweets) get a measurable reach lift — break long-form i

youtube

v1

YouTube is won on thumbnail/title click-through-rate and average view duration. Hook retention in the first 30 seconds, keep titles to 40-70 front-loaded characters, and use Shorts as a discovery funnel into long-form.

YouTube ranks on two levers: click-through-rate and average view duration:
- CTR on the thumbnail + title decides whether a video gets shown; average view duration (retention) decides whether it keeps getting shown. Optimize both.
- The first 30 seconds set retention, so open with the payoff, not a long intro.
- Shorts
Reviewed 8/19/2026The full YouTube breakdown

Recently detected changes

No changes detected in the current monitoring window. Our canary compares the last 7 days of aggregate engagement against the prior 28 days. Quiet weeks are good news.

Also in research

The profiles above are researched and attributed. The other side of publishing is what each network mechanically refuses: the limits we enforce across nine networks, and what the documentation leaves out. Those are ours rather than researched, because we had to implement them.

FAQ

Frequently asked questions

How current are these algorithm profiles?

Each profile shows the date it was last revised, and where there is no revision the page says so instead of showing a fresh-looking date. That is the part most algorithm-update pages leave out: not what changed, but when anyone last checked. A profile with no review date is a baseline value the engine ships with, not a finding.

Where do these profiles come from?

5 of the 9 carry a named external source, listed on that network's own page: a platform statement, public documentation, or a published study. The other 4 (TikTok, YouTube, Pinterest and Bluesky) record no source at all, and we label those as engine defaults rather than dressing them up as research. The split lives in our data, not only in this sentence.

How would you know that an algorithm actually changed?

By watching results rather than announcements. Our canary compares the last 7 days of aggregate engagement against the prior 28 and surfaces the platforms that moved, which is a lagging signal on purpose: a platform statement tells you what they say they changed, and this tells you what actually shifted in real posting. Quiet weeks show as quiet.

Most tools tell you to post more. We tell you what each platform actually rewards.

Content Drifter writes posts in your voice, then schedules them for the windows your audience actually shows up. The free plan covers every network. No credit card to start.