繁中 →
— views

The Threads playbook

Algorithm, numbers, viral formulas, posting times
All of it backed by data

← HomeAbout·Failure list·GitHub·Threads
?

The numbers at a glance

Data compiled 2026
21,864
posts in the sample
52 million
posts in Buffer's study
5
main sections
1

Algorithm weights

Meta patent analysis + Buffer research
1
Speed of engagement
Source: Meta patent analysis | the first 30 minutes matter most
2
Time spent reading
Source: Buffer research | a post read through is worth more than one scrolled past
3
Reply depth
Source: Mosseri | reply threads matter more than likes
4
Topic match
Medium weight | it puts you in front of the right audience
5
Posting frequency
Base weight | regular beats frequent
!

April 2026 algorithm update

Recent changes in how the algorithm behaves
The numbers on this page are an Apr 2026 snapshot. Weights change, the method still holds, so treat your own account's data as the source of truth.
1
Replies > likes
The 30-90 minutes after you post is the golden window, and the reply count decides how far it travels
2
Longer comments carry more weight
A long reply is worth more than a one-word comment like nice or +1
3
Topic consistency (content neighborhoods)
AI sorts your account into a content neighborhood, and jumping topics too often cuts your reach
4
Engagement bait gets penalized
A CTA like comment XX and I will send you the guide can get downranked
5
Dear Algo is live
People can now tell the algorithm in plain words what they want to see, which changes what gets pushed
2

The numbers that matter

From real tests and published research
10 > 100
10 real reply threads beat 100 likes
Mosseri, official statement
10.09%
Average monthly follower growth after 500+ replies a month
21,864 posts tested
+60%
Extra engagement an image post gets over a text-only post
Buffer, 52 million posts
2.34%
Engagement rate on link posts, the lowest of any format
Buffer research
+42%
Engagement lift after replying to your own comment
Buffer research
8.87%
Average monthly follower growth at 5–6 posts a week
21,864 posts tested
3

Best times to post (Taiwan)

Slot Rating Notes
Thursday 20:00 Best First choice
Thursday 21:00 Runner-up Same day, still prime time
Wednesday 20:00 Third Midweek peak
Saturday, all day New 2026 data, worth testing
8–10 in the morning Worst Commute hours, low engagement
Sunday, all day Avoid Lowest engagement overall
Best days, in order: Thursday > Tuesday > Wednesday
4

Four formulas for posts that travel

Reach

Get seen by more people

Pain everyone shares Practical how-to News
Follows

Make people want to follow

Personal experience Honest feelings A point of view
Engagement

Make people comment

Show your value Open question Shared pain point
⚠️ Apr 2026 update: comment XX and I will DM you can get downranked, an open question is safer
Connection

Make people feel it

Everyday story A line people quote Emotion
5

Hook templates

The first line decides whether anyone reads on
Flip the expectation
Show the behind the scenes
Open with something relatable
Admit something about yourself
Ask a question
Lead with a number
The first line decides whether anyone reads on.
Flips and numbers work the most reliably.
6

How to write it

  • 1
    Keep every line under 15 words, break the lines, easy to read on a phone
  • 2
    Short posts of 3–6 lines take off most often
  • 3
    Bulleted posts of 7–12 lines get shared most (the meaty ones)
  • 4
    Put links in the first reply, not in the post itself
  • 5
    Write like you talk, not like an essay, casual > formal
7

What to do after you post

  • 1
    Reply to every comment in the first 30-90 minutes. The reply count in that golden window decides how far the post travels (Apr 2026 update)
  • 2
    Reply to your own comment and add a point. Worth +42% more engagement
  • 3
    Do not post again while one is taking off. It splits the traffic
  • 4
    No same-angle content within 48 hours. Let the system redistribute
  • 5
    After a big post, share a personal reflection. That turns traffic into follows
8

Beginner mistakes

  • XTreating Threads like a classroom, only teaching, never talking
  • XPosting without engaging, never replying on other people's posts
  • XRepackaging other people's knowledge with no view or feeling of your own
  • XCopy-pasting across platforms, dropping IG captions straight in
  • XOne big push, then disappearing, no steady presence
  • XJumping topics (AI today, food tomorrow), the algorithm does not know who you are so it does not push you (added Apr 2026)
  • XPolishing forever, waiting until it is perfect (posting it is what makes it good)
9

How Threads differs from IG

Item Threads IG
Main format Text posts, image posts Reels video
Video weight Low, video is not its strength Very high, Reels is the main engine
How it gets pushed Text engagement triggers it (comments, reply threads) Reels completion rate and shares trigger it
Core behavior Post + reply to others + go deep in threads Make video + chase completion + land a big Reel
Production effort Low, typing is enough High, you need to shoot and edit
On Threads, one good text post > one beautifully produced video.
Put the time into words, not into filming.
?

Sources

References
  • Meta patent analysis: engagement speed and how the algorithm distributes posts
  • Buffer research: analysis of 52 million Threads posts
  • Mosseri's official statement: reply depth versus the weight of likes
  • 21,864 posts tested: how posting frequency, follower growth, and reply volume correlate
  • RecurPost / MomentumHive / EmbedSocial / Metricool: Q1 2026 algorithm reports
  • Playbook compiled by Leo Huang @kanisleo328, 2026 edition
  • Last updated: 2026/04/05

Want more Threads tactics?

Follow Leo for first-hand notes as they happen

Follow @kanisleo328

Useful? Cite this page and pass it on

Published 2026-04-02, last updated 2026-08-19

The Threads playbook by Leo Huang @kanisleo328

leoaido.com

One-person company + an AI team — Built with Claude Code

#leoaido #kanisleo328 Leo Huang