1. engagement velocity is key…
the algorithm tracks real-time features such as likes, retweets, & replies which immediately boost your tweet’s ranking, so get as many interactions as you can in the first 30 minutes
2. comments > likes
the algorithm weighs replies more heavily than likes because they signal higher engagement
3. quality of engagement matters
engagement from accounts with high “tweepcred” which is twitter’s account reputation score gives your tweet a stronger boost
4. bookmarks signal value
the algorithm keeps track of bookmarks & sees them as a positive signal
5. click-through matters
the algorithm tracks when users click on things, the more that ppl expand your tweet, your profile, your images, click your links, etc. the better
6. watch time is tracked
for video content the algorithm measures how long users watch for & ranks based on that
7. negative feedback is measured
the algorithm penalizes content when people press “not interested”
8. dwelling time is important
the algorithm tracks how long users spend looking at your tweet
9. tweet text quality score exists
the algorithm has a “static quality of the text” score based on readability & length
10. media tweets get preferential treatment
tweets with images or videos are weighted more favorably
11. original content is preferred
the algorithm distinguishes between original tweets and retweets… it favors original content
12. conversation starters win
tweets that initiate conversations are ranked higher than general statements
13. thread quality matters
the quality of an entire conversation thread impacts individual tweet ranking
14. trends and current events get boosted
the algorithm gives preference to tweets containing trending terms
15. language matchin
tweets in the same language as the viewer are prioritized
16. toxicity reduces reach
the algorithm has “pToxicity” scores that nerf potentially offensive content
17. shouting (ALL CAPS) hurts visibility
the text quality scorer penalizes excessive capitalization
18. link preview quality matters
the algorithm evaluates the quality of link previews for tweets with urls
19. recency has major weight
the algorithm heavily favors recent tweets, especially in the “for you” timeline
20. posting in peak activity hours helps
tweets posted when your audience is most active get initial boost
21. algorithm uses “temporal decay”
tweets lose ranking power over time, with the decay rate varying by topic/domain
22. consistency is rewarded
regular posting schedules help establish patterns the algorithm recognizes
23. author credibility scores exist
the “tweepcred” score evaluates your overall account reputation
24. strong connections boost visibility
tweets from accounts with which the viewer has strong “real graph” connections rank higher in their timeline
25. simclusters matter
the algorithm places users in “communities” and content popular within your community gets boosted
26. topic authority is measured
the algorithm tracks topic-specific authority and boosts tweets when they’re in your areas of expertise
27. prior engagement history matters
the algorithm prioritizes content from accounts the viewer has engaged with previously
28. verification has weight
verified accounts receive slight ranking boosts
29. author diversity is enforced
the algorithm prevents any single author from dominating a timeline
30. conversation-worthy topics win
topics that typically generate discussions get preference
31. niche expertise performs better
tweets demonstrating specialized knowledge in specific domains rank higher
32. accessible formatting helps
well-formatted tweets with line breaks and clear structure score better on readability
33. clear call-to-actions work
tweets that explicitly invite engagement (without seeming spammy) perform better
34. audience-specific content resonates
content that aligns with the interests of your followers simclusters performs best
35. tweet semantics are analyzed
the twhin knowledge graph embeds tweets in a vector space to find related content
36. tweet entities matter
the algorithm gives weight to properly formatted hashtags, mentions, and cashtags
37. content similarity is tracked
the algorithm prevents showing too much similar content (“feedback fatigue”)
38. “heavy ranker” uses 6,000+ features
the final scoring model considers thousands of signals about each tweet
39. content balancing occurs
the algorithm maintains ratios between different content types
40. visibility filtering impacts reach
content flagged by safety systems receives reduced distribution
41. social proof boosts tweets
showing that people in the viewer’s network engaged with your tweet increases its ranking
42. topic tagging affects distribution
the “topic-social-proof” system identifies topics in tweets for better targeting
43. conversation starters outperform replies
starting fresh threads generally gets more distribution than replies
44. properly tagged media helps
using alt text and clear descriptions for media improves ranking
45. multi-modal content performs best
combining text, media, and interactive elements increases engagement signals
46. network effects matter
content that bridges multiple simclusters can achieve viral spread between communities
47. “graphjet traversals” determine reach
the algorithm finds potential audiences by traversing user-tweet interaction graphs
48. cross-domain engagement helps
receiving engagement from users across different interest clusters signals broad appeal
49. utility content gets boosted
content that solves problems or provides clear value ranks higher
50. authenticity signals matter