Algorithm Intelligence Brief

X Algorithm
Playbook

Decoded from Twitter's open-source algorithm. 145,000 community clusters. 6,000 features per tweet. Every signal mapped.

Engagement Signal Hierarchy

Not all engagement is equal. The algorithm uses weighted predictions for each signal type. Here's the actual hierarchy from the codebase:

Rank Signal Algorithmic Impact
#1 Replies + Author Reply-Back PREDICTED_IS_REPLIED_REPLY_ENGAGED_BY_AUTHOR Highest weight. When someone replies AND you engage back, massive boost.
#2 Bookmarks PREDICTED_IS_BOOKMARKED Signals "save-worthy" content. Tracked separately from likes.
#3 Shares PREDICTED_IS_SHARED External distribution intent. High value signal.
#4 Quality Clicks (15s+ dwell) PREDICTED_IS_GOOD_CLICKED_V2 Click + meaningful time spent = quality engagement proof.
#5 Profile Click → Engagement PREDICTED_IS_PROFILE_CLICKED_AND_PROFILE_ENGAGED Shows author interest. Builds Real Graph relationships.
#6 Replies PREDICTED_IS_REPLIED Conversation value. Weighted significantly higher than likes.
#7 Retweets / Quote Tweets PREDICTED_IS_RETWEETED Distribution signal. Quote tweets add context value.
#8 Video 50%+ Watch PREDICTED_IS_VIDEO_QUALITY_VIEWED Video-specific quality metric. Completion rate matters.
#9 Likes PREDICTED_IS_FAVORITED Basic approval. Lowest positive signal in the hierarchy.

Key Insight: The most underutilized signal is replying to replies on your own tweets. Most accounts optimize for likes (lowest signal) while ignoring the highest-value behavior.

✓ DO's — Increase Reach

01

Reply to Your Repliers

Within the first hour, engage back with replies on your tweets. This triggers REPLY_ENGAGED_BY_AUTHOR — the highest-weighted positive signal.

02

Create Bookmark-Worthy Content

Tutorials, threads, resources, insights. Bookmarks are tracked separately from likes and signal "I want to return to this."

03

Optimize for Dwell Time

The algorithm tracks PREDICTED_IS_DWELLED. Longer, substantive content that holds attention (15+ seconds) scores better.

04

Build Real Graph Relationships

Consistently engage with target accounts. The algorithm builds relationship strength via mutual follows, profile visits, and direct replies.

05

Earn High-Authority Engagement

TweepCred uses PageRank — engagement from high-authority accounts carries more algorithmic weight. Target quality over quantity.

06

Write Reply-Inviting CTAs

"What's your take?" beats "Like if you agree." Replies are weighted much higher than likes in the ranking model.

07

Be Consistent in Topic Focus

SimClusters builds embeddings based on who engages with you. Topic consistency = stronger algorithmic identity matching.

08

Maintain Healthy Follow Ratio

More followers than following. Reputation.scala explicitly penalizes accounts with inverse ratios.

✗ DON'Ts — Decrease Reach

01

Ignore Your Replies

You're leaving the highest-value signal on the table. Every ignored reply is a missed opportunity for algorithmic boost.

02

Optimize Only for Likes

Likes are the lowest positive signal. Content that earns only likes underperforms content that earns replies, bookmarks, or shares.

03

Mass-Follow or Follow-Unfollow

Unfollows are tracked as NEGATIVE_SIGNAL. Pattern detection flags coordinated behavior. Damages Real Graph scores.

04

Post Repetitive Content

"Feedback Fatigue" system down-ranks repetitive authors by up to 80%. Penalty persists for 140 days.

05

Use Engagement Pods

Graph analysis detects coordinated engagement patterns. Artificial engagement undermines authentic Real Graph building.

06

Trigger Trust & Safety Filters

Models scan for toxicity (pToxicity), harassment (pAbuse), NSFW content. Triggers visibility filtering and downranking.

07

Hashtag Stuffing

The algorithm uses semantic understanding via SimClusters, not hashtag counts. Stuffing looks spammy without adding value.

08

Post Without a Hook

First line determines if content gets ranked. Weak hooks = low initial engagement = algorithmic death spiral.

Penalty Severity Matrix

Signal Penalty Weight Duration Recovery
Reports -20,000 (Extreme) Persistent Difficult — account-level impact
Blocks / Mutes -1,000 (Severe) Persistent Gradual — through positive signals
"Not Interested" Clicks -95% Similar Content 140 days Time-based decay
Unfollows After Engagement Real Graph Damage Decaying Rebuild through consistent engagement
Tweet Unliking Negative Signal Tracked Minor — offset by new engagement
The Core Formula
Reach = Σ(Positive Signals × Weights) − Σ(Negative Signals × Penalties)

Team-Specific Actions

Marketing

  • Write threads optimized for replies, not likes
  • End posts with reply-inviting CTAs
  • Always engage back with repliers in first hour
  • Create bookmark-worthy resource posts
  • Front-load value in first line (hook)

Community

  • Build Real Graph through consistent engagement
  • Reply substantively to community members
  • Quote-tweet community wins with added context
  • Monitor for negative signals (unfollows, mutes)
  • Seed engagement from target community first

BD / Partnerships

  • Tag partners in announcements (mutual boost)
  • Create content that earns partner audience replies
  • Focus on quality engagement from high-authority accounts
  • Use quote-tweets for partnership amplification
  • Build Real Graph with target investor/builder accounts

Quick Reference Checklist

Before Every Post

  • Does this content invite replies, not just likes?
  • Is this bookmark-worthy (saves value)?
  • Does the first line hook AND contain topic keywords?
  • Am I prepared to engage with replies in the first hour?

Daily Habits

  • Post at least once during audience peak hours
  • Engage with replies on recent posts
  • Interact with 5-10 accounts in target community
  • Check for and respond to DMs/mentions

Weekly Audit

  • Review follower/following ratio
  • Check for any visibility restrictions
  • Analyze which content earned replies vs. just likes
  • Identify high-authority accounts engaging with you

Monthly Review

  • Topic consistency — building clear SimClusters embeddings?
  • Graph growth — are the right accounts engaging?
  • Negative signal check — unfollows or mutes trending up?
  • Content format testing — what earns highest-value engagement?