Key takeaways
- X (then Twitter) open-sourced portions of its recommendation algorithm, which showed reply engagement weighted many times more heavily than a like — an unusually direct public confirmation.
- Algorithms are proprietary and change often — treat public signals as directional, not guaranteed.
- Consistency and safe pacing matter more than gaming any single ranking factor.
X algorithm: publicly known ranking signals
- Reply engagement, weighted far more heavily than likes per the open-sourced code
- Engagement velocity shortly after posting
- Content that keeps users on-platform, versus posts that link away
- Account-level signals like follower relevance and past interaction history
What this means for a posting schedule
Posts that genuinely invite replies tend to outperform ones optimized only for likes. Consistent, well-timed posting for early velocity still matters, but content that sparks conversation is the stronger lever per X's own disclosed weighting.
What SkedCast can and cannot influence
SkedCast schedules and paces posts reliably; it does not influence how X's recommendation system scores an individual post's replies or engagement.
algorithmplatform
FAQ
- Are replies weighted more than likes on X?
- Yes — X's own open-sourced ranking code showed replies weighted many times more heavily than likes.
- Does posting time affect reach on X?
- Early engagement velocity is a known factor, so posting when your audience is active still helps, alongside content that invites replies.
- Do links out of X hurt reach?
- Public reporting and the open-sourced code suggest content that keeps users on-platform is generally favored over posts that primarily link away.