xiry.

Marketing & Sales · Framework

X Algorithm Posting Strategy

Tactical cheat sheet for maximizing post reach under the open-sourced X algorithm

From the Xiry library · Updated 11 October 2026

This framework breaks down how the X recommendation system scores and filters content, based on xAI's open-sourced code. It details the specific positive signals, such as replies and direct message shares, alongside negative penalties like mutes and rapid-fire posting. It serves as a practical reference for tailoring post formats and timing to algorithm weights.

The framework

HOW SCORING WORKS

The algorithm uses a Grok-based transformer that predicts engagement probabilities. Score = Σ (weight × P(action)).

Positive signals (boost): likes, replies, reposts, quotes, shares (especially via DM & copy link), dwell time, profile clicks, follows, video quality views.

Negative signals (hurt): "Not interested" clicks, blocks, mutes, reports.

HIGH-IMPACT TACTICS

1. Optimize for Replies > Reposts > Likes — replies and quotes are stronger signals (deeper engagement). Ask questions, create controversy/debate, use hooks that invite responses.

2. Maximize Dwell Time — the algorithm tracks dwell_time; use longer-form content, threads, images that take time to consume, or videos. For video: must exceed minimum duration threshold for the VQV (Video Quality View) bonus.

3. Trigger Shares (especially DM shares) — sharing via DM and copy link are separately weighted signals; create "send this to someone" content (relatable memes, useful threads, shocking facts).

4. Avoid Negative Actions — block/mute/report signals have negative weights that actively push down your content; avoid engagement bait and divisive content that triggers blocks.

5. Post Timing Matters — an AgeFilter removes posts older than a threshold; fresh content gets considered, old content filtered out.

6. Build In-Network Reach First — the OONScorer applies a discount factor (OON_WEIGHT_FACTOR) to out-of-network content; posts from followed accounts have an advantage, so growing genuine followers helps.

7. Vary Your Posting Cadence — the AuthorDiversityScorer attenuates repeated posts from the same author; each additional post in a feed gets a decaying multiplier; quality over quantity.

8. Use Media Strategically — photo expand and video quality view are distinct signals; high-quality images users tap to expand generate signal; videos hitting the quality view threshold (watch time) get bonus weight.

TL;DR PRIORITY ORDER

1. Create reply/conversation-worthy content

2. Make it shareable (especially DM-worthy)

3. Maximize time spent viewing (dwell)

4. Avoid triggering blocks/mutes

5. Post fresh, space out your content

What it's for

  • Format posts to trigger high-weight algorithm actions like replies and DM shares
  • Pace posting frequency to prevent feed penalties from the author diversity filter
  • Structure media and thread length to maximize user dwell time

Key points

  • Scoring breakdown of positive signals (replies, reposts, DM shares, dwell time) versus negative signals (blocks, mutes, reports)
  • Engagement prioritization showing replies and quotes carry higher weights than standard likes
  • Dwell time optimization guidance for threads, expandable images, and video duration thresholds
  • Explanation of out-of-network discounts applied by the OONScorer
  • Author diversity rules explaining decay penalties for posting too frequently in a short window
  • Prioritized five-step checklist for day-to-day posting

Pricing

Free

Details

Type
Framework
Category
Marketing & Sales
Topic
Social Media Strategy
Open source
No
Platforms
Web
For
Content creators, social media managers, and founders growing an audience on X
Tags
x algorithm, social media, twitter, growth strategy, audience building, content strategy

Why it was saved

  • @weswinder, 20 January 2026: A complete, actionable reference for what boosts or hurts a post's score in the X algorithm.

Related

All 20 in Social Media Strategy →

Updated