X (Twitter) algorithm signals cheatsheet
Quick-reference card for how the X algorithm scores content in 2026. Engagement weights, penalties, recency signals, recommendation triggers.
Printable PDF
Get the X (Twitter) algorithm signals cheatsheet as a printable PDF
Drop your email — we'll send the PDF version (printable + shareable) plus the weekly X growth playbook each Sunday. Unsubscribe anytime.
We'll never share your email. Unsubscribe with one click.
The X algorithm has stable patterns that compound. This cheatsheet is the reference card — the values you want at hand when planning a viral attempt or auditing why a post underperformed.
Engagement weights (per X open-sourced ranking)
| Reply (substantive) | 27x a like |
| Quote tweet | 18x a like |
| Retweet | 11x a like |
| Like | 1x baseline |
| Bookmark | ~0.5x a like (less weight, but signals save value) |
| Profile click after viewing tweet | 12x a like |
| Time spent on tweet | Significant — long dwell time boosts |
Penalties (down-weight signals)
| External URL in tweet body | -10 to -40% reach |
| Reported as 'not interested' | -74x a like (large penalty) |
| Reported as spam / abuse | Account-level suppression possible |
| Muted by user | -100% reach to that user |
| Block by user | -100% reach + removes from search results |
| Very low engagement first 30 min | Reach decays sharply |
Recommendation triggers (what gets you into For You)
| Engagement velocity (first 30 min) | Most important single signal |
| Engagement rate (eng / impressions) | Boost if above 3-5% |
| Reply quality (length, sentiment) | Substantive replies multiply |
| Cross-network connection | Engagement from accounts you don't yet follow |
| Topic affinity | Posts in topics user has engaged with before |
| Author authority | Higher-authority accounts amplify faster |
| Recency | Decay starts ~2-4 hours, sharp drop at 24h |
Format effects
| Image attached | +30-60% impressions (mixed by niche) |
| Video attached | +50-150% impressions, +200% time-on-tweet |
| Long-form (250+ chars) | Separate format, different distribution rules |
| Thread (4+ tweets) | Top tweet treated like a singleton; subsequent tweets get progressively less reach |
| Poll attached | +10-30% engagement, no impression boost |
| Reply to your own tweet | Extends thread freshness signal |
Account-level signals
| X Premium (Basic) | Reply ranking boost (10-30% lift) |
| Verified | Trust signal, slightly amplified reach |
| Follower count under 500 | Limited recommendation surface |
| Account age < 30 days | Reduced reach (anti-spam measure) |
| Engagement rate > 5% | Account-level boost on subsequent posts |
| Posting cadence consistency | Posting daily increases overall account weight |
Common questions
Are these signal weights official or estimated?+
Engagement weights (reply 27x, RT 11x, etc.) are from X's open-sourced ranking model from 2023; X has updated the model since but the relative ordering remains. Other signals are inferred from operator-observed patterns across thousands of accounts. Treat as practical guidance, not academic precision.
Does paying for X Premium really help my reach?+
Modestly. The biggest concrete benefit is reply ranking boost (10-30% lift in reply visibility). Original tweet reach increase is smaller. For accounts that reply heavily, Premium ROI is positive; for occasional posters, it's marginal.
How much do penalties matter?+
External URL penalty (-10 to -40% reach) is the most operator-relevant. The 'not interested' penalty is large per click but rare in aggregate. Most accounts don't get reported enough for spam/block penalties to matter; if you are, the algorithm is the symptom, not the cause.