Answers from the X algorithm, not Twitter folklore.
Direct answers to late-night reach questions, account labels, and cleanup rules — grounded in the open-source recommendation code.
Core Pillar Guides
What are X account labels, and which ones actually kill reach?
Account labels drop all posts from non-follower recommendations. Post labels usually hit one tweet. The ones that kill discovery are SpamHighRecall, DoNotAmplify, and Compromised.
Read Answer → Labels & VisibilityWhat's the difference between SpamHighRecall and SPAM_HIGH_RECALL?
One is the account. One is a single post. Only the account label suppresses discovery across your entire profile.
Read Answer → Deleting & CleanupWill deleting old spam posts make my account alive again?
Sometimes. Deleting live bait, affiliate, and airdrop posts removes fuel for labels, but it will not flip reach overnight. Targeted cleanup + weeks of clean posting is the real path.
Read Answer →All 17 Sourced Questions
What does DoNotAmplify mean on X?
X thinks the account may violate Terms or safety policies. Posts stay up for followers, but are dropped for strangers in For You.
Full answer → Labels & VisibilityHow do I check my X account labels in Under the Hood?
Official portal: x.com/i/under_the_hood. XPurge maps the same label concepts onto your local archive file.
Full answer → Deleting & CleanupShould I mass-delete my entire X archive?
No. Mass-nuking deletes positive historical graph signals and can trigger bot heuristics. Surgical deletion of flagged spam is far safer.
Full answer → Deleting & CleanupWhich old posts are actually worth deleting?
Focus on 4 categories: engagement bait, dead/broken affiliate links, crypto airdrop templates, and burst reply dumps.
Full answer → Deleting & CleanupCan bulk deletion itself look like spam?
Yes, if you delete thousands in minutes via unthrottled scripts. Paced, rate-limit safe deletion avoids botmaker cadence alarms.
Full answer → Scanner SignalsWhat is engagement bait, and why does X drop it?
Posts that solicit superficial replies ("drop your @", "comment YES") trigger Grox classifiers and get dropped from non-follower discovery.
Full answer → Scanner SignalsDo links in posts kill reach?
Not as a direct penalty, but links lower dwell time and copy-link shares, reducing final weighted score by 40–60%.
Full answer → Scanner SignalsWhat is burst cadence, and why does it get flagged?
Publishing 10 posts in 5 minutes triggers Author Diversity decay and botmaker rate alarms, cutting per-post impressions.
Full answer → Scanner SignalsWhy do airdrop and affiliate posts get marked dropped?
Crypto airdrop templates, contract address strings, and affiliate links trigger Grox spam models with high confidence.
Full answer → Scanner SignalsIs “92% organic” enough if I still have hundreds of bait posts?
No. Having 400 live bait posts provides a massive surface area for mutes and reports that can easily push you over the label threshold.
Full answer → Ranking vs FilteringWhy can followers see my post but For You strangers can’t?
Because Visibility Filtering rules split by in-network vs out-of-network. Labels like SpamHighRecall only drop strangers.
Full answer → Account RecoveryHow long after cleanup does reach come back?
Typically 14 to 30 days of consistent, clean posting after removing flagged spam tweets.
Full answer → Account RecoveryWhat should I post after I purge spam?
High-dwell insights, diagrams, deep text breakdowns, and mutual conversation. Avoid external links and promo bait.
Full answer → Account RecoveryShould I put affiliate links in the reply instead?
Yes, if you must share a link. Keeping the root post clean maximizes For You discovery while letting interested users find the link in the thread.
Full answer →