# WordPress Lost Market Share (MemberPress) > [!question] Lingering Questions > - The causal claim connecting Google's spam-policy purge to the specific market-share dip is explicitly labeled unprovable by the author himself — no CMS breakdown of deindexed sites exists. Has one been published since? > - The "90%+ of world webspam runs on WordPress" figure is a personal estimate, not a study finding — worth flagging every time this article gets cited. > - The piece's own commercial interest (MemberPress sells WordPress plugins, including a new MCP add-on built on the exact infrastructure the article promotes) is disclosed only once, briefly, near the end. > [!tip] Article Ideas > - "The metric that measures a market by counting who left it" — the market-share-denominator reframe as a general case study in vanity-metric design, cross-domain with the DevRel/community-metrics critique already in the vault. > [!link] Cross Reference > - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — the caller/callee framing for AI-agent-compatible platforms > - Cross-check the caller/callee framing against WP 7.1 coverage in the same reading batch for consistency before citing it elsewhere > [!quote] Vocabulary > - **caller vs. callee** — developer Jorijn Schrijvershof's distinction: a platform that *calls out* to an AI provider needs re-integration every time that provider relationship shifts; a platform that exposes itself as a standardized, permission-gated *callee* (WordPress's Abilities API + MCP Adapter) stays compatible as providers change > - **Abilities API** — a typed, permission-gated registry of site capabilities, matured in WordPress 6.9/7.0, that AI agents can operate against safely ## Metadata - Source: [WordPress Lost Market Share and Got Stronger. Here's Why.](3.%20Resources/Readwise/Articles/WordPress%20Lost%20Market%20Share%20and%20Got%20Stronger.%20Here's%20Why..md) - Author: MemberPress - Type: article (vendor blog) - URL: https://memberpress.com/blog/wordpress-lost-market-share-got-stronger-heres-why/ ## Author's thesis WordPress's headline market-share dip (43.6% peak in early 2025 to 41.5% by July 2026) is not platform decline but a denominator effect: Google's 2024–2026 spam-policy enforcement deindexed a large tail of PBN/content-farm sites that were disproportionately built on WordPress because it was free and cloneable at scale — while WordPress's share of the top of the web (59.2% of known-CMS sites, ~58% of the top 10,000 sites by traffic) stayed dominant and Wix/Shopify remained "almost absent" from that tier. A parallel "vibe coding" security crisis is pushing some founders back toward WordPress, and WordPress 7.0's AI Client, Abilities API, and MCP Adapter position it as both a caller and callee of AI agents for the AI era ahead. ## Key arguments (≤3) 1. **The market-share ratio is a composite metric whose denominator changed**: "percent of all websites" treats a Fortune 500 site and an abandoned affiliate blog as equal units; when Google's spam enforcement disproportionately removed WordPress-built junk-tail sites, WordPress's share of the total can fall without any healthy site leaving. 2. **WordPress dominates exactly where health would show up first**: two independent sources (W3Techs, HTTP Archive's 2025 Web Almanac) converge on WordPress holding ~58–59% of the high-value/high-traffic tier, while Wix and Shopify are nearly absent from it despite meaningfully higher overall-web share. 3. **Vibe-coding's documented security crisis is real and well-sourced independent of the WordPress argument**: Veracode's ~45% OWASP Top 10 vulnerability rate in AI-generated code, Escape.tech's scan finding 2,000+ high-impact vulnerabilities across 5,600 live vibe-coded apps, and a vendor estimate that 8,000+ of ~10,000 AI-built startups needed $50K–$500K "rescue engineering" — this section holds up on its own merits regardless of the WordPress framing built on top of it. ## Claim atoms ### Market share denominator shrinks as spam sites deindex > [!quote] ID:rw_h1 > So the real question is: what got removed from the web between 2024 and 2026? Reframes the entire market-share debate as a denominator problem, not a platform-health problem — the load-bearing move the rest of the article builds on. ### WordPress dominates the top-of-web despite overall dip > [!quote] ID:rw_h2 > Roughly 58% of CMS usage among the top 10,000 sites (HTTP Archive's 2025 Web Almanac) The strongest evidentiary counterweight to "WordPress is declining" framing in this batch, corroborated independently by W3Techs' 59.2%-of-known-CMS figure. ### Vibe coding produced a documented security and rescue-cost crisis > [!quote] ID:rw_h3 > roughly 45% of AI-generated code samples introduce OWASP Top 10 vulnerabilities Directly relevant to any AI-code-security item read in the same batch — same underlying theme (AI-assisted code producing exploitable, unreviewed vulnerabilities) from a market-adoption angle rather than a direct-exploit angle. ## Permanent candidates - **Callee interfaces outlast caller integrations under provider churn** — the article's single most reusable concept, generalizable well beyond WordPress: a platform that exposes a stable, permission-gated interface for agents to call *into* survives model-provider churn better than one that only calls *out* to a specific provider. Promoted below. ## Processing checklist - [x] Verify no existing Literature Note covers this doc (checked, none found) - [x] Extract claim atoms and permanent candidate from the existing deep-read - [ ] Check whether any independent CMS-fingerprinting study of deindexed domains has been published since, to settle the 90%-webspam-on-WordPress claim - [x] Connect into the relevant MOC — [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] ## Links - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — the caller/callee architecture question is a live instance of how platforms position themselves for agentic AI