ENTITY

Lily Ray

Properties1
tagsseo-researcher, ai-content, algorithm-updates, scaled-content-abuse, substack

3 min read

Lily Ray

Type: person | Domain relevance: Independent SEO practitioner and researcher; author of the 220-site AI-content study that is this wiki’s main practitioner evidence on Scaled Content Abuse.

Summary

Lily Ray publishes on Substack (lilyraynyc.substack.com) and describes “more than a decade of helping companies recover from Google algorithm updates.” In It Works Until It Doesn't: AI Content Strategies That Backfire (2026-05-13) she reports monitoring 220+ sites named as customers by more than a dozen AI content platforms, using Ahrefs estimates corroborated against the Sistrix Visibility Index. 54% lost 30% or more of peak organic traffic, 39% lost half or more, 22% lost three quarters or more, typically after 6–12 months of page growth and a traffic peak 3–6 months later. She labels the shape “Mount AI,” crediting Glenn Gabe.

She is unusually careful about limits, which is why she is useful here: third-party estimates only, a long list of confounders, no claim that any tool caused any outcome, and no vendors or domains named. Her explanation (formulaic, competitor-repeatable templates become a detectable footprint once enough sites publish them) is offered as professional opinion. She says she spent 2024 working with site owners hit by Google’s September 2023 helpful content update and March 2024 core update, and that AI-tool output now “looks and feels” like what those updates removed.

She pitches no product or service in the piece, which distinguishes it from most of this wiki’s SEO sources. Caleb Ulku relays the study in They Lied About AI Content (This Proves It) and adds a local-SEO layer that is his, not hers.

Key Facts

  • 220+ sites / subfolders; 12+ vendors’ public customer stories; Ahrefs + Sistrix; data as of May 2026.
  • 54% / 39% / 22% lost ≥30% / ≥50% / ≥75% of peak organic traffic. No causation claimed.
  • Identified eight recurring templates on declining sites; says all of them work until scaled.
  • Observed an unconfirmed late-January 2026 update: 40+ sites, 40–95% declines, centered on self-promotional listicles.
  • Recommends E-E-A-T (Experience, Expertise, Authoritativeness, Trust), Information Gain, AI-use transparency, and an expert human in the loop.
  • Source of the “could a competitor publish a near-identical page tomorrow” test.
  • Earlier work referenced but not in this wiki: a February 2026 Substack piece on self-promotional listicles; a March 2024 Amsive article against scaled FAQ pages.

In the Sources

SourceContext
It Works Until It Doesn't: AI Content Strategies That BackfirePrimary: the study, the eight templates, the January 2026 pattern, safe-use guidance.
They Lied About AI Content (This Proves It)Caleb’s accurate relay of her numbers, plus his own local-SEO extension.

Entities: Glenn Gabe, Caleb Ulku, Google, Ahrefs Concepts: Scaled Content Abuse, Information Gain, Programmatic SEO, GEO / AEO (Getting Recommended by AI), E-E-A-T (Experience, Expertise, Authoritativeness, Trust)