a builder's codex
codex · operators · Eli Schwartz · ins_eli-schwartz-personalized-search-dissolves-position

Keyword position reporting is measuring a construct that is dissolving as Google personalizes results per user

By Eli Schwartz · Product-led SEO consultant and author · 2026-05-07 · essay · The Personalized Internet Is Here

Tier A · TL;DR
Keyword position reporting is measuring a construct that is dissolving as Google personalizes results per user

Claim

Keyword position reporting is measuring a construct that is dissolving. Google has crossed into genuine context-aware personalization: the same query returns different results for two people in the same neighborhood at the same moment. There is no longer a single stable position to rank for.

Mechanism

Google uses two decades of behavioral data to serve personalized results. Personalization currently lives inside Gemini and will expand to AI Mode, then AI Overviews, then all search surfaces. A keyword-position metric reports a statistical aggregate of an increasingly fragmented distribution. As personalization deepens, that aggregate becomes less representative of any individual user's experience. The construct being measured, position for a given query, is losing its referent.

Conditions

Holds when: Google continues expanding context-aware personalization across more surfaces. The trajectory is confirmed; the pace is not publicly disclosed.

Fails when: personalization rollout stalls, or Google introduces position-reporting tools that account for personalization variance.

Evidence

Schwartz argues Google has already crossed the threshold:

There is no longer a single stable answer to rank for.

Personalization is active in Gemini-powered surfaces and tracking for expansion to AI Mode and AI Overviews. A dashboard showing "position 4 for [query]" is reporting one user's experience, not a market-level fact.

Signals

Counter-evidence

Google has not formally confirmed position personalization as a broad policy for non-logged-in users. Some practitioners argue personalization applies primarily within logged-in sessions, leaving anonymous searches more stable. Personalization depth varies by query type.

Cross-references

Open the interactive view → View original source → Markdown source →