IEO ENGINE · FIELD REPORT Nº 1 · MEASURED DATA, NOT THEORY

90 Days of Measured AI Citations: What Answer Engines Actually Retrieve

Server-side citation data from a live, instrumented three-site portfolio — including a working local-service business — measured from the only vantage point that can't be gamed: the server's own access logs.

PUBLISHED JULY 12, 2026 · UPDATED MONTHLY · MEASUREMENT WINDOW FEB 23 – JUL 12, 2026 · AUTHOR DREW M., IEO ENGINE™ (USPTO SERIAL 99676324) · PDF EDITION ↓
150+
AI citations measured
(lifetime, all platforms)
140
Days of continuous
measurement
190+
Distinct pages pulled
into AI answers
1,660
Pages crawled by AI,
never yet retrieved
LIVE TOTALS · updated in delayed batches by design — this report never exposes real-time, per-event data. Methodology note at the end explains why.

Every number on this page was measured, not estimated. It comes from reading raw server access logs on three live websites — a Sarasota pressure-washing company, a consumer photo app, and this methodology site — and identifying, verifying, and counting each time an AI answer engine pulled a page to build a response for a real user. Nothing here is a survey, a projection, or a vendor's dashboard. It is a ledger.


FINDING 01The Zero-Click Gap Is Real, and It Is Enormous

The most important comparison in this report is two instruments describing the same sites on the same day.

Readout · Same 24 hours · July 11, 2026 · all three sites✓ Measured
Google Search Console: 321 impressions · 0 clicks · 0.0% CTR
Server-side citation ledger: ~821 answer-surface retrievals
One instrument counts clicks in a world that has largely stopped clicking. The other counts where the audience actually went: into AI-generated answers built from these sites' pages. One of these numbers is the reason a business believes its website "does nothing."
In a single 24-hour period, AI answer engines retrieved these sites' pages roughly 2.5 times more often than Google showed them to human searchers — and Google delivered zero clicks.

FINDING 02Each Platform Has a Different Clock

How long does it take between an AI crawler ingesting a page and an answer engine actually using it? Measured across every crawl-to-citation round trip in the archive, the platforms behave like different species:

Readout · Ingestion → first citation, median lag✓ Measured
PlatformRound trips observedMedian lag
Perplexity311 minutes
OpenAI / ChatGPT109.2 days
Google (AI answers)1620.3 days
All platforms306.9 days
Perplexity's near-instant loop (small sample, n=3, and growing) makes it the fastest feedback instrument in the industry: content published in the morning can appear in its answers the same day. Google's answer layer is the slowest to adopt — but holds what it adopts.
Median time from AI ingestion to first citation: 11 minutes on Perplexity, 9.2 days on OpenAI, 20.3 days on Google.

FINDING 03Rank and Citation Are Nearly Unrelated

The single most consequential finding for anyone still optimizing purely for rank. Comparing Google Search Console position data with measured citations, page by page, across the portfolio:

Readout · Rank vs. citation, correlation & exhibits✓ Measured
Correlation between Google rank and AI citation: r = 0.29 — weak
Query (local service site)Google positionTimes cited by AI
"roof cleaning lido key"3.00 — never
"roof cleaning bayshore gardens"5.30 — never
"hoa paver sealing siesta key"436
Pages ranking in Google's top 5 that have never once been pulled into an AI answer — while a page ranked #43 was cited six times. Answer engines select for clear, extractable answers backed by consistent facts — not for rank. Being #1 and being the answer are different achievements, earned by different work.

FINDING 04What Answer Engines Actually Want From a Local Business

Classifying every measured retrieval on the local-service site by page type produces an unambiguous winner:

Readout · Retrievals by page class · local-service site✓ Measured
Page classAI retrievals
Cost & pricing guides205
Homepage202
Long-form articles96
Service & location pages70
"How much does X cost" is the query class answer engines are built to serve, and honest, specific pricing content is what they reach for. The traditional service-page-per-town architecture — the backbone of local SEO — is the least retrieved class on the site.

FINDING 05Being Crawled Is Not Being Used

The most sobering number in the ledger. AI crawlers are voracious — but ingestion is a warehouse, not a shelf:

Readout · The ingestion-retrieval gap✓ Measured
1,660 pages crawled by AI platforms across the portfolio have never once been retrieved for an answer
Documented extreme: on July 11, one platform's crawler ingested an entire 340+ page content library in 23 minutes, every request verified against the operator's published IP ranges. Answer-surface retrievals from that library since: zero, with the clock still running. Getting eaten is free. Getting used is earned.
Across three instrumented sites, 1,660 AI-crawled pages have produced zero answer-engine citations — the gap between being ingested and being useful is where the actual optimization work lives.

FINDING 06The Training Channel Is Separate From the Citation Channel

On July 11, Common Crawl — the nonprofit dataset that major LLM training corpora are built from — performed a verified crawl of the local-service site: 181 pages in 104 minutes, one polite request every 35 seconds, robots.txt honored. Cross-checking against Common Crawl's public index confirmed 23 pages already in the May 2026 corpus, including two PDFs.

Two lessons worth publishing. First, this is a different door than citation: content taken by Common Crawl doesn't get quoted — it gets learned, becoming part of what future models simply know. Second, Common Crawl discovers by following links, and it missed four of the site's six most-retrieved pricing pages — the same under-linked pages the answer engines have to work hardest to find. Internal linking, dismissed in some AI-optimization circles, still decides what enters the training corpus.

FINDING 07You Cannot Trust a User-Agent String

During this measurement window, the instrument caught live impersonation: a Google Cloud VM announcing itself as an Anthropic answer agent, rented cloud machines claiming to be Google's and Anthropic's crawlers, and a probe session cycling through user-agent variants to test the site's defenses.

Readout · Identity verification policy✓ Enforced
Every count on this page is verified against operator-published IP ranges (OpenAI, Google, Perplexity, Anthropic — 1,100+ prefixes, refreshed weekly) or forward-confirmed reverse DNS
A user-agent is a claim; the network is the evidence. Impersonators are logged and counted nowhere. Agents that cannot be verified are flagged, never assumed genuine. These are floor numbers, not estimates.

Method — named, not disclosed

Measurements are produced by a private server-side instrument reading raw access logs across the portfolio: citation events are detected by proprietary log forensics, de-duplicated so multi-page fetches within one answer count once, filtered of all operator traffic and tooling, tagged and excluded during any operator self-testing, and identity-verified as described in Finding 07.

Detection mechanics, scoring weights, and instrumentation fingerprints are deliberately not documented here: publishing how the measurement works would let subjects game the measurement. The live totals above update in delayed batches for the same reason — a real-time counter is an oracle that can be probed. Findings are published; the instrument is not.

This page is updated monthly with new measured findings. The measurement continues. A citable, self-contained PDF edition of this report is maintained at a stable URL alongside this page.