SEO Echo Report - 2026-10-25

20 new studies · 1 disputed

In this report
Grade	Study	Skeptic	Red flags	Original
Tier B	Ahrefs survey of 301 marketers: 69% publish a larger share of AI content than a year ago	—	selection bias	no primary source
Tier B	Insurance AI-search benchmark: Google AI Mode drives 43.9% of citations; AI Overviews on 40.7% of searches	—	correlation causation, vendor metric	no primary source
Tier B	Nonprofit AI-search benchmark: Google AI Mode drives 50% of citations; AI Overviews on 44.2% of searches	—	correlation causation, vendor metric	no primary source
Tier B	Search Signals Index, Aug 10-31 2026: index -2.4% across 1,842 companies; Amazon -16.9% to a record low	—	correlation causation, confounded window, vendor metric	no primary source
Tier D	Orbit Media survey: marketers reporting strong content results fell to 14% in 2026 (from 20-28%)	—	correlation causation	no primary source
Tier B	ChatGPT query fan-out across 66.6M responses: only 14.4% search the web, ~1.6 queries each	—	vendor metric	no primary source
Tier D	Ahrefs signup survey: 21.7% of signups credit ChatGPT or Claude, more than YouTube	—	correlation causation, anecdote generalisation	no primary source
Tier B	AI Overviews on branded Google searches: 61.3% in July to 73.3% in September 2026 (82.91% on Sept 29)	—	cherry pick, headline overclaim	no primary source
Tier B	AI Overviews on 400 commercial keywords: 45.75% to 88% in a year; 'review' queries now ~90%	—	vendor metric, selection bias	no primary source
Tier D	Zapier case study: cited in ~80% of its bottom-funnel AI prompts; 25% of signups attributed to AI	—	correlation causation, anecdote generalisation, selection bias	no primary source
Tier B	1,000 transactional prompts: 52.5% of the most-cited domains in LLM answers are affiliate or sponsorship sites	—	correlation causation, cherry pick, headline overclaim, vendor metric	no primary source
Tier B	Hidden AI traffic in Direct: only 0.003% of Direct sessions across 12 sites; no reliable AI-visibility link in 632 tests	—	none	no primary source
Tier B	LLMs.txt across ~300,000 domains: 10.13% adoption, no measurable effect on AI citations	—	headline overclaim	no primary source
Tier C	Local SEO overhaul for a French paving firm: ~100 to 1,100+ monthly visits and 2-5 to ~12 quote requests in 3 months	—	headline overclaim, no baseline, selection bias	no primary source
Tier C	Content clusters for a French cruise site: 2.7x organic traffic and 20x qualified leads in 9 months	—	correlation causation, no baseline, selection bias	no primary source
Tier B	AI Overviews in France at launch: 52.63% of 100,000 keywords, about twice the first-month rate in Germany and the Netherlands	—	none	no primary source
Tier B	GPT-5.6 Luna vs GPT-5.5 on 346 matched prompts: site: searches 22% to 76%, inline citations -12.9%	—	headline overclaim	no primary source
Tier B	AI answers about one brand: platforms declined 31.2% of prompts but were inaccurate only 2.8%	—	anecdote generalisation, selection bias	no primary source
Tier B	Organic rank vs AI Overview citations (168M citations): position 1 cited 24.9%, 57% of citations from outside the top 10	—	correlation causation	no primary source
Tier D	Same ChatGPT prompt rerun: 14 to 40 retrieved URLs in two weeks, zero shared queries 8 minutes apart	—	anecdote generalisation	no primary source
Tier D	Updating stale pages and requesting indexing flipped a wrong AI Overview answer within about an hour (one brand)	D	anecdote generalisation, confounded window	no primary source

How to read this table

Tier A Tier B Tier C Tier D strongest to weakest evidenceTier ? ungradedSkeptic: the blind second grade, — when it agrees≠disputed#numbers not found in the original∅original not reachable

New studies

Tier Bvendor

Ahrefs survey of 301 marketers: 69% publish a larger share of AI content than a year ago

Ahrefs' second annual survey (2026) found that 93% of 301 respondents use AI for content, up from 87%. 69% publish a larger share of AI content than a year ago, and 85% of AI users produce content faster. The answers are self-reported and probably come from Ahrefs' own audience. Last year's sample was larger and used different questions, so the year-over-year changes are only indicative.

Red flags: selection bias

Conflict of interest: Ahrefs sells AI content tooling (Letaido) and runs the survey.

Measured: share of marketers publishing more AI content

Research: 2026 (survey)

Part of: Ahrefs AI content marketing survey (editions 2025, 2026)

Subtopic: Programmatic & AI content

Question: What share of marketers publish more AI-assisted content than a year ago?

About: Claude, ChatGPT

Published: 2026-09-30

Coverage: Ahrefs Blog

Tier Bvendor

Insurance AI-search benchmark: Google AI Mode drives 43.9% of citations; AI Overviews on 40.7% of searches

Conductor analysed over 3.6 million US insurance-domain citations across seven AI engines (January-May 2026) and 2.9 million Google searches (May 20-June 20). AI Mode produced 43.9% of the citations and Perplexity 33.6%. AI Overviews appeared on 40.7% of searches, with 53.3% of their citations from outside the organic top 10. The vendor uses its own taxonomy, and some subindustry figures appear inconsistent.

Red flags: correlation causation, vendor metric

Conflict of interest: Conductor sells AI visibility tracking software.

Measured: AI citations

Research: January 1 - May 31, 2026 (engines); May 20 - June 20, 2026 (AIO)

Part of: Conductor AI Search & AI Overview Benchmarks: Insurance (editions 2026)

Subtopic: AI Overviews & AI search

Question: Which AI engines account for the most citations of an industry's domains? · 1 study · 2 articles

About: Google AI Mode, Google AI Overviews, Perplexity, Progressive

Published: 2026-08-12

Coverage: Conductor Academy

Tier Bvendor

Nonprofit AI-search benchmark: Google AI Mode drives 50% of citations; AI Overviews on 44.2% of searches

Conductor analysed over 14 million US nonprofit-domain citations across seven AI engines (January-May 2026) and 10.7 million Google searches (May 20-June 20). AI Mode produced 50% of the citations, ahead of Perplexity (27.9%) and ChatGPT (12.2%). AI Overviews appeared on 44.2% of searches, and 48.1% of their citations came from outside the organic top 10. Shares and intents use the vendor's own taxonomy.

Red flags: correlation causation, vendor metric

Conflict of interest: Conductor sells AI visibility tracking software.

Measured: AI citations

Research: January 1 - May 31, 2026 (engines); May 20 - June 20, 2026 (AIO)

Part of: Conductor AI Search & AI Overview Benchmarks: Nonprofit (editions 2026)

Subtopic: AI Overviews & AI search

Question: Which AI engines account for the most citations of an industry's domains? · 1 study · 2 articles

About: Google AI Mode, Google AI Overviews, Perplexity, ChatGPT

Published: 2026-08-12

Coverage: Conductor Academy

Tier Bindependent

Search Signals Index, Aug 10-31 2026: index -2.4% across 1,842 companies; Amazon -16.9% to a record low

Growth Memo's Growth Intelligence Brief #24 tracks SEO visibility for a fixed panel of about 2,600 companies. In the 21 days to August 31, 2026, the index slipped 2.4%, Amazon fell 16.9% to its lowest reading since tracking began, the IRS lost 59.6% as Google served fewer PDFs, and YouTube gained 8.7%. The window also covered Google's August spam update, so the causes are not isolated.

Red flags: correlation causation, confounded window, vendor metric

Measured: Rankings · SEO visibility index points

Research: August 10-31, 2026

Part of: Growth Intelligence Brief (editions #24)

Subtopic: Other

Question: How is SEO visibility shifting across a fixed panel of companies?

About: Amazon, Google August 2026 spam update, Reddit, YouTube

Published: 2026-09-08

Coverage: Growth Memo

Tier Dagency

Orbit Media survey: marketers reporting strong content results fell to 14% in 2026 (from 20-28%)

In a Siege Media podcast, Orbit Media's co-founder reports that the share of marketers reporting strong content results in Orbit's annual survey fell to 14% in 2026, after a decade at 20-28%, and that AI use showed no correlation with performance. The sample size is not given in the episode summary, and the results are self-reported.

Red flags: correlation causation

Conflict of interest: Orbit Media is a content agency that runs the survey.

Research: 2026 survey edition; 13 years of surveys

Part of: Orbit Media annual blogger survey (editions 2026)

Subtopic: Content depth & updates

Question: What share of content marketers report strong results?

Published: 2026-10-20

Coverage: Siege Media Blog

Tier Bvendor

ChatGPT query fan-out across 66.6M responses: only 14.4% search the web, ~1.6 queries each

Ahrefs analysed 66.6 million free, logged-out GPT-5.6 Luna responses and their 15 million fan-out queries from August and September 2026. Only 14.4% of responses contained fan-out queries, and 6.8% of fan-outs used the site: operator, mostly on brand sites. In separate Brand Radar citation data (2.4-3.8M US answers a month, March-September), Reddit fell from about 1 in 5 cited answers to 1 in 43. The citation data comes from Ahrefs' own product, and there is no control group.

Red flags: vendor metric

Conflict of interest: Ahrefs sells Brand Radar, the AI visibility tool whose data is used and promoted.

Measured: AI citations · share of responses with fan-out queries; site: share; citation share

Research: August-September 2026 (fan-outs); March-September 2026 (citations)

Subtopic: AI Overviews & AI search

Question: How does ChatGPT search the web (query fan-out) before answering?

About: ChatGPT, Ahrefs Brand Radar

Published: 2026-10-12

Coverage: Ahrefs Blog

Tier Dvendor

Ahrefs signup survey: 21.7% of signups credit ChatGPT or Claude, more than YouTube

Ahrefs reports that customers self-attribute 21.7% of its signups to ChatGPT and Claude, and that AI search visitors convert to a free account at 6%, about four times its organic rate. The figures come from one company's own analytics and onboarding survey. The collection period is unstated and there are no controls, so treat them as one company's experience.

Red flags: correlation causation, anecdote generalisation

Conflict of interest: Ahrefs measures its own signups while promoting its analytics products.

Measured: Conversions & revenue

Research: unstated

Subtopic: AI Overviews & AI search

Question: What share of a company's signups or visitors comes from AI assistants?

About: ChatGPT, Claude

Published: 2026-10-06

Coverage: Ahrefs Blog

Tier Bvendor

AI Overviews on branded Google searches: 61.3% in July to 73.3% in September 2026 (82.91% on Sept 29)

Ahrefs analysed 232.8 million US desktop SERP crawls from July to September 2026. The monthly share of branded keywords with an AI Overview rose from 61.3% to 73.3% (non-branded: 73.5% to 78.3%), and a single-day reading on September 29 hit 82.91%. In a separate 100-brand panel, 48 of 63 AI Overviews (76.2%) sat in position one. The headline 83% is that single-day figure, not the monthly average.

Red flags: cherry pick, headline overclaim

Conflict of interest: Ahrefs sells SEO tools that track AI Overviews.

Measured: AI Overview presence · share of SERPs showing AI Overview

Baseline: 61.3% of branded keywords in July 2026

Research: July-September 2026

Subtopic: AI Overviews & AI search

Question: How often do Google AI Overviews appear in search results?

About: Google AI Overviews

Published: 2026-10-02

Coverage: Ahrefs Blog

Tier Bvendor

AI Overviews on 400 commercial keywords: 45.75% to 88% in a year; 'review' queries now ~90%

Semrush tracked a fixed panel of 400 commercial-modifier keywords (banking, software, ecommerce) monthly from September 2025 to August 2026 using US desktop data. AI Overview coverage rose from 45.75% to 88% overall and to 97.08% when weighted by search volume. 'Reviews' searches went from about 30% to about 90%. How the keywords were chosen is not explained.

Red flags: vendor metric, selection bias

Conflict of interest: Semrush sells AI Overview tracking tools.

Measured: AI Overview presence · share of keywords triggering AI Overview

Baseline: 45.75% in September 2025

Research: September 2025 - August 2026

Subtopic: AI Overviews & AI search

Question: How often do Google AI Overviews appear in search results?

About: Google AI Overviews

Published: 2026-10-21

Coverage: Semrush Blog

Tier Dagency

Zapier case study: cited in ~80% of its bottom-funnel AI prompts; 25% of signups attributed to AI

Siege Media reports, in a podcast case study, that Zapier appears in roughly 80% of the bottom-funnel prompts it tracks and ranks fifth among cited B2B domains. Per Zapier's own numbers, 40% of visitors find it through LLMs and 25% of signups are attributed to AI. This is one hand-picked success story with no before/after comparison and no stated collection period.

Red flags: correlation causation, anecdote generalisation, selection bias

Conflict of interest: Siege Media sells GEO services and promotes its book.

Measured: AI citations

Research: unstated

Subtopic: AI Overviews & AI search

Question: What share of a company's signups or visitors comes from AI assistants?

About: Zapier

Published: 2026-09-29

Coverage: Siege Media Blog

Tier Bagency

1,000 transactional prompts: 52.5% of the most-cited domains in LLM answers are affiliate or sponsorship sites

Siege Media and Peec AI found that 21 of the 40 most-cited domains across 1,000 transactional prompts were affiliate or sponsorship sites rather than brands. Siege's AI buying-advice study puts Reddit in 62% of bottom-funnel LLM responses. The collection period is unstated, method details are thin, and the claim that category leaders lose ground rests on argument rather than data.

Red flags: correlation causation, cherry pick, headline overclaim, vendor metric

Conflict of interest: Siege Media sells GEO and digital PR services; Peec AI sells AI visibility tracking.

Measured: AI citations

Research: unstated

Subtopic: AI Overviews & AI search

Question: Which kinds of domains do LLMs cite most for buying-stage prompts?

About: Reddit, Peec AI

Published: 2026-09-25

Coverage: Siege Media Blog

Tier Bindependent

Hidden AI traffic in Direct: only 0.003% of Direct sessions across 12 sites; no reliable AI-visibility link in 632 tests

Growth Memo analysed 12 GA4 properties from January to September 2026. Total Direct traffic rose 140.9% while engaged Direct stayed flat (-1.0%), which points to bots, and sessions with AI traces made up only 0.003% of Direct. Across 632 lagged correlation tests on 4 brands, with shifted-week placebo checks, no reliable link between AI visibility and later Direct or branded traffic held up. The sample is small and skews B2B.

Measured: AI referral traffic · AI-attributable direct traffic; correlation of AI visibility with direct/branded traffic

Research: January-September 2026

Subtopic: AI Overviews & AI search

Question: Does higher AI visibility raise later Direct or branded search traffic?

About: Google Analytics 4

Published: 2026-10-19

Coverage: Growth Memo

Tier Bvendor

LLMs.txt across ~300,000 domains: 10.13% adoption, no measurable effect on AI citations

SE Ranking analysed nearly 300,000 domains and found that only 10.13% have an LLMs.txt file, at similar rates across traffic tiers. An XGBoost model of AI citation frequency became more accurate once the LLMs.txt feature was removed, which suggests the file has no detectable effect. This is an observational null result, the data period is unstated, and method details are thin.

Red flags: headline overclaim

Conflict of interest: SE Ranking sells SEO and AI visibility tools.

Measured: AI citations

Research: unstated

Subtopic: AI Overviews & AI search

Question: Does having an LLMs.txt file raise a domain's AI citations?

About: LLMs.txt

Published: 2026-10-15

Coverage: SE Ranking Blog

Tier Cagency

Local SEO overhaul for a French paving firm: ~100 to 1,100+ monthly visits and 2-5 to ~12 quote requests in 3 months

Organic Pareto reports, in a case study published by SE Ranking, that rebuilding OTPE's priority service, quote and area pages lifted organic visits from about 100 to over 1,100 a month within three months. Quote requests rose from 2-5 to about 12 a month and added an estimated EUR 200,000 in revenue. It is a single site with no control, the revenue had no tracked baseline, and the year is unstated.

Red flags: headline overclaim, no baseline, selection bias

Conflict of interest: Agency case study published by SE Ranking as promotion for its tools.

Measured: Organic traffic

Baseline: ~100 organic visits/month, 2-5 quote requests/month

Research: September-December (year unstated, likely 2025)

Subtopic: Content depth & updates

Question: Does a focused SEO overhaul of priority pages raise organic traffic and leads? · 1 study · 2 articles

About: OTPE

Published: 2026-08-25

Coverage: SE Ranking Blog

Tier Cagency

Content clusters for a French cruise site: 2.7x organic traffic and 20x qualified leads in 9 months

Organic Pareto reports, in an SE Ranking-published case study, that building content clusters around OKCroisiere's flagship ships raised organic traffic by 50% in four months and 2.7x over nine months, with qualified leads up 20x. Only multipliers are given, with no absolute baseline. The engagement dates are unstated and there is no control, so the result is one vendor-promoted case.

Red flags: correlation causation, no baseline, selection bias

Conflict of interest: Agency case study published by SE Ranking as promotion for its tools.

Measured: Conversions & revenue

Research: unstated (9-month engagement)

Subtopic: Content depth & updates

Question: Does a focused SEO overhaul of priority pages raise organic traffic and leads? · 1 study · 2 articles

About: OKCroisiere, ChatGPT

Published: 2026-08-25

Coverage: SE Ranking Blog

Tier Bvendor

AI Overviews in France at launch: 52.63% of 100,000 keywords, about twice the first-month rate in Germany and the Netherlands

SE Ranking analysed 100,000 French keywords across 20 niches shortly after the July 22, 2026 launch and found AI Overviews on 52.63% of searches. First-month rates in Germany, Spain and the Netherlands were 26.56-33.23%. YouTube appeared in 51.31% of French AI Overviews, and 60.46% of cited pages ranked in the organic top 10. The study is descriptive and has no control.

Conflict of interest: SE Ranking sells AI Overview tracking tools.

Measured: AI Overview presence · share of keywords triggering AI Overview

Research: shortly after July 22, 2026 launch

Subtopic: AI Overviews & AI search

Question: How often do Google AI Overviews appear in search results?

About: Google AI Overviews, YouTube

Published: 2026-08-18

Coverage: SE Ranking Blog

Tier Bagency

GPT-5.6 Luna vs GPT-5.5 on 346 matched prompts: site: searches 22% to 76%, inline citations -12.9%

Seer Interactive compared the two ChatGPT models on 346 matched prompts from six industries. Total fan-out queries barely changed (733 vs 729), but site: searches rose from about 22% to 76% of fan-outs, consulted sources fell 43%, and inline citations fell 12.9% (2,054 to 1,790). The exact run dates are unstated (after Luna became the default on July 9, 2026), and the explanations go beyond the data.

Red flags: headline overclaim

Conflict of interest: Seer Interactive sells AI search visibility services.

Measured: AI citations

Baseline: GPT-5.5 on the same 346 prompts

Research: after July 9, 2026 (unstated exact dates)

Subtopic: AI Overviews & AI search

Question: Does GPT-5.6 Luna cite fewer sources than GPT-5.5?

About: ChatGPT, GPT-5.6 Luna, GPT-5.5

Published: 2026-09-24

Coverage: Seer Interactive

Tier Bagency

AI answers about one brand: platforms declined 31.2% of prompts but were inaccurate only 2.8%

Seer Interactive ran 1,562 branded prompts about itself daily on six AI platforms from May 1 to June 9, 2026 (28,123 responses). When the platforms answered, they were right 95.9% of the time. They declined 31.2% of prompts on average, and 80.3% of comparative prompts. All prompts concern one brand, and Seer labelled accuracy itself.

Red flags: anecdote generalisation, selection bias

Conflict of interest: Seer measured AI answers about its own brand and sells GEO services.

Measured: AI share of voice & mentions

Research: May 1 - June 9, 2026

Subtopic: AI Overviews & AI search

Question: How accurately do AI platforms answer prompts about a brand?

About: Google AI Mode, Google AI Overviews, ChatGPT, Claude, Gemini

Published: 2026-08-27

Coverage: Seer Interactive

Tier Bvendor

Organic rank vs AI Overview citations (168M citations): position 1 cited 24.9%, 57% of citations from outside the top 10

Conductor measured 167,867,680 US AI Overview citations from June 15 to August 15, 2026 against the organic results of the same searches. The citation rate falls from 24.9% at position one to 10.2% at position ten, and 57% of AIO citations go to pages outside the organic top 10. The result is a correlation, which the study itself says it does not prove as causation.

Red flags: correlation causation

Conflict of interest: Conductor sells AI visibility and SEO tracking software.

Measured: AI citations

Research: June 15 - August 15, 2026

Subtopic: AI Overviews & AI search

Question: Are higher-ranking organic pages cited more often in AI Overviews?

About: Google AI Overviews

Published: 2026-09-04

Coverage: Conductor Academy

Tier Dagency

Same ChatGPT prompt rerun: 14 to 40 retrieved URLs in two weeks, zero shared queries 8 minutes apart

Seer Interactive reports that rerunning one identical 148-character prompt on September 3 and 16, 2026 changed the exposed search data. Pipe-style instructions dropped from 4 to 0, unique retrieved URLs rose from 14 to 40, and 34 of the 40 were not cited. Two runs 8 minutes apart shared no exact query strings. This is a single-prompt test with a few captures.

Red flags: anecdote generalisation

Conflict of interest: Seer Interactive sells AI search visibility services.

Research: September 3-17, 2026

Subtopic: AI Overviews & AI search

Question: Does ChatGPT search the same way when an identical prompt is rerun?

About: ChatGPT

Published: 2026-10-22

Coverage: Seer Interactive

Disputed (verdict and skeptic disagree; both grades shown)

Tier Dindependent

disputed

Updating stale pages and requesting indexing flipped a wrong AI Overview answer within about an hour (one brand)

The Semrush blog reports that Feature Upvote's founder updated old blog posts claiming the product had no REST API and submitted them for indexing on July 28, 2026. Within about an hour, Google's AI Overview answered that the API existed, and it was consistent across query variants a few days later. This is a single, hand-checked case. Announcement posts went out in the same window, so the fix cannot be isolated.

Disputed: no quoted number found in the text.
Skeptic's doubts: One brand's AI Overview answer flipped from no to yes after page updates and indexing requests, checked by hand in a few browser tabs, but an announcement post, an X post and a newsletter went out in the same window, and the ending generalises that wrong answers 'can be fixed'.

Red flags: anecdote generalisation, confounded window

Conflict of interest: Published on the Semrush blog, which promotes Semrush AI visibility tools.

Research: July 2026

Subtopic: AI Overviews & AI search

Question: Does updating outdated pages and requesting indexing correct a wrong AI Overview answer?

About: Google AI Overviews, Feature Upvote

Published: 2026-10-21

Coverage: Semrush Blog