The State of Search

Can I block AI training without losing search traffic?

settled The record agrees. A change here would itself be news.

With Google and OpenAI, yes — both state in their own documentation that the training and search crawlers are independent. Anthropic runs separable crawlers but publishes no equivalent guarantee. In every case the block is a request rather than a lock — and on Cloudflare-proxied sites, the layer enforcing it can undo the operator's own separation: taking Cloudflare's September 15 default blocks Google's, Apple's and Bing's search crawlers too, because those crawlers are multi-purpose.

Verified against the record 2026-10-12 · record last moved October 11, 2026

The three operators the question actually turns on each run a search crawler and a training crawler that can be addressed separately in robots.txt. Two of them say in public what blocking one does to the other; the third does not.

Compliance is a separate question from policy. TollBit measured roughly 15% of the AI page fetchers it tracked in the first half of 2026 reaching European publisher URLs marked disallowed, with OpenAI's ChatGPT-User among the bots most often crossing blocks aimed specifically at it. No operator publishes a compliance rate against its own declared behavior, so a block is worth logging against rather than trusting.

Since August 21, 2026 a fourth party sits in this decision for Cloudflare-proxied sites. Bot Preference Sync conditions continued access for crawlers that both search and train on four public disclosures, which converts what was an operator's private practice into something a publisher can check.

The standard underneath any of this is itself still being drafted. Revision 08 of the IETF's AI Preference Vocabulary, published September 16, 2026 and co-authored by Google's Martin Thomson, split its single "AI Training" section into separate "AI Training" and "AI Use" sections while keeping "Search" as its own category — giving sites a more precise vocabulary to state the training-versus-search distinction this claim already tracks operator by operator. The draft expires March 18, 2027 unless renewed, and no operator has said whether it will implement the new categories.

A second measurement, at site-declaration scale rather than per-fetch, adds to the compliance gap TollBit found — and surfaces a migration failure distinct from any operator's crawling behavior. HasData tested 10,894 domains directly and found 39.5% of 592 sites with a declared GPTBot ban in robots.txt still served GPTBot a live page. The same study found 118 sites lost their existing Cloudflare-managed opt-out outright when Cloudflare retired its older "Managed Robots.txt" and "Block AI Bots" features on September 15, 2026 — the replacement, Bot Preference Sync, leaves robots.txt blank until the site owner reviews and confirms new rules, a step Cloudflare's own announcements didn't flag to existing customers.

aeo · engines · law-money

Is AI-crawler access now a paywall rather than a block?

moving Actively changing — a rollout, a deadline, or a live fight.

On the Cloudflare-proxied publishers the wire tracks, yes for some — access is starting to split by whether the crawler's operator pays rather than by whether the site allows AI crawling at all, and the split is now moving in both directions on the same sites within days. Four large publishers block OpenAI's GPTBot outright while continuing to bill Anthropic's ClaudeBot; Perplexity's crawler has now lost that paid option at one of them. Other large sites reject the split entirely and block every crawler, paying or not — though one of those blocks reversed itself four days later even as its stated robots.txt policy did not change. A fourth capture adds a new failure mode on top of the split: a publisher's own robots.txt can say one thing while Cloudflare's enforcement in front of it does another.

Verified against the record 2026-09-30 · record last moved September 30, 2026 — nothing has moved it in 14 days

A different kind of payment surfaced September 16, on the answer side rather than the crawler side. Google confirmed an early-stage pilot that pays a small set of publishers when their content "contributes significantly" to an answer in Gemini, AI Overviews or AI Mode — content that only confirms a fact or shows up as a post-answer link doesn't qualify. Participants see a monthly earnings figure inside a Search Console panel and can opt out any time; Google hasn't disclosed a rate or eligibility criteria, and sources described the earnings calculation as "quite black box" with early payouts small next to ad revenue. It meters something different from the Cloudflare split above: Pay Per Crawl charges for access before an answer is generated, this pays for contribution after one is.

The wire runs its own crawler access panel, probing 55 publisher domains daily as GPTBot, ClaudeBot, PerplexityBot and a browser user-agent. Between the September 12 and 13 captures, three People Inc. properties (Allrecipes, Investopedia, People) and The Atlantic made an identical change: GPTBot and the browser probe went from a 200 to a Cloudflare-managed 403, while ClaudeBot and PerplexityBot kept drawing the same 402 Pay Per Crawl response as before — the same site charging two crawlers for the access it now refuses a third.

That split sits on top of Cloudflare's Pay Per Use mechanism, announced July 1, 2026, which bases publisher compensation on how AI search actually uses a page rather than a flat per-crawl fee.

Not every large site is taking the paywall option. Quora hardened the same day from a redirect to a flat 403 for all four probes — GPTBot, ClaudeBot, PerplexityBot and the browser alike — and its robots.txt, readable to the panel for the first time, disallows all three crawlers by name and bars AI-training use outright. Reddit moved separately, on its own infrastructure rather than Cloudflare's, adding a 403 to browser and PerplexityBot requests that had drawn a 200 the day before.

Three days later the panel caught a second large publisher rejecting the split outright, and one moving the other way on a single crawler. El País's homepage flipped from a 200 to a 403 for GPTBot, ClaudeBot and PerplexityBot alike between the September 13 and 16 captures. Stack Overflow moved just Anthropic's ClaudeBot — off Cloudflare's Pay Per Crawl paid track and onto a flat 403 block — while GPTBot and PerplexityBot stayed behind the same Cloudflare challenge as before: the panel's first capture of a crawler moving from paid access to blocked rather than the reverse.

This is now two capture days from the wire's own panel — September 13 and September 16 — not yet corroborated by another outlet's measurement.

A third capture, September 20, moved two of the four September-16 findings in different directions. El País's homepage returned to a 200 for both a browser request and ClaudeBot — reversing its own September 16 block — even though El País's robots.txt still disallows ClaudeBot by name; GPTBot and PerplexityBot there remain blocked. The Atlantic moved PerplexityBot off Cloudflare's Pay Per Crawl track onto a flat 403, while keeping Anthropic's ClaudeBot on the paid track unchanged — the first capture of a publisher ending one crawler's paid access while leaving another's in place. Separately, Politico's robots.txt became readable to the panel for the first time, disclosing disallow-all rules for GPTBot, ClaudeBot and PerplexityBot by name, and its homepage dropped Cloudflare's browser challenge (403 to 200) while continuing to challenge all three bots; The New York Times' homepage began returning a 403 to the panel's browser probe, now served by the bot-detection vendor DataDome rather than the Envoy server it ran on September 16, with its existing bot blocks unchanged.

A fourth capture, September 27, caught stated policy and live enforcement pointing opposite directions at the same site rather than a new split. The Atlantic's robots.txt became readable again for the first time since the wire's September 13 capture — every capture in between had returned a 403 — and now names GPTBot, OAI-SearchBot and ChatGPT-User as allowed while disallowing Claude-SearchBot, anthropic-ai and PerplexityBot outright. Its live homepage still returns Cloudflare's 403 to GPTBot itself, unchanged since September 13: the published policy and the enforced one now disagree, the mirror image of El País's September 20 reversal, where enforcement reopened ahead of an unchanged robots.txt. Search Engine Land's robots.txt, readable throughout the panel's tracking, started returning the same Cloudflare challenge on September 27, closing off the one part of the site the panel could still read after its homepage went behind an identical wall on September 16. Separately, NPR's robots.txt added a disallow-all rule for Claude-SearchBot, a token distinct from the ClaudeBot and anthropic-ai rules already there, and Press Gazette's homepage began returning a 403 to GPTBot and ClaudeBot while continuing to let PerplexityBot through unblocked — a third access pattern, alongside outright blocking and Pay Per Crawl billing, of admitting one operator's crawler while shutting out two others.

The Google pilot's black-box rate got its first number September 30. The Information reported that the roughly 100 publishers in the pilot are paid an average of under 0.1% of their advertising revenue — payouts ranging from under $1,000 over several months to more than $1 million a year — with participants still unable to explain how Google calculates the figure month to month. It remains a narrower mechanism than the Cloudflare split above: a disclosed rate and publisher count now exist, but eligibility criteria and the calculation method do not.

tools · aeo

Why don't AI visibility tools agree with each other?

settled The record agrees. A change here would itself be news.

Because they measure different engines, query sets and windows. Two trackers reporting opposite trends for the same subject are usually both correct, and the disagreement is methodological rather than factual — which is why the fact of disagreement stays stable while every number under it moves.

Verified against the record 2026-10-14 · record last moved October 13, 2026

The clearest worked example the record holds is Reddit in August 2026. BrightEdge reported on August 16 that Reddit's AI citation surge had plateaued at a baseline several times higher than a year earlier, and that Google AI Overviews accounts for roughly 88% of Reddit's AI citation volume. Three days later a tracker reported an 86.4% collapse. Both were right.

Promptwatch's own per-engine breakdown resolved it: ChatGPT Search fell 86.4% (3.83% to 0.52%), while AI Overviews slipped 11.3% and AI Mode 30.5% over the same window. The surface carrying ~88% of the volume barely moved. A decline reported widely as Reddit disappearing from AI search was one engine changing how it assembles background queries.

A separate, single-source measurement of that query-writing step shows why the change moved so much: brands ChatGPT named in its own background search query reached the final answer 68.9% of the time, against 2.1% for brands surfacing only on retrieved pages — a roughly 33x gap — and only 3.1% of 3,554 retrieved pages earned a citation at all. Read against Promptwatch's finding, the two measurements describe the same mechanism nine days apart: the query-writing step, not the page, is what the Aug 8 change actually altered.

Engine-to-engine spread inside a single study is wide enough to swamp most between-study comparisons: Writesonic measured "ghost citation" rates from 52% (Perplexity) down to 19% (Microsoft Copilot) across seven engines and roughly 16 million brand appearances. A figure quoted without its engine is not a figure.

The model underneath moves fast enough that naming the current one dates the paragraph. AI Mode's underlying Flash model has swapped four times since July 2026 — 3.5 Flash-Lite, then 3.6 Flash, 3.7 Flash on August 14, and 3.8 Flash on September 2 — each arriving roughly three weeks apart, which is faster than most visibility studies are conducted and published.

A day after the 3.8 Flash swap, two SEO practitioners reported AI Mode responses running on it showing markedly fewer citations than before — screenshots of several top-of-funnel queries returned no source links at all. Google confirmed it the next day: Robby Stein, Search's VP of product, said on X that citations were not working as intended and a fix is coming. Still no query sample or scope from either side, but the mechanism is now acknowledged rather than anecdotal — a bug, not a redesign.

A same-day finding complicates measurement further, upstream of any tracker: ChatGPT doesn't answer from one index. Peec AI reports OpenAI has built its own retrieval system, internally named Labrador, with separate vertical indexes for web, PDF, video and news — but a since-removed `result_source` field visible in ChatGPT's server-side events between May 21 and July 21, 2026 showed answers still drawing on Google, Microsoft's Bing, and two third-party scraping services alongside it. A dashboard built to approximate ChatGPT exposure from Bing rankings alone is reading one of at least four backends, not the whole surface.

A later study puts a number to that gap for the most common tracking approach. Replaying the same queries through OpenAI's Responses API and Google Search — the two external proxies most GEO tracking tools lean on in place of real ChatGPT Web citations — recovered only about a quarter of the exact URLs ChatGPT Web actually cited, combined. A model combining both signals tracked hostname- and entity-level visibility far better (AUC above 0.74 for GPT-5.3 and GPT-5.5) than exact-URL matching, which argues for weighting domain- and entity-level visibility over URL-level counts in any tool built on these proxies.

aeo · research · tools

Once I'm cited, do I stay cited?

settled The record agrees. A change here would itself be news.

Not reliably. For most prompts a small stable core of domains holds while everything around it rotates fast — the largest published measurement found Google replaces 56% of its AI-answer citations every week, ChatGPT as much as 74%. The instability is structural, not a sign anything is broken, and three separate measurements had already caught pieces of the same mechanism before this one supplied the base rate.

Verified against the record 2026-09-30 · record last moved September 30, 2026

SISTRIX re-sampled 82,619 prompts weekly for 17 weeks (1,548,213 snapshots, six countries, three surfaces) and found the headline churn sits on top of real structure: 86% of prompts keep a stable core of a few domains, with everything outside it rotating at 89% per week. On Google AI Overviews specifically, 53% of prompts saw no source change at all across the full window — aggregate churn and per-prompt permanence are both true, describing the tail and the core of the same answer.

The structure has consequences the record had not connected before. AI Overviews and AI Mode — one company, the same query, the same week — cite different domains 83% of the time, so a citation-share number measured on one Google surface is not a reading of the other. And permanence varies sharply by content type: only 1.4% of cited news articles held their spot across all 17 weeks, against 43% of brand queries that kept the brand's own domain present throughout even as the co-citations beside it rotated at 70% per week.

Three items on this wire measured pieces of this before the wire had a name for it. Kevin Indig's Search Signals analysis found AI Mode's publisher mentions falling nearly by half in thirty days while retailer mentions tripled, across 15 of 20 verticals — the engine's preference changing. Promptwatch traced Reddit's 86.4% four-day collapse in ChatGPT citations to a change in how the engine assembles its background queries — the retrieval method changing. And Profound shipped a feature called Citation Decay on August 13, 2026 that tracks week-over-week citation counts per URL — a vendor building the instrument this research argues is necessary. Three different mechanisms; SISTRIX supplies the base rate they were all sampling from.

A model swap can also break citation outright rather than just shift its rate. Two days after Gemini 3.8 Flash rolled out to Google AI Mode on September 2, SEO practitioners found some top-of-funnel queries returning no source links at all; Google's Search VP Robby Stein confirmed it was a citation bug on September 4 and said a fix was coming. It is the sharpest confirmed instance yet of the standing caveat that citation share moves when the model underneath does — this time to zero, briefly, on one model.

A follow-up BrightEdge measurement narrows the mechanism to a single vertical and a single engine, and answers a question the churn rates alone couldn't: what happens to a slot once it's vacated. Reddit's share of ChatGPT's Education-category citations fell from 17.2% to 0% over two weeks in August, and the domains that took its place — state.gov, asu.edu, apa.org, aacnnursing.org — still held those positions five weeks later. The slot didn't revert to variety; it was claimed by a narrower, more institutional set and stayed claimed. BrightEdge attributes the drop to Reddit's robots.txt policy; this wire's own August 27 reporting found a ChatGPT query-behavior change that preceded the decline by about six days, so the cause is still unsettled even as the outcome — persistence, not reversion — replicates.

Profound's Citation Decay tracker, announced Aug. 13, 2026, produced its first published measurement on Sept. 30: across 883,000 pages tracked over a year on seven engines, the median page's citation share loses half its peak within 11 days, with the full rise-to-half-peak arc running 42 days. This is a different axis from the weekly substitution rates above — a per-page decay curve rather than aggregate turnover — and durability on it is engine-specific: a sustained Claude citation is 2.8 times as likely as a ChatGPT one to still hold half its peak two months out, and a page cited by five or more engines is 5.5 times as likely as a single-engine page to stay above half its peak after eight weeks.

aeo · research · tools

What can a publisher actually control about AI answers?

moving Actively changing — a rollout, a deadline, or a live fight.

Appearance and training are separate controls and no single lever covers both. Google now offers the only control that removes a site from AI Overviews and AI Mode without touching ordinary Search — it reached all sites worldwide on August 31, 2026, and it does not cover training.

Verified against the record 2026-08-31 · record last moved August 31, 2026 — nothing has moved it in 43 days

Four levers exist and each has a different scope. Choosing wrongly costs more than doing nothing: three of the four take ordinary Search results with them.

The price of the one that does what publishers asked for is stated plainly by Google — opt out and "you won't receive any traffic or impressions from these features" — against AI Overviews appearing on 43% of Google searches in July 2026, up from 15% a year earlier.

A regulator is already pushing on the gap. The UK CMA's Publisher Conduct Requirement, imposed June 3, 2026 and legally binding under the digital markets regime, requires Google to let publishers opt out of training, fine-tuning **and** the grounding behind AI Overviews and AI Mode, with nine months to comply. The Search Console control answers one of those three.

Google's own guidance has not caught up with its own control. "AI features and your website" — the Search Central page telling publishers how to approach inclusion — named `nosnippet` 18 times and the Search generative AI control zero times when checked on August 27, 2026. A publisher following the documentation rather than opening Search Console would not learn the control exists.

google-search · aeo · law-money

Does llms.txt do anything?

thin One measurement or one source, without corroboration.

For Google, no — its documentation says outright that no AI text file is needed to appear in its AI features. For every other major engine there is no published answer either way, and the one large-scale measurement found 97% of llms.txt files were never read by an AI crawler.

Verified against the record 2026-10-05 · record last moved October 5, 2026

The evidence for whether the file is read is one vendor study and one vendor's documentation. That is thin for a tactic this widely recommended, and the thinness is the finding: nobody has published a reason to make the file, and only one company has published a reason not to.

Ahrefs, across roughly 15 million data points, found 97% of sites' llms.txt files were never read by an AI crawler, and that adding schema markup produced no measurable lift in AI citations across 1,885 pages over 30 days. The same research found 88.46% of AI citations still traced to pages in the general search index — the tactic that works is the old one.

The spec itself is still moving. Jeremy Howard published a version 2 update on August 10, 2026, its first revision since the format launched in 2024, adding `rel="alternate" type="text/markdown"` and `rel="describedby"` link relations so agents can find a page's Markdown version.

Google's John Mueller added a behavioral data point rather than a measurement: on his own test sites, "the only crawlers who claim to accept markdown are SEO tools" — no major AI bot did. He recommends logging accept headers before building a markdown version at all, which is the same conclusion the Ahrefs read-rate finding points to by a different route.

Common Crawl's analysis of 584,107 llms.txt files in its July 2026 archive found 68% were produced by templates or SEO plugins — Wix alone accounts for 41% of the corpus — with only about half fully following the spec's structure. It doesn't measure whether crawlers read the files, but it corroborates the same direction from a different angle: adoption is running mostly on CMS-plugin autopilot rather than deliberate authoring.

A third independent dataset agrees on the one point all three can agree on even though none of them measure read rates: adoption is still low. LintLab's October 2026 crawl of the top 1,000 most-visited sites found just 6.6% serving a valid llms.txt file, with nearly two-thirds of attempts returning a 404 instead.

Mueller and Martin Splitt added a third Google statement in the same direction, on the Oct. 1 episode of Google's own "Search Off the Record" podcast: Google's search systems do not process llms.txt as a sitemap at all, and Mueller said of it, "I would not treat that as a priority."

aeo · tools · llmeo

Are AI answers costing me traffic?

contested Credible sources disagree, and the disagreement is unresolved.

It depends on what you publish, and part of the disagreement is instrumentation. Commerce reports AI referrals as growth; nonprofit and health publishers report the steepest declines; general-news outlets diverge — Le Monde reports none seven weeks after AI Overviews launched in France, while People Inc.'s Google search traffic fell 40% — and Google's own figure arrives without methodology. The one non-vendor academic estimate, a University of Washington study of Wikipedia, lands far below the self-reported figures at roughly 5%. The record's first randomized experiment lands much higher: removing the AI Overview nearly doubled outbound clicks in a controlled two-week test. A second randomized experiment, on AI Mode rather than AI Overviews, found the same direction of effect on clicks plus something the first didn't test for: forced AI Mode use also lowered searchers' trust and satisfaction, worst among heavy searchers. Before any of it is compared, 22.4% of AI Overview traffic is misattributed away from Organic Search.

Verified against the record 2026-09-29 · record last moved September 29, 2026 — nothing has moved it in 15 days

The disagreement is real and it sorts by what a site sells. Shopify told its Q2 earnings call that AI-driven traffic and orders tripled year over year, with half of AI-referred sessions landing directly on product pages — 2.5 times the rate of traditional search. Over the same period Blood Cancer UK's leukaemia information page lost 53% of its page views. Both are first-party reports from parties with no incentive to overstate against their own interest.

One widely-cited framing does not survive the record. Zero-click searches hit a record high in June 2026 — 40% of U.S. searches sent a click to the open web — but the same report put Google AI Mode at 0.13% of U.S. search-related visits, usage it described as plateaued or dipping. A surface used by roughly one in a thousand visits cannot be the cause of a decline that broad, whatever else is.

Measurement sits underneath all of it. A nine-month GA4 study of 51,200 clicks found an average 22.4% of AI Overview traffic recorded as Direct rather than Organic Search, and Similarweb found 58.8% of ChatGPT referral traffic lands on publisher homepages even though 65% of cited URLs point two or three folders deep. A citation and the session it produces are not measured in the same place, which is why two honest parties can count the same channel and disagree. See the measurement-disagreement claim.

The response has a price tag now too. Similarweb's Top 100 Media publishers spent an estimated $113 million on Google paid search in July 2026, up 41% year over year and 274% over three years — Forbes alone spent $72.2 million, about two-thirds of the total, even as its own organic traffic fell 26.7% year over year. Buying back the clicks AI answers are keeping is a growing spend line, not an absorbed loss.

Le Monde's CEO complicates the uniform-decline framing directly: seven weeks after AI Overviews launched in France, the outlet logged 24,000 new subscribers in each of July and August 2026 and 7% growth in English-language traffic for H1 2026 versus H1 2025 — while French publishers' group APIG reported a 33-38% search-referral loss industry-wide. Louis Dreyfus's own account of the gap: search drives only about half of a major news site's visits, and the AI Overviews damage he's tracking is concentrated in service-content categories — health, food, automotive — down as much as 50%, not general news.

The only independent, non-vendor measurement in the record lands well under every self-reported figure above. A University of Washington paper estimates AI Overviews cut monthly search referrals to English Wikipedia by roughly 5% since the feature became Google's US default in 2024 — a figure the authors themselves revised down from an initial draft estimate near 15% after switching from daily pageviews to monthly referrals and adding international controls. Google disputes even the smaller number, saying Wikimedia's public referral data mixes in traffic from every search engine and can't isolate Google's effect; the authors counter that other engines carry too little share to account for the gap. The revision is itself informative: the same underlying effect produced a threefold-different headline depending on methodology, which is reason to treat every self-reported percentage above as provisional in the same way.

France now has independent, site-level evidence rather than one CEO's account. Ahrefs tracked 963 high-traffic French domains through Google Search Console for 28 days before and 9 days after AI Overviews launched in France on July 22, 2026: sites most exposed to AI Overviews (over 20% of their queries triggering one) lost 23.1% of CTR, against a 5.7% median CTR loss and 8.8% median click loss across all tracked domains, with nearly 48% of sites losing more than 10% of daily clicks within nine days. This is not in conflict with Le Monde's CEO — it corroborates his own account of where the damage sits. Ahrefs' worst-hit category was health (22.2% AI Overview rate, 20.4% CTR decline), a service-content vertical, matching the categories — health, food, automotive — Dreyfus named as carrying the losses he does see, while general news is where he reports none.

Every figure above is observational — a before/after comparison, a self-report, or a vendor's own tracking. The record's first randomized test changes that: Saharsh Agarwal (Indian School of Business) and Ananya Sen (Carnegie Mellon) randomly assigned 1,065 U.S. desktop searchers to see Google results with or without AI Overviews over two weeks in early 2026. Removing the AI Overview raised outbound clicks per search from 0.37 to 0.62 — a nearly 40% reduction when it's shown — while self-reported search satisfaction held flat at 4.0 out of 5 either way. Citations inside the AI Overview recovered little of what it took: just 7.1% of outbound clicks on AI Overview-bearing search pages came from links inside the AI-generated answer itself, with the rest going to ordinary organic results on the same page — a direct answer to Google's own undocumented "billions of clicks" claim above. The result sits well above the Wikipedia study's ~5% estimate, but it measures a different thing over a different window: all outbound clicks rather than one site's referrals, two weeks of 2026 desktop search rather than a multi-year trend.

A second randomized experiment corroborates the first on clicks and adds a dimension it didn't measure: trust. Stephanie Wang, Jeffrey Gleason, Yakov Bart, Christo Wilson and Danaé Metaxa (Northeastern and Penn) tracked 1,100 participants' Google searches over one week under three conditions — AI Overviews hidden, normal search, and a forced AI Mode experience. Removing AI Overviews increased click-through without reducing satisfaction, matching Agarwal and Sen's clean-removal result; the forced AI Mode condition went further, cutting click-through and also lowering users' trust in, satisfaction with, and sense of the relevance of the results, with effects strongest among heavy searchers. Only 51.1% of AI Overviews could be reliably hidden in the no-AI condition, per the authors, because of Google's changing page structure — a limit on how clean even a randomized removal can be.

seo · google-search · research

What on-page work actually moves AI citation?

settled The record agrees. A change here would itself be news.

The levers that measure are the ones that were already SEO — being in the search index, ranking well, being reachable in plain HTML, and keeping pages current. The AI-specific file formats measure near zero.

Verified against the record 2026-10-14 · record last moved October 13, 2026

The single largest number in the record points back at classic search: 88.46% of AI citations traced to pages already in the general search index, and 76% of the passages Google reused 100 or more times came from pages already ranking first organically. Google's own position is the same — John Mueller said there is "nothing really special you need to do for generative AI responses in search".

That is not the same as saying nothing is specific to AI. Three levers measure, and all three are mechanical rather than semantic: whether a crawler can reach the page at all, how recently it changed, and whether the passage is shaped to be lifted. The AI crawlers are stricter than Googlebot on the first — GPTBot, ClaudeBot and Bingbot found none of a site's JavaScript-injected internal links across 41 days, and recovered 250 pages within 48 hours of the links being converted to HTML.

The tactics named after the category are the ones that do not measure. An analysis spanning roughly 15 million data points found 97% of llms.txt files were never read by an AI crawler, and adding schema markup produced no measurable citation lift across 1,885 pages over 30 days. See the llms-txt-effect claim, which holds that question's own open threads. A separate study re-tested the field's original 2023 citation-lever effect sizes directly against ten modern AI engine families and found none of them moved citations any longer; a content score built from those levers correlated with actual citations at just 0.11 within-query, positioning it as a quality filter rather than a citation predictor.

A third study found why schema and similar signals keep testing null: analyzing roughly 2 million citations across 10,000 pages from 19 B2B SaaS workspaces, domain-level AI authority outweighed the next-strongest page feature sixfold in SHAP value, and positive correlations between citations and schema markup, FAQ blocks or Core Web Vitals reversed or fell to zero once that domain effect was controlled for. The strongest surviving per-page predictor was how closely a page's wording matched real user prompts (beta +0.37), not any traditional SEO signal.

A fourth study supplies the mechanism behind that domain-level confound. Tracking 90 days of ChatGPT, Gemini and Perplexity answers (158,099 answers, 468,037 citations), University of Toronto researchers found that whether ChatGPT's own rewritten search queries retrieved a page within a given site explained 32% of which pages got cited, against just 2.8% for page content and structure — a gap the study calls statistically decisive (95% CI 24-42%). Content still mattered, but mainly through matching the wording of those rewritten queries rather than through FAQ blocks, structured data or page speed, which only showed an effect comparing across different websites, not between pages of the same site.

aeo · geo · seo · research

Where does AI now appear in Google Search?

moving Actively changing — a rollout, a deadline, or a live fight.

On the results page itself, and increasingly around it. AI Overviews sat on 43% of Google searches in July 2026 and on nearly all People Also Ask boxes by September; AI Mode is one tap from the homepage and, in a test, from the results-page search bar; Search Live runs on Gemini 3.8; and Google has tested AI-written summaries atop local business panels, a "Dive deeper" AI overview inside Discover, an "add to calendar" link on event dates, and — on hotel searches — an AI Overview moved below the knowledge panel instead of above results, all unconfirmed. AI is moving from a separate mode into a layer across Search, Local and Discover, not always toward the top of the page. Measured use has not followed yet: AI Mode held at 0.13% of U.S. search-related visits in June.

Verified against the record 2026-10-07 · record last moved October 8, 2026

The surface that matters most by volume is still the one that arrived first. Similarweb put AI Overviews on 43% of Google searches in July 2026, up from 15% a year earlier, and AlsoAsked found them answering 97% of People Also Ask boxes in the first week of September — the last major results-page feature they had not absorbed.

The week of September 14 added four entry points in five days, three of them tests Google has not confirmed. An AI Mode button appeared inside the search bar on the results page, not only the homepage. Local knowledge panels for Google Business Profiles showed an AI Overview label with a "Show more" that opens an AI Mode-style chat — and the AI-written description of the reporter's own business used information from his site that was already out of date. Discover's "Dive deeper" button, which Google described as starting with video cards, carries code flags for article cards too. Search Live, the audio-and-video conversational mode, moved to Gemini 3.8 Live.

Web Guide, the opt-in results layout that groups links by AI-generated subtopic, is still a beta with a working exit as its open question: on September 20 users reported being unable to return to standard results, and some links rendering as "null://null". Most searchers are not in it.

A different kind of shift appeared September 24, on hotel searches: screenshots shared by an SEO professional, reported by Search Engine Roundtable, showed the AI Overview moved below the hotel knowledge panel instead of above search results — the first sign of Google testing AI Overview placement moving down the page rather than adding a new entry point. Google has not confirmed the test.

Use is the missing column. Datos and SparkToro's Q2 2026 report put AI Mode at 0.13% of U.S. search-related visits in June, flat, in a quarter when zero-click searches set a record. Google reports AI Mode's reach in queries and countries, not as a share of search visits; no measurer has published a later share.

The 43% aggregate rate hides a large gap by query type. On searches containing a brand name, Ahrefs measured an AI Overview on 82.91% of US desktop results on September 29, up from a 61.3% July average, and DemandSphere's separate tracker read 82.06% the same day, up from 26.12% on September 1 — a jump DemandSphere says it cannot explain and Google has not addressed. On the bare brand name alone, two measurements disagree on both rate and placement. Researcher Chris Long tested roughly 100 large brands once — Reddit, Salesforce, Amazon, Adobe, Wikipedia, Samsung, Apple, Yahoo among them — and found an AI Overview on 93%, most mid-page or lower. Ahrefs' weekly panel of 100 brand names found 63 by September 30, against 5 of 98 on July 1, and 48 of those 63 sat in first position, above the brand's own site. Google showed none for searches on its own name or on news publishers' names, per Long.

google-search · aeo · seo

Can I buy my way into AI answers?

moving Actively changing — a rollout, a deadline, or a live fight.

You can buy a slot, but not a citation — they are separate games. Ads run on about a quarter of ChatGPT's commercial prompts and nearly one in three commercial AI Mode queries, and only 3.63% of the advertisers whose ads ran were also cited as a source in the answer above them. The ad stack is scaling faster than the reporting under it.

Verified against the record 2026-10-13 · record last moved October 13, 2026

The ad formats keep widening inside AI Mode, not just below it. Google confirmed testing regular Search-campaign text ads — for now limited to exact and phrase match keywords carrying "explicit and direct" purchase intent — inside AI Mode responses, on top of the shopping ads carousel that arrived August 31. Google has not said how long the test will run or how many advertisers are included. A third format surfaced October 2: a sitelinks-style ad spotted within AI Mode results, apparently testing click-through against the regular links already shown there — single-source, and Google has not confirmed this one either.

The paid layer arrived quickly and is still moving. ChatGPT Ads launched in the US on February 9, 2026, reached nine markets by August 17 — its first Spanish- and Portuguese-language ones — and roughly 40 countries two days later when 31 European markets went live at once. The business crossed $1 billion in annualized revenue run rate in under 200 days after launch, per OpenAI, and self-service Ads Manager access opened to India, Europe, the Middle East and North Africa on August 31 — previously a managed-sales and partner-led market. Ads remain limited to Free and Go plans. On September 10, Amazon Ads began piloting a third route to that inventory: buying ChatGPT Ads placements through Amazon's own DSP for select US advertisers, with Amazon managing the buy and OpenAI still controlling how and where ads appear. No timeline was given for expanding beyond the pilot. The next day, Adform became the first European-headquartered name on OpenAI's technology platform partner roster, letting its clients manage ChatGPT campaigns alongside their other Adform-run media; Volkswagen and Vodafone were named as early clients. On September 16, OpenAI moved distribution into the software marketers already run their day in: HubSpot became ChatGPT Ads' first CRM integration and Shopify its first ecommerce integration (U.S. merchants first, other countries from September 23), and OpenAI began testing Sponsored Agents — a labeled ad format that opens a separate conversation with the advertiser's own AI agent instead of sending a click to its site — with select US advertisers. Ads Manager also gained per-platform targeting and reporting (Android, iOS, web, desktop) and 7/14/30-day attribution windows, up from a fixed one day. On September 23, OpenAI added Taiwan, Singapore, Indonesia, Malaysia, the Philippines, Thailand and Vietnam, pushing the platform past 60 countries. On September 28, LiveRamp became ChatGPT Ads' first outside identity partner, expanding from conversion measurement (added in June) into pre-campaign targeting: brands can match RampID-linked CRM, loyalty, behavioral, web and app data into Custom Audiences for targeting and suppression, in 11 of OpenAI's now 63 listed ad countries — neither company named the 11 markets or said whether RampID itself is shared with OpenAI's ad systems. On October 5, OpenAI added 20 measurement, attribution and incrementality partners — including AppsFlyer, DV Rockerbox, Triple Whale and Kantar alongside LiveRamp — plus brand-safety evaluation pilots with DoubleVerify and Integral Ad Science, bringing pixel and conversions-API integrations, an Event Quality Score and configurable attribution windows. The three advertiser case studies OpenAI cited alongside the announcement — a 15.3% lower cost-per-acquisition for WeightWatchers via DV Rockerbox, a 2.3x lift in incremental orders for Dose via WorkMagic, and a 93% new-visitor rate for Portland Leather via Triple Whale — carried no disclosed spend, sample size or campaign dates.

The format itself is moving into the answer, not just below it. As of August 31, 2026, Google began showing a shopping ads carousel inline within AI Mode's response, alongside the existing separate ad unit at the bottom of results — reported from a single ad-format tracker, with Google neither confirming the change nor stating a rollout scope.

Paid presence and cited presence are close to unrelated, which is the finding that matters for anyone treating ads as an AEO shortcut. Across more than 50,000 commercial prompts in 20 niches, 3.63% of advertisers running an ad were also cited as a source in the answer above it, and 14.35% of the ads shown were unrelated to the surrounding conversation. The same separation shows on Google: an analysis of AI Mode text ads found advertisers rarely among the answer's cited sources.

Two coverage figures circulate and they are not in conflict — they use different denominators. SE Ranking found ads on 25.94% of commercial prompts; Adthena, measuring all US queries rather than commercial ones, found 4.47% at an average of 1.06 ad items per response, against a 3.53-item average on Google's AI surfaces. Read the first as ad load on the queries advertisers want and the second as ad load across everything.

A second pair of coverage figures circulates and this one is not reconciled. Adthena counted 12,075 distinct weekly ChatGPT advertisers as of August 31, 2026 — up 65% from 7,306 eight weeks earlier — by sampling which advertisers' ads appear across AI search queries; Sensor Tower's spend-panel estimate for the same window put active ChatGPT advertisers around 1,200. The two trackers count different things (query-sample appearances versus active spend accounts), and nothing in the record says which better answers how many advertisers are actually on ChatGPT.

On Google's side, the ad targeting itself is shifting under the query. Adthena found one- and two-word searches generating more than half of Google AI Overview ad appearances in every market it measured by late August 2026, reversing a June 2026 pattern in which three- and four-word queries had the majority — a sign Google's AI Overview ad matching is moving toward shorter, more generic queries rather than longer ones.

Category concentration is shifting under retail's early lead. Sensor Tower's April-August window found financial services growing from one top-100 advertiser to four of the top ten spenders, and from 2% to 13% of US ad spend, while shopping and retail's share fell from 37% to 21% and travel rose from 5% to 11%. Comscore's new panel-based tracking corroborates the travel shift from a different angle: sponsored links on hotel-related ChatGPT prompts rose from 6% to 24% between March and May, measured from real opted-in user prompts rather than synthetic tests.

Reporting lags the levers throughout. Campaign-level platform targeting shipped in August into an Insights view that still groups results as Mobile and Desktop, so a web-only campaign reports entirely as Mobile. View-through conversions appear as their own column but are excluded from CPA, bidding and billing. Automated bidding became the default for new ad groups carrying no guarantee against a CPA, CPC or ROAS target. Third-party tooling is currently the only cross-platform view of what is actually running.

A new regulated category opened on Google's side. Google confirmed to Adweek on October 12 that it had begun testing ads for certified pharmaceutical advertisers, including prescription drugs, in AI Mode and AI Overviews — limited to the US. Healthcare ads, along with finance and gambling, had been excluded from Google's AI surfaces until a narrower test for select healthcare advertisers began in June. A day later, ad-intelligence firm Adthena said it had already caught a live one: an Eli Lilly ad for an oral GLP-1 weight-loss pill running in an AI Mode answer on October 11, before Google's own confirmation. The ad never names the drug, using an unbranded, help-seeking format that sidesteps the risk disclosures a named-drug ad would carry under FDA rules. Adthena also said Google Ads reporting still gives advertisers no way to separate AI Mode or AI Overviews placements from the rest of a campaign, and that none of the roughly 50,000 US AI Mode ad placements it tracked in August carried a prescription brand.

engines · tools · aeo · google-search

Will the courts stop AI engines from using my content?

moving Actively changing — a rollout, a deadline, or a live fight.

Not so far. The one case that has ended produced a price, not a ban — roughly $3,000 per work — while 143 AI copyright suits remain pending nationwide, 27 against OpenAI and 15 against Microsoft. Courts are reaching opposite results on near-identical arguments. Three sets of defendants have raised a First Amendment (free speech) defense in six months. The Justice Department has filed its first formal position on fair use — the rule that allows some copying without permission, and the question underneath nearly all of these suits — and it sides with fair use. A new suit has added a trademark-dilution claim (that made-up AI answers credited to a newspaper damage its brand), which stands apart from copyright altogether. And the first federal appeals ruling on a copyright question about AI output, from the Ninth Circuit, went to the AI company.

Verified against the record 2026-10-13 · record last moved October 13, 2026

One case has produced a number. The Bartz v. Anthropic settlement over pirated books — a class settlement, covering a whole group of authors at once — became effective August 20, 2026. The group's lawyers (class counsel) now say the first payout — $2,203.56 per work — goes out on or before Nov. 15, 2026. That is later than the Sept. 17 date estimated when the settlement took effect. Combined payout statements go out Sept. 4, and class members then have 30 days to dispute them before money moves. A second payment, funded by a further $450 million Anthropic owes plus interest, is expected to bring the total toward the roughly $3,000 per work estimated earlier. Two appeals filed since challenge only attorneys' fee awards; under the agreement's terms they do not affect the payout. It is the first figure in the record that puts a price on training AI on pirated books. It came from a settlement, not a court ruling, so it binds nobody else.

The legal arguments are splitting rather than converging, sometimes within days. A federal judge threw out most of Google's claims against SerpApi under the law against breaking technical locks (the DMCA's anti-circumvention rule), reasoning that Google's anti-bot system protects ad revenue rather than a copyrighted work. Days later a different judge let Reddit's near-identical DMCA claims against Perplexity and SerpApi go forward. Claims that a company helped others infringe (contributory infringement) have moved the other way, and that route is now closed: the New York Times, Daily News and Ziff Davis had those claims against Microsoft thrown out for good (dismissed with prejudice) after the Supreme Court's Cox Communications decision ruled the argument out.

AI agents that browse the web for a user got their first answer from an appeals court, and it favored the agent. The Ninth Circuit, a federal appeals court, overturned a court order (an injunction) that barred Perplexity's Comet browser from Amazon. It found Comet unlikely to violate the federal anti-hacking law, the Computer Fraud and Abuse Act, because Comet acts on the user's direction rather than accessing servers on its own.

A shared defense is forming at the same time. Musk, Tesla and Warner Bros. Discovery raised a First Amendment (free speech) defense in an AI-copyright dispute in March 2026; Anthropic added one in the Gilbert case on August 20; Perplexity raised one in Reddit's DMCA suit on August 28. Separately, the pressure that has moved fastest is not copyright but competition law. Judge Mehta, who already found Google's search business an illegal monopoly, said in a hearing that Google's use of publisher content in AI Overviews "seems really unfair".

The federal government took a side for the first time on September 1. The Justice Department filed a Statement of Interest — a filing setting out the government's view — in the OpenAI copyright litigation. It argued that training on copyrighted text is a transformative fair use: copying for a new purpose, which the law allows without permission. It called the plaintiffs' "market dilution" theory — that AI outputs harm a work's market merely by resembling its genre — "deeply flawed." A Statement of Interest under 28 U.S.C. § 517 states the government's view; it does not bind the court deciding the case.

A new legal theory arrived September 4, outside copyright entirely. The Seattle Times and Newsday's joint suit against OpenAI and Microsoft adds a trademark-dilution claim under federal trademark law (the Lanham Act). It argues that when ChatGPT and Copilot make up content (hallucinate) and wrongly credit it to the papers, that damages the papers' trademarks (tarnishes their marks). That claim would stand on its own however the fair-use question — whether AI companies may copy without permission — is decided.

The DOJ's position didn't stay in the case it was filed in. On September 4, ROSS Intelligence filed a letter pointing the appeals court to new authority, citing the DOJ's Statement of Interest. ROSS is the defendant appealing in Thomson Reuters v. ROSS Intelligence, one of the earliest fair-use rulings on AI training, now before the Third Circuit federal appeals court. ROSS argued the DOJ's position applies broadly to AI training, not only to generative AI. It is the first sign of the DOJ's position being used outside the case it was filed in — by a party to another case, not yet by a court.

A new track opened September 9, aimed at company boards rather than at the copying itself. Katelyn Gray filed a shareholder derivative suit — a suit by shareholders against the board on the company's behalf — against Satya Nadella, other Microsoft directors and officers, and Microsoft itself. It argues they approved or were exposed to the risk of copyright infringement and misrepresented that risk to shareholders. It is the third such "copyright shareholder derivative" suit against Microsoft's board, and it is expected to be combined with the two before it into one stockholder derivative action. The theory tests the board's oversight of copyright risk, not whether the training itself infringed.

A dispute opened September 11 over whether a promise not to sue can head off a fair-use ruling. Daniel Gilbert is suing Anthropic without a lawyer (pro se) over his book *Hacking World of Warcraft*. Anthropic filed its own claim back against him (a counterclaim), saying its copying was fair use. Gilbert asked the court to throw out that counterclaim (a motion to dismiss). Anthropic opposed him, arguing that his promise not to sue over "training copies" (a covenant not to sue) does not make the counterclaim moot — no longer needing a decision. Its reason: the counterclaim rests on Gilbert's infringement claim, which is still pending, and on his position that the court cannot weigh why Anthropic copied — not on fear of a second suit.

A second appeals-court answer arrived September 16. It again favored the AI company, but for a narrower reason than the lower court gave. The Ninth Circuit upheld (affirmed) the dismissal of a claim that GitHub Copilot broke copyright law by stripping copyright management information (CMI) — the ownership and licensing tags attached to code — when it generated new code. The court agreed that producing a new AI output isn't the same as removing tags from an existing copy. But it rejected the lower (district) court's own reasoning for getting there: it called that court's "identicality requirement" — a test of whether the output was identical to the original — a misnomer, and declined to endorse it. It is the first federal appeals-court (circuit-court) ruling on this DMCA question for AI outputs.

The litigation calendar moved in two directions at once on September 15-16, both in cases against Anthropic. In Concord Music v. Anthropic I, both sides finished their written arguments on summary judgment — a request for a ruling without a trial — with a hearing set for October 21, 2026 before Judge Lee. It is one of only three AI copyright cases nationwide to reach that stage this year, alongside In re Mosaic LLM Litigation and In re OpenAI Copyright Infringement Litigation, and the first of the three scheduled to be heard. Separately, Judge Pitts set a timetable (a scheduling order) for Cambronne v. Anthropic, the case brought by authors and publishers who left (opted out of) the Bartz book-piracy settlement. Non-expert summary judgment motions — requests for a ruling without a trial — are due May 13, 2027, and trial is not until 2028. That is years behind the Bartz class members now waiting on their first settlement payout.

A second newspaper publisher sued OpenAI and Microsoft in as many weeks. Times Publishing Company, the corporate publisher of the Tampa Bay Times, filed suit September 16 in federal court in the Southern District of New York. It alleges copyright infringement; vicarious copyright infringement (responsibility for someone else's infringement); and removal, under the DMCA, of copyright-management information — the ownership and licensing tags on a work. Those are the same claims the Seattle Times and Newsday raised jointly on September 4. Per the complaint's own count, it is the 27th copyright suit against OpenAI and the 15th against Microsoft, out of 143 filed industry-wide, up from the 24-against-OpenAI, 137-total count in the record three weeks earlier. The case is expected to be assigned to Judge Stein and folded into the combined federal case against OpenAI (the In re OpenAI Copyright Infringement Litigation MDL). It would then be paused (stayed) while the summary judgment briefing already underway there — written arguments for a ruling without a trial — plays out.

Anthropic raised a second defense on September 18 against authors who left the Bartz settlement (opt-out plaintiffs), this one about timing rather than fair use. Shakespeare v. Anthropic is a separate case from Cambronne, brought by another group of authors who opted out of the Bartz settlement; it was amended in August to name CEO Dario Amodei personally. In it, Anthropic asked the court to throw out the copyright claims (a motion to dismiss) as too late under the Copyright Act's three-year statute of limitations — the deadline for suing after discovering the infringement — per Chat GPT Is Eating the World, which obtained the brief.

The publishers suing OpenAI pushed back on the government directly on September 28. The New York Times and the other News Plaintiffs, suing in the same combined case the DOJ's Statement of Interest was filed in, filed a response arguing the DOJ's fair-use position contradicts the trial evidence, conflicts with the Copyright Office's own analysis, and leans on an unsupported national-security argument. News/Media Alliance CEO Danielle Coffey said there is no credible argument that OpenAI and Microsoft's use of copyrighted news content is fair use, given the companies' resources to pay for a license instead — comparing it to a utility giving away electricity for free.

law-money

What do Google's own reports tell me about AI search?

moving Actively changing — a rollout, a deadline, or a live fight.

Impressions without clicks, and position as a block. Search Console's AI performance reports, available to every site since August 31, 2026, show impressions across AI Overviews and AI Mode but no click data, and report position for the whole AI Overview rather than for a link — because isolating a per-link position, in Google's words, "is hard to do". Merchant Center's AI Performance Insights adds intent, search-term and attribute views for products. Underneath both, the analytics instrument undercounts: a nine-month study found about 22% of AI Overview traffic recorded as Direct rather than Organic Search.

Verified against the record 2026-10-02 · record last moved October 2, 2026

Search Console reports AI Overviews position "as a block" rather than per link, John Mueller said in September 2026; impression counts do not confirm a user saw the link, and links revealed only behind a "Show more" are undercounted. He asked what better tracking would look like and described no fix in progress.

The AI performance reports and the Search generative AI control reached all sites worldwide on August 31, 2026; the reports show impressions across AI Overviews and AI Mode without clicks.

Merchant Center's AI Performance Insights, launched July 2026, added three sections in September: AI Search Intent, AI Search Terms and AI Attributes — the first first-party view that tells a merchant which product attributes the AI surfaces are missing.

The measurement problem starts below the platforms. A nine-month GA4 study of 51,200 clicks found an average 22.4% of AI Overview traffic recorded as Direct rather than Organic Search, worst at 29.3% in May 2026. Vendors are patching the gap: Peec AI's AI referrals pairs GA4's native AI-referral grouping with its own detection for referrals the default grouping misses.

Google has started closing part of that gap with link tags rather than new reports. Gemini began appending UTM parameters to its outgoing links as of October 2026, confirmed by John Mueller, letting site owners separate Gemini referrals from GA4's Direct bucket for the first time — the tag mostly survives only on desktop web, with roughly 9% pass-through from Gemini's mobile app, and Google hasn't documented when it triggers. A day later, a marketer spotted the same idea applied to the results page itself: a `?st_source=ai_overview` parameter on AI Overview and AI Mode links, which would let site owners separate that click traffic the same way — but Google hasn't confirmed it, and Search Engine Roundtable's own attempt to reproduce it across several browsers didn't surface the parameter.

tools · google-search · seo

I'm getting cited. Am I getting recommended?

settled The record agrees. A change here would itself be news.

Not necessarily — being cited, being named and being recommended are three different outcomes with different drivers, and the record separates them. Across 1,094 ChatGPT categories the most-cited domain was the most-mentioned brand only 20.8% of the time. Citation measures whose page an engine used; recommendation tracks how well the engine already knows your brand.

Verified against the record 2026-10-13 · record last moved October 13, 2026

The three outcomes come apart at every step. An engine can lift a passage from your page and name a competitor in the sentence it supports; it can name you without linking you; it can recommend you having never cited you at all. A dashboard reporting one number cannot tell you which of the three you bought.

Being correctly identified is not the same as being offered. A Recrawled test of 40 UK B2B software firms across ChatGPT, Gemini, AI Overviews and AI Mode found 9 never appeared in any of 24 shortlist-style answers each was eligible for, even though all 40 were named correctly when searched directly by brand. The exact wording of the buyer's question moved results far more than any site-level factor did: a UK-specific shortlist question named 63% of the 40 firms, a market-leader-versus-rival question named none, and a firm's review-site volume showed no correlation with how often it was named.

What moves recommendation is mostly not on-page work. Models searched for brands they already knew 3.2 times more often than unfamiliar ones — 55.7% versus 17.4% across 66 buyer prompts — and 63% of brand-specific searches surfaced one of each model's five most-familiar brands. Where competing products were otherwise identical, the well-known brand was recommended 100% of the time.

Two findings look like they disagree about reviews and do not. A market census of 4,776 venues found star rating had no effect on whether a venue appeared at all; a controlled study found a rival needed less than a 0.1-star edge to overturn an incumbent's recommendation. Those are different stages of the same funnel — rating does not get you into the candidate set, and decides between candidates once you are in it. Having a business website nearly doubled the odds of the first.

The uncomfortable finding for this discipline sits in the same study: when every brand adopted the same marketing tactics, the incumbent's advantage collapsed from a payoff of +0.802 to +0.007. Tactics that everyone runs stop being an edge and become the price of entry — while brands that opted out received no recommendations at all.

aeo · research · engines