Latest Technical SEO News - Page 2
Crawling changes, algorithm updates, rendering fixes, and structured data developments as they happen.
-
Google changed AI Mode's model two days into the core update
Gemini 4 Hydrogen is the new AI Mode default and answers the hardest AI Overviews questions, so November tracking data spans a model change on 12 November.
-
Google now lists the referrer values its search traffic sends
Google's help page names an android-app:// Referer for the Android app and explains why Search Console clicks will not match analytics sessions on the same site.
-
Google undercounted AI impressions as the core update began
Google confirmed a logging error undercounted impressions in Search Console's Generative AI report on November 9 to 11, the days the core update began rolling out.
-
Only a quarter of ChatGPT's cited sites were on Bing page one
A 38-keyword test found 60.5% of the sites ChatGPT cited were on neither search engine's first page, partly because ChatGPT rewrites the keyword first.
-
Claiming your site in ChatGPT now needs a government ID
OpenAI's Business Hub asks for a DNS TXT record, a Persona ID check, legal name and tax ID, then reports mentions and website clicks without improving placement.
-
Position 4 to 3 is worth 19 clicks per 1,000, 8 to 7 is worth one
iPullRank's 166-brand Search Console dataset measures single-position climbs in clicks per 1,000 impressions and tracks unbranded position-1 CTR from 7% to 5%.
-
Google's free audit tells AI agents what to fix in your repo
Modern Web Guidance 1.0 adds a static audit of HTML, CSS, JS and package.json. Google's tests still show agents failing on loading priority and security headers.
-
Google's core update puts your comparison week on Black Friday
Google's November 2026 core update started on 10 November and may run two weeks, so the recommended post-rollout comparison week falls on Black Friday.
-
A CDN switch speeds up Googlebot while the cache is still empty
Google's Search Central blog explains why crawl capacity rises after a CDN cutover, and recommends pre-warming the cache, verified-bot allow rules, and 429s.
-
Google adds 'things to do' queries to its EEA search units
The 9 November changelog entry puts tour and attraction queries in the aggregator and supplier unit docs, where the requirement is a feed or plain crawlable pages.
-
Google counts AI Overview clicks inside your total Search clicks
Ziff Davis said organic Google referrals fell about 37% year over year and set 2.7 billion AI impressions against every Search click, a ratio that is not a CTR.
-
AI Overviews cut clicks on three of five sites in France
Glenn Gabe's France-filtered Search Console data shows CTR falling from 2% to 1.1% on one site, a second YMYL drop weeks later, and two sites with no loss.
-
Sites left 43% of agent questions unanswered in one vendor's logs
LightSite AI logged what AI agents ask its client sites through WebMCP tools. 43% got nothing back, but the figure pools blocked requests with missing answers.
-
OpenAI pays Cloudflare for hints on which pages changed
Matthew Prince told Cloudflare's Q3 call the crawl-efficiency deal is already booking revenue, with no word on whether the pilot's websites get a share of it.
-
Four SERP APIs return real URLs again, Bright Data still does not
A 7 November re-measurement of 50 German and US queries puts Bright Data at 86.8% unresolved results while DataForSEO, SerpApi, Serper and Oxylabs sit near zero.
-
OpenAI is building impression and click charts into ChatGPT
OpenAI's Business Hub counts daily impressions and clicks for verified businesses but no prompts, users or rankings, and Shopify is the one named verification path.
-
Facebook's AI-written answer pages rank in Google and AI Mode
Meta has more than 6 million facebook.com/fb-answers/ URLs indexed, built from public Facebook posts, and Search Engine Roundtable reports no response from Google.
-
Brave sells your pages to AI and robots.txt can't block it
Brave's crawler advertises no distinct user agent and borrows Googlebot's permissions, so noindex is the only exit from the index it licenses to AI companies.
-
Search Console click spikes from 31 October are bot traffic
Marie Haynes traced the early-November click spikes to automated clicks on page 2 to 6 results, and a Position over 10 filter isolates the affected URLs.
-
Google says Crawled - currently not indexed may need no action
The rewritten Page Indexing report entry names crawler blocks as its one example, says nothing about quality, and sets a site-wide bar for filing an indexing bug.