Pular para o conteúdo
← Back to Skalablog

Published article

5 Answers From Google Search Office Hours, Sept 2025 — Part 5

Software Engineering

Seven questions got answers at Google Search Office Hours on September 25, 2025. The most useful: sample URLs in Search Console crawl stats do not reflect real crawl distribution, and 301 redirects do not lose PageRank. The next session is planned for October 30, 2025.

Part 5 of a series. Start with 7 answers from google search office hours 2025, then 8 Answers from Google Search Office Hours, May 2025.

What Was Answered in Google Search Office Hours

Google Search Office Hours on September 25, 2025 was a 20-minute Japanese-language session where Google staff answered seven site-owner questions. Hosted by Anna Ogawa, it is part of a series Google has run for Japanese webmasters since 2013, rebranded in May 2023 as Google Search Office Hours.

The questions covered, in order: whether crawl stats sample URLs reflect real crawl distribution, what happens when invalid HTML elements appear in the head tag, whether 301 redirects during a domain move split ranking signals, why crawl requests dropped sharply after August 9, how positions are counted when AI Overviews rank first, why a long-running photo blog stopped appearing, recurring phishing warnings in Search Console, and a case where a manga page vanished only for its exact title query.

Three answers relied on site-specific checks Google performed after the submitter attached concrete site details. The recurring lesson from the session is that Google can investigate individual cases during office hours only when a URL is included in the question.

Search Central Live Tokyo Returns on November 7, 2025

Search Central Live Tokyo returns on November 7, 2025, held at the Google Tokyo office in Shibuya, its first Tokyo edition in two years. Google announced the event in a blog post and will keep popular elements from past editions while adding new content.

Three participation details matter if you plan to attend or speak:

  • Lightning talk proposals are open, with a submission deadline of October 5, 2025.
  • Registration asks which search topics you want covered, so Google can build an agenda relevant to attendees with different backgrounds.
  • Sign-up happens through the registration form in the Search Central Live announcement.

The next office hours session was scheduled for October 30, 2025, and Google invites questions in advance through its question form.

Do Crawl Stats Sample URLs Reflect Real Crawling?

No. Google said the sample URLs shown in the Search Console Crawl Stats report are not guaranteed to match the distribution of actual crawling across the whole site.

The question was concrete: if a day's sample shows 70 percent URLs under /aaa/ and 30 percent under /bbb/, does real crawling follow that ratio? The answer was that the sample may come entirely from one subdirectory or from many different ones, so no such inference is safe. A sudden shift in the sample's composition also does not prove Google changed how it crawls the site as a whole.

To measure real crawl behavior, the Google employee said, you must check your own server logs. That advice matches the documentation, which describes the report as showing crawl activity Google has recorded on your site, with the hosted sample being a subset. For log-based analysis, tools such as the open-source Screaming Frog log file analyser can complement server-side access logs.

What Happens When Invalid Elements Sit Inside the Head Tag

The browser opens the body automatically. Per the HTML standard, when a parser encounters an element not allowed in the head, such as a div, it implicitly ends the head and starts the body.

A site owner asked whether a div or iframe placed in the head invalidates everything after it, including a canonical link written below a script-injected element. Google's answer was general rather than case-specific: technically, once the parser sees an element not permitted in the head, everything below it is treated as body content, and much of the head's parsing control is lost.

The practical takeaway for SEO work is ordering. Place title, meta description, canonical, and robots meta tags before any script that could inject non-head elements, in server-rendered and JavaScript-generated HTML alike.

Do 301 Redirects During a Site Move Split Ranking Signals?

They do not. Google reaffirmed that server-side redirects such as 301 and 302 cause no PageRank loss, pointing to the official site move guide.

The submitter saw URLs from a retired domain still appearing in search results and redirecting to the current domain, even though the current site, its pages, and its sitemap never mention the old domain. Google's guidance was to read the site move documentation; the residual listings likely come from external links that Google still crawls, which is expected during a migration.

For migrations in 2025 and beyond, the documented checklist still applies: implement server-side redirects, update internal links, submit updated sitemaps, and monitor both properties in Search Console.

Why Did Crawl Requests Drop, and Where Do AI Overviews Rank?

Two separate questions got direct answers in this part of the session, both backed by documentation.

Crawl Request Drop After August 9

A site saw crawl requests fall sharply starting August 9 and found no firewall or blocklist on its side. Google explained that crawling volume is sometimes adjusted by changes on Google's systems and usually recovers, and confirmed after checking the submitted site that this case matched that pattern.

Position Counting With AI Overviews

When an AI Overviews block appears at the very top, URLs cited inside the block count as position 1. A normal blue link in first position below the block, if it does not appear in the AI Overviews, counts as position 2. The calculation rules live in the help page About impressions, position, and clicks.

Security Warnings, a Missing Blog, and a Vanished Manga Page

The last three answers dealt with individual cases, and two of them ended with Google taking action rather than publishing a general rule.

Recurring Phishing Warnings in Search Console

A site received four phishing warnings in 2025, each cleared by a review request only to reappear. Google's guidance, confirmed with the responsible team: such notices appear when Google's systems judge the site may be a phishing site. Check your login page for elements that could mislead users and for third-party-injected content, then request a review through the Security Issues report. If a warning returns after a clean review, verify the site is still clean and request review again. That is the current correct process.

The submitter had updated a photo-plus-caption blog almost daily for 13 years but noticed it dropped out of search. Checking the submitted URL showed it redirected to a domain without the blog, suggesting an unnoticed domain change, and the redirected site's top page did appear on page one for the site name query. Google suggested the owner re-verify the situation in their own environment.

A Manga Page Invisible Only for Its Exact Title

A page indexed under a manga title appeared for an abbreviated form of the title but not the exact one, and the results page showed a DMCA removal notice even though the official URL was absent from the public Lumen Database removal requests. Google reviewed the submitted details, reported the case to the responsible team, and thanked the submitter for the feedback. Search Console feedback had already been filed, and no additional remedy was offered in the session.

FAQ

  • Do crawl stats sample ratios match real Googlebot crawling? No. Google stated the samples are not guaranteed to reflect overall crawl distribution and may come from one or many subdirectories. Verify actual crawling with your own server logs.
  • Do 301 redirects lose PageRank during a domain migration? No. Google's site move documentation states that server-side redirects such as 301 and 302 cause no PageRank loss.
  • What position does a URL get when AI Overviews appear above it? URLs cited inside a top-position AI Overviews block count as position 1, and the first regular result below the block, if absent from it, counts as position 2.
  • When is the next Google Search Office Hours session? The session following the September 25, 2025 broadcast was planned for October 30, 2025, with questions accepted in advance through Google's form.
  • What should I do if a phishing warning in Search Console keeps coming back? Check the login page for misleading elements or third-party content, then submit a review request via the Security Issues report. If the warning reappears, confirm the site is clean and request review again each time.

Source video