Pular para o conteúdo
← Back to Skalablog

Published article

Google Search Office Hours August 2025 Recap — Part 3

Software Engineering

Google Search Office Hours August 2025 answered nine site-owner questions on hashtag search, subdomain favicons, ranking drops, knowledge panels, indexing, robots.txt, Googlebot crawl size, structured data, and mobile-first indexing. Google employees Daniel Waisberg and company ran the session on August 7, 2025, and every answer below links to Google's own documentation.

Part 3 of a series. Start with 7 answers from google search office hours 2025, then Google Search Office Hours: June 2025 Recap.

What Did Google Answer in Its August 2025 Office Hours?

Google Search Office Hours August 2025 answered nine site-owner questions and announced two product updates. The session ran on August 7, 2025, hosted by Anna Ogawa of Google, and continues a series Google has run for Japanese site owners since 2013, rebranded as Google Search Office Hours in May 2023. Each question below is summarized with the official answer.

The nine questions, in the order Google addressed them:

#QuestionGoogle's answer
1Is hashtag search still active?Results differ from before; search keeps updating
2Why do subdomain favicons fail to display?Subdirectory-level homepages are unsupported; follow favicon guidelines
3Why did a big-keyword ranking drop?Past rankings are never guaranteed to persist
4Why won't the knowledge panel update after a rebrand?Verified as updated; check Google Business Profile and all name sources
5Pages crawled but not indexedNo resubmission needed; see the page indexing report
6Do encoded commas in robots.txt disallow?Comma is not a reserved character; avoid confusing punctuation in URLs
7Googlebot download size dropped sharplyNot a problem if indexing and results are unaffected
8Should markup be added for AI Overviews?Markup helps understanding but does not guarantee AI Overview inclusion
9Must desktop and mobile URLs both be indexed?Yes; desktop users still search and would see no results otherwise

Two announcements preceded the Q&A. First, Google released a new version of Search Console Insights, the report inside Search Console that shows site owners how content performs. Google says the new report gives a more consistent experience and value closer to what other Search Console tools provide, so content creators and site owners can understand performance without data expertise.

Second, Google announced an alpha release of the Google Trends API, a new programmatic way to access search trends data. The announcement was made at a Bangkok event shortly before the session. Access is limited to a small group of testers, and Google points site owners to its blog post for the data scope and application steps. Google notes that Trends previously offered website access only, and growing user demand drove the API.

Both announcements are early-stage: the Insights report is new in its current form, and the Trends API is explicitly an alpha limited to testers, so treat either as subject to change before relying on it in workflows.

Is Hashtag Search Still Working, and Why Do Subdomain Favicons Break?

Hashtag search results have changed and Google would not confirm the feature's current state. A questioner said hashtag-driven traffic to their owned media stopped appearing recently even though the feature launched the previous year. Google's team confirmed only that searching with a # prefix now produces different results than before, because Search constantly updates to serve useful results. No end-of-feature announcement exists, and Google offered no commitment either way.

Subdomain favicon failures usually trace to unsupported URL structure. The questioner's subdomain homepage sat in a specific directory rather than at the subdomain root. Google said that looks like a 'subdirectory-level homepage', which the favicon guidelines do not support, so missing displays follow the specification rather than a bug. Google recommends building favicons exactly as specified in its favicon guidance for search results.

Why Did a Big Keyword Ranking Drop, and Why Won't the Knowledge Panel Update?

A top-page ranking around position 10 that slid out of the top 100 is not a technical fault. Google checked the submitted sample page and found it indexed and appearing in results. The stated reality: holding top positions for broad, competitive queries is hard, and Google cannot guarantee that a site ranked highly in the past keeps that position, because sites and user expectations change constantly. A site's size or regional footprint does not preserve a rank.

The knowledge panel question involved a brand rename in April 2024. Title links showed the new name, but the panel's detail and review searches still used the old brand. When Google checked, the profile appeared already updated, so the issue seemed resolved. Google's advice for panel issues: follow Google Business Profile guidelines, expect propagation delays, and thoroughly check every system that publishes the name, including Business Profile and Wikipedia. Panel information can also be edited by verified owners. Google added one explicit warning: using the disavow tool against low-quality sites would have no effect on this problem.

What Should You Do About 'Crawled - Currently Not Indexed' Pages?

'Crawled - currently not indexed' means Google fetched the page but chose not to index it, and resubmitting will not change that. Google pointed to the page indexing report documentation, which states the page may be indexed later or may never be, and that there is no need to request recrawling.

Two practical points followed. Adding such URLs to a sitemap accomplishes nothing, because Google already knows them, having crawled them. And when many pages in a sitemap show 'Crawled - currently not indexed' with no 'Referring sitemap' detected in URL inspection, Google reframed the problem: the real question is why the pages are not indexed, not why the sitemap field is blank, so investigate page quality and value rather than sitemap mechanics.

How Does robots.txt Treat Encoded Commas in URLs?

An encoded comma is not automatically disallowed, because the comma is not a reserved character in robots.txt path elements. Google's team cited RFC guidance: only spaces and hash marks are reserved and require encoding in robots.txt paths, so a Disallow rule written for literal commas does not automatically cover percent-encoded variants.

Google still recommends avoiding confusing characters altogether. Commas and similar punctuation can confuse CMS platforms, crawlers, robots.txt parsers, and users, even when technically valid. Shared URLs with such characters also sometimes fail to link correctly on social media and chat apps, a risk Google says is not worth taking. Its blanket recommendation: keep punctuation and symbols out of URLs entirely. The governing rules live in Google's robots.txt documentation.

Why Did Googlebot Download Size Drop, and Should You Mark Up Content for AI Overviews?

A sharp drop in Googlebot download size with stable crawl request counts is not itself a problem. One site reported the drop starting around May 8, 2025, with no internal changes, though its crawl statistics showed the decline only for purpose-based updates, not discovery. Google's team said that if site updates in results and indexing are unaffected, there is no issue. Crawlers are designed for efficiency, so fewer downloads can still cover the important pages, and a changed crawl type with a similar page count needs no worry. Google advised checking raw server logs rather than judging from aggregated dashboard values alone.

On structured data for AI Overviews, Google's answer was measured. Markup generally helps Google understand content better, so implementing reasonable markup is a good idea. But pages that fail the eligibility requirements for structured data features simply will not use them, and Google noted that markup describing wholesale-market products that ordinary consumers cannot buy looks unnatural for shopping results. Structured data supports understanding; it does not purchase AI Overview placement.

Do You Still Need to Index Desktop Pages Under Mobile-First Indexing?

Yes, index both desktop and mobile URLs when they differ. For setups like example.com paired with example.com/sp, Google answered generally: many users still browse from desktop devices, and if the desktop version is not indexed, those users cannot find the site in results, cutting traffic. Desktop pages can also carry more information that fits larger screens, which can improve how search engines understand the site.

Google recommends responsive design as the configuration least prone to setup mistakes, and points to its official guidance on mobile sites and mobile-first indexing within the Search Central documentation.

Frequently Asked Questions

  • When did Google run this office hours session? The session took place on August 7, 2025, and Google announced the next one for August 28, 2025. Google has held these sessions for Japanese site owners since 2013.
  • Do you need to resubmit a sitemap for crawled but unindexed pages? No. Google's page indexing documentation states that recrawl requests are unnecessary and that sitemap inclusion is meaningless for URLs Google has already crawled.
  • Does structured data guarantee inclusion in AI Overviews? No. Google said markup helps Google understand content, but pages that fail the eligibility requirements for a structured data feature will not use it, and AI Overview coverage was not promised.
  • Is a sudden drop in Googlebot download size a penalty? Google's team said no, provided the site's updates and indexing behave normally. They recommended verifying from server logs instead of relying on aggregated crawl statistics.
  • Does the disavow tool fix knowledge panel brand issues? Google explicitly said disavowing low-quality sites would have no effect on a knowledge panel showing an old brand name. Panel data comes from sources such as Google Business Profile and Wikipedia, not from link disavowal.

Source video