Tuesday’s Search Shift tracks a ranking spike Google still has not logged, alongside a hard August 31 deadline for Bing’s legacy API and new documentation on how Google’s crawlers share one crawl budget. Two independent datasets on AI citations, one covering local search and one from Microsoft’s own analytics tool, round out what changes this week’s measurement stack. One unconfirmed sighting on Google’s captcha behavior closes out today’s Quick Hits.
A Second Spike Hits a Ledger That Still Isn’t Moving
Ten third-party trackers just flagged the second unconfirmed ranking event in two weeks, and Google’s own status dashboard has not logged a single entry for either one. The story here is not what caused the spike, it is the verified silence on Google’s side.
- A Second Ranking Spike Hits, Google’s Status Log Still Hasn’t Moved Ten third-party trackers flagged a new ranking spike from August 1 to 3, the second such event this publication has tracked in two weeks, and Google has confirmed neither one. Our research desk checked Google’s Search Status Dashboard directly: the most recent logged ranking entry is still the June 24 spam update, with nothing recorded for either the July 23 to 26 spike or this one.
Two Deadlines Technical Teams Need to Calendar
One change reshapes how a shared crawl budget gets spent, the other sets a hard cutoff date for a legacy API. Both are operational, not conceptual, and both carry a concrete task to close out this month.
- Google Says All Crawlers Share One Crawl Budget, Urges 304 Support Updated Google documentation confirms Googlebot, Googlebot-Image and every other Google crawler draw against one shared crawl capacity pool rather than separate budgets, and now recommends serving HTTP 304 responses for unchanged pages to protect that shared pool. Sites that treated image and content crawling as independent budgets were working from a false assumption.
- Bing Webmaster Tools Sets API Deadline, Promises Filtering Fix Microsoft will retire the legacy SOAP and POX Bing Webmaster Tools APIs on August 31, 2026, and any integration still calling the old endpoints stops returning data that day. Microsoft separately acknowledged a long-standing request to add page and query level filtering to Bing Webmaster Tools reports, without giving a timeline.
Two Ways to Stop Guessing at AI Visibility
One dataset traces exactly where local AI answers pull their citations from, corroborating an independent study this publication covered yesterday. The other hands site owners a free way to separate visibility that already knows the brand from visibility that has to earn it.
- Local AI Answers Cite 6 to 27 Domains, Rarely the Business Itself State of AI Search data from AthenaHQ, alongside a separate Uberall benchmark, show local AI answers cite between 6 and 27 domains depending on the engine, and a tracked business’s own website appears in only about 16 percent of those responses. The finding lands a day after this publication covered Aleyda Solis’s cross-vertical citation study, an independent dataset from a different vendor reaching a compatible conclusion for local search specifically.
- Microsoft Clarity splits branded from non-branded AI visibility Microsoft added branded and non-branded segmentation to Clarity’s free AI Citations dashboard, letting site owners see whether an AI system cites their pages only when a searcher already names the brand. The split exposes a measurement gap that blended visibility scores have hidden since AEO reporting began.
Today’s Quick Hits
- Google May Be Testing Sign-In Prompts In Place Of Captchas A single searcher reported Google showing a sign-in prompt instead of the usual captcha deep in results pages, a sighting Search Engine Roundtable could not reproduce and Google has not addressed. If the behavior ever expands beyond this one report, any rank tracker paging deep into results without an authenticated session would need to adapt.