WordPress /random/, Branded API Data and CDN Risks for SEO
WordPress 7.2 adds a default-on /random/ redirect, Search Console's API now splits branded queries, and Google warns that CDN migrations can accidentally block Googlebot. Most of the fixes sit outside the SEO team.
Rajesh Kumar · 22 Nov 2026 · 3 min read

- WordPress 7.2 (Dec. 8) adds a default-on /random/ redirect with no dashboard off switch, so audit custom post types before updating.
- Search Console API now offers branded query classification from Feb. 21, 2026, but keep your regex for earlier dates and for cross-checking.
- Google referrers do not separate Search, Discover and Google Images; use the Discover report instead.
- CDN migrations can trigger Googlebot blocks and origin load, so allow verified bots and pre-warm caches first.
- AI agent access and Ask Maps calls are policy decisions that sit in legal terms, bot protection and Business Profile settings.
What happened
This week's SEO Pulse from Search Engine Journal covers several changes that look unrelated but share one problem: the person who can act on each is rarely the SEO.
WordPress 7.2, scheduled for Dec. 8, adds a /random/ URL that redirects visitors to a randomly chosen published post. It is on by default, and there is no dashboard setting to turn it off. It uses a 302 redirect with an X-Robots-Tag: noindex, follow header on the redirect itself, but not on the post the visitor lands on. Turning it off takes the wp_enable_random_content_redirect filter in code.
Google also added a brandedClassification dimension to the Search Console API. It only works for eligible properties, and its data starts on Feb. 21, 2026. A webMultimodal search type now returns Google Lens, Circle to Search and image-upload traffic, and a filterLists field allows multiple countries or devices in one request.
Google published documentation on referrers too. A google.com referrer will not tell you whether the visit came from Search, Discover or Google Images. Separately, Google explained that a new CDN starts with an empty cache, so the first request for every URL hits your origin server, and blanket bot protection can make Googlebot fail CAPTCHA challenges.
Why it matters
The /random/ redirect can surface content you never meant to show. Custom post types such as client testimonials may not be in your navigation, but if they are publicly queryable and have archives, /random/ can serve them. That holds even if an SEO plugin marks them noindex. Pages and attachments are excluded, so the risk sits with custom post types and blog posts.
Branded versus non-branded reporting gets easier, with caveats. Many teams maintain brand regex lists that drift out of date. Google says some queries may be misidentified, so the API split is not perfect either. For dates before Feb. 21, 2026, you still need your own regex.
Analytics attribution stays blurry. If Discover and Search both arrive as google.com, your reports will blend them. The Discover report in Search Console is the reliable place to separate them. If Google traffic shows up as referral or direct, custom channel rules or filters may be reclassifying it.
CDN migrations are an SEO risk, not just an IT project. If bot protection blocks Googlebot, pages can drop out of search results. A cold cache can also push heavy crawl load onto your origin server.
Local and AI agent access are now policy questions. Google's Ask Maps has started calling some U.S. beauty, pet care, auto and home-repair businesses to check appointment availability. United Airlines' terms allow search indexing but require written permission for AI agents that search, book or buy. In HUMAN Security's October data, Meta's and OpenAI's agents made about 90% of agent requests, and Meta's Muse publishes no user agent for robots.txt to name.
What to do now
- WordPress owners: list your public custom post types before updating to 7.2. Ask a developer to exclude any that should not be reachable through /random/, using 'randomable' => false or the filter.
- Reporting: run the new branded split next to your regex for the same period and investigate the differences. Keep the regex for anything before Feb. 21, 2026.
- Analytics: check channel groupings, referral exclusions and filters. Use the Discover report rather than referrer data to measure Discover.
- Before a CDN move: enable verified-bot allowances in the CDN and firewall first. Pre-warm the cache by fetching your most popular and most crawled URLs before switching traffic.
- Local businesses: check the Bookings and inquiries from customers setting in Advanced settings of your verified Business Profile. Google's post does not say whether Ask Maps calls follow that setting, so brief whoever answers the phone that these calls may come from Google on a customer's behalf.
- Ecommerce and travel sites: decide your AI agent policy in your terms and bot protection, not only in robots.txt. If an agent publishes no user agent, robots.txt cannot address it.
The bigger picture
None of this needs a new strategy. It needs clear ownership. Write down who controls the WordPress code, the CDN and firewall, the analytics configuration, the Business Profile and the legal terms. When something changes, you will know who to call.
Raj's takeNone of this is dramatic, but all of it bites when nobody owns it. I've seen CDN switches and security plugins quietly cut organic traffic because the SEO wasn't in the room. Make one ownership list this week: code, CDN, analytics, Business Profile and legal terms.
Want this applied to your business?
Related

Why Real Experts Beat SEO Loopholes in the AI Search Era
Lily Ray argues that every SEO shortcut eventually expires, while investing in real, identifiable people has held up through each cycle. Her case rests on 20+ years of consistent Google guidance and the E-E-A-T framework.

Google to CMOs: Stop Generic 'Top 10' Posts, Show Real Expertise
Google's Think with Google account told CMOs that content, not AI, makes brands invisible in Search. The advice targets generic tips posts, not every listicle, and it leans on first-hand expertise and business metrics.

Google Search Console Click Spikes Are Bot Traffic, Not Demand
Some sites saw Search Console clicks jump 2x to 6.5x between Oct. 31 and Nov. 2 while impressions stayed flat. The pattern points to automated activity, not people, so don't report it as growth.
