WordPress 7.2 /random/ Redirect SEO Guide

The WordPress 7.2 /random/ redirect impacts SEO. Check your setup now to prevent unindexed content from leaking to search engines via this default feature.

A website administrator sits at a desk reviewing code on a laptop screen while coffee sits nearby

Intro

The WordPress 7.2 /random/ redirect is not a minor feature update; it is a default-on behavior that changes how unlinked content surfaces on your site. Scheduled for December 8, this version introduces a /random/ endpoint that automatically selects a published post or eligible custom post type and sends the visitor there. The core issue for site owners and developers is simple: there is no toggle in the WordPress dashboard settings to disable it. You will not find a checkbox in the admin area to turn this feature off. If you need it disabled, you must intervene at the code level.

This default state creates an immediate risk for sites that use custom post types for content that is not intended for general public discovery. Consider a common setup: a portfolio site with a custom post type called “Testimonials” or a SaaS platform with “Case Studies” that are currently set to noindex via an SEO plugin. Under normal navigation, these items are often hidden behind specific URLs or internal links. However, if the post type is publicly queryable and has an archive page, the /random/ function considers it eligible. This means a visitor or bot hitting /random/ could be redirected to a case study that your SEO strategy specifically excludes from search indexation. The redirect itself uses a 302 status and includes an X-Robots-Tag: noindex, follow header, but that header does not attach to the destination page. So if the destination URL is technically indexable by other means, the redirect path creates a new access route that bypasses your usual containment controls.

The friction here is that the standard WordPress dashboard does not expose this setting. The developer note for this release confirms that the feature is enabled by default. To disable it, you must use the wp_enable_random_content_redirect filter and set it to false. Alternatively, if you want to keep the feature for standard blog posts but exclude specific post types, you register those custom post types with 'randomable' => false during their registration. There is a catch, though. The November developer news roundup referenced a filter spelling that does not match the actual code implementation found in the development branch. That discrepancy causes confusion for anyone trying to follow the written guide versus the live code.

If you run a site with mixed content types, this is not just a technicality. It is a potential leakage point for content you intended to keep semi-private. Before the update lands, you need to audit which custom post types meet the eligibility criteria: they must be publicly queryable and have archives. If your “Hidden Gems” or “Internal Docs” post types fit that description, they are fair game for the randomizer. The good news is that the fix is a one-line filter or a parameter in register_post_type. The bad news is that if you forget to implement it, your unlinked, noindexed content becomes reachable through a new, unexpected URL pattern. As detailed in Search Engine Journal’s recent pulse (opens in new tab), this is the primary takeaway for site operators: check your post type registrations now, or be prepared to add a filter to your theme or plugin code before the December update pushes this behavior to your live servers.

Background

Understanding the mechanics behind the WordPress 7.2 /random/ redirect requires looking at how the core handles content selection. The feature is not a blanket randomizer for all database records. It specifically targets published posts and eligible custom post types. Crucially, it excludes pages and attachments from the pool. This means your About page or portfolio images are safe from this specific redirect, but any standard post type or registered CPT that meets two criteria is in play: it must be publicly queryable, and it must have archives enabled.

When a user hits the /random/ endpoint, the system performs a lookup and issues a 302 status code to the chosen URL. This is where the technical nuance that trips up many developers lies. The redirect response itself includes an X-Robots-Tag: noindex, follow header. This tells crawlers that the intermediate redirect step should not be indexed, but it also tells them to follow the link. However, the destination page, the actual content the user lands on, does not inherit this header from the redirect. It serves with its own standard headers. If that destination page is a custom post type that you previously set to noindex via an SEO plugin or manual meta tags, the noindex directive on the destination page is what protects it from indexing. But if you forgot to add that noindex tag to the CPT in the first place, or if your SEO plugin has a bug that overrides it on certain views, the /random/ endpoint becomes a discovery mechanism for content you wanted to keep out of search results.

There is another layer of complexity introduced by the developer documentation. The WordPress developer news from November listed a specific filter name intended to control this behavior. When I looked at the actual code implementation in the development branch, the spelling provided in the official news roundup did not match the code. The filter in the code is wp_enable_random_content_redirect. If you try to use the name from the news report, your code will do nothing because WordPress will not recognize that hook. This discrepancy is not just a typo in a blog post; it is a functional difference that will break any attempt to disable the feature globally using the documented method. For site owners, this means reading the source code rather than relying on secondary reporting. The system is designed to be permissive. It assumes you want this feature on. The responsibility falls on the developer to actively suppress it for sensitive content types. If you have a testimonials CPT that is publicly queryable but you only link to it from a specific case study page, it is now technically reachable via a random URL. The /random/ endpoint does not check your internal navigation structure. It only checks if the post type is eligible. This shift in how URLs are generated changes how you need to think about content exposure in the WordPress ecosystem.

What’s happening now

While the WordPress update is still weeks away, the Search Console API has already shifted how we measure brand relevance. Google added three new Performance report features to the Search Analytics API on November 16. The most significant addition is the brandedClassification dimension. This allows teams to segment data into branded and non-branded searches directly from the API. Previously, this required maintaining complex regular expressions to match your brand name against query strings. Now, the API handles the classification natively. However, there is a hard cutoff. According to the API reference, data for this dimension only starts on February 21, 2026. Any report covering dates before that threshold will still require your own brand regex. You cannot retroactively apply this new dimension to older data. If your reporting pipeline covers a wide date range, you will need to stitch together two different methodologies. Use the new dimension for future data and your existing regex for historical records. This creates a continuity risk if the two methods classify queries differently. Google’s announcement notes that “some queries may occasionally be misidentified.” Comparing the API split with your legacy regex is the only way to spot where they diverge.

At the same time, a new Search Central page clarified how referrer values work for Google traffic. The page, logged in documentation updates on November 12, explains that Google Search uses an origin referrer policy. This means the referrer header contains only the scheme and host (like https://google.com), but it strips the specific search query string. The same applies to Google Discover and Google Images. If a user clicks from Discover, the referrer might look different depending on the surface, but it does not reveal the exact query or even clearly distinguish it from standard Search traffic without additional context. This effectively kills any hope of using referrer headers in your server logs to distinguish between Search, Discover, and Image clicks. The recommendation is to rely on the Search Console Discover report instead of trying to parse referrer values in analytics. If you see Google traffic appearing as “referral” or “direct” in your analytics, the issue is likely custom channel rules or referral exclusions reclassifying it, not a new feature in how Google sends users to your site.

These API and referrer updates matter because they change the operational baseline for tracking. You can no longer assume that a google.com referrer gives you insight into user intent. The query data is hidden by design. This makes the brandedClassification dimension more valuable, but the February 2026 start date means you are in a limbo period. For now, keep your regex pipelines running. Do not delete them. When the new API data becomes available, use that period to validate whether Google’s automatic classification aligns with your specific brand nuances. If you have hyphenated brand names or multiple regional variants, the API’s “branded” definition may not match your specific business logic. Verify the overlap manually before you automate the switch.

Source: WordPress Random Redirects, Branded API Data & CDN Advice, SEO Pulse (opens in new tab)

What it means in practice

For the /random/ redirect, the fix is code-based, not visual. Since there is no toggle in the WordPress dashboard, you cannot simply click a setting to disable it. You must intervene at the code level. If you are running a site with custom post types like testimonials, client portfolios, or internal notes that are set to noindex but remain publicly queryable, they are currently vulnerable to this redirect. A visitor hitting /random/ could land on a hidden testimonial page that is not linked anywhere in your navigation. The 302 redirect itself carries an X-Robots-Tag: noindex, follow header, but that tag does not apply to the destination page. So, Googlebot follows the redirect to your custom post type page and evaluates that page based on its own indexing status. If your SEO plugin or theme has not explicitly blocked indexing at the destination, you risk accidental indexing of content you intended to keep hidden.

To disable this feature entirely, add the following filter to your functions.php file or a site-specific plugin:

add_filter( 'wp_enable_random_content_redirect', '__return_false' );

If you want to keep the feature for standard posts but exclude specific custom post types, you should update their registration arguments. During the register_post_type call, include 'randomable' => false. This ensures those post types are skipped by the random selector. Be careful with the filter name. The November developer news mentioned a spelling that does not match the actual implementation. Trust the code in your trunk or the filter name above, not the typo in the press release.

Consider a client who runs a real estate agency. They have a custom post type for “Open Houses” that is public but also a separate post type for “Agent Bio Notes” that is internal. The “Agent Bio Notes” are set to noindex via a meta tag, but because the post type is publicly queryable and has an archive, it is eligible for the /random/ redirect. An SEO might assume that because the content is noindex, it is safe. It is not. The redirect can serve that internal note to a random user. The 302 does not protect the destination’s metadata. You must explicitly exclude the post type from the random pool.

Regarding CDN caching, the recent guidance from Google suggests that when you switch CDNs, the cache starts empty. Googlebot will hit your origin server directly for the first requests until the cache is populated. This is actually a temporary boost in crawl activity for your origin. Do not interpret high origin traffic during a migration as an attack. Instead, it is just the CDN warming up. Ensure your origin server can handle this spike. Also, if you use bot protection, verify that you are allowing verified Googlebot IPs. Blocking all bots with a blanket rule will cause Googlebot to fail challenges, and your pages will drop out of search results.

In analytics, stop trying to split Google Search and Google Discover using referrer headers. The google.com referrer hides the query string and does not identify the surface. Use the Search Console Discover report for that specific segmentation. Your server logs will only show google.com as the referrer for all Google surfaces. This is a blind spot you have to accept.

Source: WordPress Random Redirects, Branded API Data & CDN Advice, SEO Pulse (opens in new tab)

What to expect next

Based on the operational details in this release, I expect developers to immediately audit their codebases for custom post types that are publicly queryable but intended to be hidden from random access. The default behavior of WordPress 7.2 changes the security posture of these assets. Any post type that lacks the public => false flag and does not have archives enabled is now a candidate for the /random/ redirect unless explicitly excluded. I have seen too many plugins register internal data types without considering this new surface area. You should assume that any publicly accessible custom post type is now part of the random selection pool. If you are running a site with testimonials, team member profiles, or case studies, verify their registration arguments now. Specifically, look for the 'randomable' => false argument in your register_post_type calls. If it is missing, your content is exposed to random redirections on December 8. This is not a theoretical risk. It is a functional change in the core that will alter how your URLs resolve for any user or bot hitting that path.

Once the February 21, 2026 data cutoff for the new API dimension passes, I anticipate a significant shift in how agencies handle branded search reporting. The brandedClassification dimension will remove the need to maintain brittle regex lists for brand names, which is a major source of reporting error today. However, the transition period will be messy. You will have to manually cross-reference old regex-based reports with the new API outputs to establish a baseline. Expect to see more tools integrating this specific API call to automate the split between branded and non-branded queries. The value is in the accuracy, but the implementation requires patience during the data overlap period. Do not discard your current regex patterns until you have confirmed the new API data aligns with your historical figures for the eligible date range.

Finally, look for an increase in local businesses receiving contact attempts from AI agents like Ask Maps. Google is expanding its AI agent interactions, and these systems will begin proactively engaging with local entities to gather information or facilitate bookings. This is a new vector for inbound communication that does not follow traditional web traffic patterns. It is not just a bot crawling your site; it is an agent initiating a dialogue. For small business operators, this means your contact forms and chat widgets may receive more automated, conversational inquiries that are harder to filter. You should prepare your support teams for this shift in interaction type. I do not have a precise timeline for when this volume becomes significant, but the infrastructure is being laid now. The practical takeaway is to monitor your inbound channels for unusual patterns of automated conversation and ensure your human agents are trained to handle them. The line between standard web crawling and active agent engagement is blurring, and your operations need to keep up with that technical reality.

Next Steps

Update your staging environment to the latest WordPress trunk or the specific 7.2 beta build to test the /random/ behavior before December 8. Do not wait for the production release to discover that a client’s portfolio items are now accessible via a random URL. Create a simple list of all custom post types registered in your theme or plugins. For each one, check if it has archives enabled and if it is publicly queryable. If both are true, you are in scope for the redirect. If the content is sensitive, like private testimonials or internal case studies that were previously protected only by noindex headers, you need to act now. Add 'randomable' => false to the register_post_type arguments for those specific types. This is a one-line code change, but it requires touching your codebase. If you are using a page builder or a third-party plugin that registers these types, check their documentation or issue a support ticket to see if they are planning to update their registration arguments for 7.2. Relying on the SEO plugin’s noindex setting is no longer a safe boundary because the 302 redirect exposes the content to any visitor who hits /random/, and the X-Robots-Tag header on the redirect does not stop a human or a scraper from following that link.

Run a manual test on your local or staging site. Navigate to /random/ in your browser. Observe which URL you land on. Repeat this ten or twenty times. If you see any post type you did not intend to expose, your configuration is incomplete. Also, check the response headers using a tool like curl or your browser’s developer tools. Confirm that you are seeing a 302 status and the X-Robots-Tag: noindex, follow header on the redirect itself. This confirms the feature is active. If you decide to disable the feature entirely for your specific deployment, implement the wp_enable_random_content_redirect filter in your functions.php or a site-specific plugin. Note that the filter name differs from what was reported in some developer news roundups, so double-check the actual source code if you encounter issues. A spelling error in a filter name means the code is ignored, and the feature remains on.

While you are in the code, review your analytics setup. Open your Google Analytics 4 property and look at your traffic source grouping. Ensure that traffic from google.com is not being misclassified as direct traffic due to referral exclusion settings. Google’s new documentation clarifies that the referrer will only show the origin, not the specific query. If you have custom channel rules that try to separate Search from Discover based on referrer strings, those rules are now redundant or potentially incorrect. You should rely on Search Console’s Discover report for that specific split. Clean up any regex patterns in your reporting tools that assume you can distinguish between Search and Images based on the referrer header. That data is no longer available. This cleanup will save you time when you start comparing your historical branded search data against the new brandedClassification API dimension. The goal is to have a clean baseline of your brand search performance before the API data becomes the standard for reporting. This ensures you can spot any discrepancies between your old regex-based reports and the new API-driven ones without confusion.

Source: WordPress Random Redirects, Branded API Data & CDN Advice, SEO Pulse (opens in new tab)