How to Find and Handle Soft 404s and Expired Pages Without Losing Rankings

Rows of old library card catalog drawers, some pulled open and some empty

A retail client once asked us why Google Search Console showed thousands of "Soft 404" entries when, as far as anyone on their team could tell, the site had no broken pages. They had checked. Every URL in the report loaded a perfectly normal-looking page with the header, the footer, and a friendly message saying the product was no longer available. The problem was exactly that friendliness. The server announced success with a 200 status code while the page itself said "nothing here."

Soft 404s are one of the more frustrating technical SEO problems because nothing looks broken to a human. Visitors see a polite dead end. Search engines see a contradiction, and they resolve it in ways that rarely help you. This guide covers how soft 404s happen, how to find them, and how to decide what each retired page should become.

What a Soft 404 Actually Is

A real 404 is an honest answer. The server tells the crawler that the requested URL does not exist, using the HTTP 404 status code, and the crawler can drop the URL from its index with confidence.

A soft 404 is a page that looks like an error to a person but returns a success status to a machine. The classic examples are an empty search results page, a product page with the stock replaced by "sorry, this item is gone," a category with zero items, or a custom "page not found" template that was configured to return 200 instead of 404. Google also flags pages as soft 404s on its own when the content is so thin that it reads like an error, even if nobody intended it that way.

The distinction matters because status codes drive how search engines treat a URL over time. A 200 says "keep this in the index and come back often." A page that is effectively empty but keeps asking for that treatment gets in the way of the pages that deserve it.

rows of library card catalog drawers partly open
Photo by Tima Miroshnichenko on Pexels

Why Soft 404s Quietly Cost You

The most direct cost is crawl budget. Every request a crawler spends on a dead-end page is a request it did not spend on a page that earns money. On a small brochure site that barely registers. On a store with a hundred thousand SKUs that rotate seasonally, it adds up to real delays in getting new and updated pages discovered.

The second cost is signal confusion. When Google decides a page is a soft 404, it treats it much like a hard 404 and removes it from search results, but it does so on its own schedule and with its own judgment. You lose control over what happens to the inbound links and ranking history attached to that URL. A page that should have redirected to a close replacement just evaporates instead.

The third cost is blindness. If your reports are full of soft 404 noise, it becomes much harder to notice a real regression, such as a template change that accidentally empties a whole section of the site. Clean error reporting is a monitoring tool, and soft 404s pollute it.

The Usual Sources of Soft 404s

Most soft 404s come from a small number of patterns, and once you recognize them they are easy to spot in a crawl.

Expired inventory is the biggest one. An ecommerce platform or listings site keeps the URL alive after the product, job posting, event, or property is gone, and swaps the main content for an "unavailable" notice. The template still returns 200 because the page technically rendered.

Empty listing pages are next. Filter combinations that match nothing, tag archives with no posts, and pagination beyond the last real page all produce a valid-looking shell with no content inside. Internal search result pages with zero hits behave the same way.

Misconfigured error pages round out the list. A custom 404 template served with a 200 status is surprisingly common, especially on single-page applications where the router handles unknown paths on the client and the server never has a chance to say otherwise. If your framework renders "not found" in the browser after the page has already loaded, the HTTP response was a 200 by then.

How to Find Soft 404s on Your Own Site

Start with Search Console. The page indexing report lists URLs that Google classified as soft 404s, along with a sample you can export. This is the most reliable list because it reflects Google's own judgment rather than a guess about it. Check it on a regular schedule, since new ones appear every time inventory changes.

Next, crawl the site yourself and flag any URL that returns 200 but contains very little text, or that contains phrases such as "no longer available," "no results found," or "page not found." Most crawling tools let you set custom extraction or search rules for this, and the output gives you a list you can sort by template. Sorting by template matters, because the fix usually belongs in the template, not on each URL.

Finally, compare against your sitemap. The sitemaps protocol expects the file to list URLs you want indexed. Any URL in the sitemap that returns a soft 404 is a direct contradiction in your own signals, and a quick win to clean up first.

magnifying glass resting on printed documents
Photo by Feyza Daştan on Pexels

Choosing the Right Response: 404, 410, 301, or Keep the Page

Not every retired page deserves the same treatment. A short decision process saves a lot of guesswork, and it is the core of this whole exercise.

Ask first whether the content has a true successor. If a discontinued product has a direct replacement, or an old guide was merged into a newer one, send the old URL there with a 301 redirect. A redirect passes most of the accumulated link value to the new page and sends visitors somewhere useful. The key word is "true." Redirecting everything to the homepage or a loose parent category tells Google you are hiding a missing page, and it often treats those redirects as soft 404s too.

If there is no true successor and the page is gone for good, return an honest error. A 404 says "not found" and leaves the door open. A 410 Gone says the removal is deliberate and permanent, which tends to get the URL dropped from the index a little faster. For most sites the difference is small, so either is fine as long as the status code matches the message on the page.

If the page might come back, such as a seasonal product or a recurring event, keep it live with real content. Add the expected return date, related alternatives, and a way to be notified. A page with substance is not a soft 404, and it preserves the ranking history for the day the item returns.

Building a Useful Custom Error Page

A good 404 template does two jobs at once. It returns a genuine 404 status, and it gives the visitor a way forward. Include a search box, links to your main sections, and a short note on what probably happened. Keep the header and navigation intact so the page feels like part of the site rather than a crash.

What it must not do is mimic a normal page. No product grids that pretend nothing is wrong, no recommended content that gives the crawler a reason to index the URL. The page exists to be an error, so let the status code say so, and spend your design effort on helping people recover.

If your front end is a single-page application, this needs engineering attention rather than a copy change. The server or edge layer has to know which routes are valid and return a real 404 for the rest, instead of serving the app shell for every path. We see this gap often in projects that began as pure client-side builds, and it is typically a small change in routing and server configuration once someone owns it. Our web development team treats status code handling as part of the initial architecture for exactly this reason.

Handling Expiring Content at Scale

Manual decisions do not scale when you have thousands of listings that expire every week. The answer is to encode the policy in the application so that status codes follow the data.

A workable rule set looks like this. While an item is live, serve it normally. When it expires, check for a replacement or parent in your data model, and if one exists, issue a 301 to it. If none exists, return a 404 or 410 and remove the URL from the sitemap on the next build. If the item is temporarily out of stock but expected back, keep the page live with clear messaging and structured availability data.

The sitemap step deserves emphasis because it is often forgotten. Sitemaps generated once and never rebuilt keep pointing crawlers at dead URLs for months. Generating the file from the current database on a schedule, or on every publish and unpublish event, keeps your declared intent and your actual responses in agreement.

"Search engines punish contradictions more than they punish absences. If a page is gone, say so with the status code, the sitemap, and the internal links all at once." - Dennis Traina, founder of 137Foundry

Verifying the Fix and Monitoring Afterward

After you change a template or a redirect rule, verify with real requests instead of trusting the code. Use a command line client or a crawler to confirm the status code a retired URL returns, and check the full redirect chain, since a chain of three hops can reintroduce the problem you were trying to solve. The Google Search documentation explains how Googlebot treats status codes and redirects, and it is worth rereading whenever you change your policy.

Then use the validation feature in Search Console to ask Google to recheck the affected group of URLs. Expect it to take days or a couple of weeks, depending on how often those pages are crawled. Watch the soft 404 count trend downward and confirm that the pages you redirected show up as redirected rather than lingering as errors.

Finally, make this part of routine monitoring. Add a check to your release process that fails if a known-missing URL returns 200, and review the Search Console report monthly. If the count creeps back up, a new template or a new inventory source is probably creating them again, and catching it early is far cheaper than finding it after a traffic dip.

Bringing It Together

Soft 404s are a communication problem between your server and the crawlers that read it. The fix is the same each time: make the status code, the page content, the sitemap, and the internal links all say the same thing about each URL.

If you want help auditing a large site where this has piled up, that is the kind of work our technical SEO practice does, and you can see the rest of what we build at 137Foundry.

Need help with Technical SEO?

137Foundry builds custom software, AI integrations, and automation systems for businesses that need real solutions.

Book a Free Consultation View Services