--- title: "Do 404 errors hurt SEO? What Google says in 2026" description: "404 errors do not cause penalties or drain crawl budget, Google says. When to redirect, when to leave the 404 alone, and why 500 and 429 errors matter more." date: 2026-11-03T00:00:00.000Z --- **404 errors do not hurt your site's SEO.** A 404 is the HTTP status code for "page not found", and Google treats it as a normal part of the web: when Googlebot hits a 404, it drops the URL from the index over time, with no penalty and no effect on your other pages. That comes from Gary Illyes, an analyst at Google, in a [video from The Non-Canonical Cabinet series](https://youtu.be/zJcuRhx08ek) with Cherry Prommawin, summarized by [Search Engine Roundtable](https://www.seroundtable.com/google-404-errors-42405.html) in October 2026. The video takes down three myths at once: 404s do not get you penalized, do not always need fixing and do not drain crawl budget. What actually worries Google is 500 errors and the 429 status code. ## Are 404 errors in Search Console a sign of trouble? 404 errors in Search Console are not, on their own, a sign of an SEO problem. "We don't care at all," Illyes said in the video. Pages die, products sell out, URLs get deleted, and Google expects that. The 404 list in the indexing report looks scary because it carries the word "error". But it is an inventory, not a diagnosis. The right question for each URL on the list is a different one: does anyone still land on it? I have seen a client spend a full development sprint to "clear the 404s" in Search Console while dozens of important pages on the same site were responding slowly. The effort went to the wrong place. ## Do I need to fix every 404 on my site? No. Fixing a 404 is a business and user-experience decision, not an SEO requirement. Illyes used the example of a store deleting a popular product page: if the URL still gets visits from external links or social media, letting customers land on an error page means losing sales. Back when I worked in sales, the rule was simple: a customer who walks in looking for a sold-out product will stay if someone shows them the replacement. If nobody does, they leave. A bare 404 is a store with no salesperson. So triage by value, not by count: | The 404 URL gets... | What to do | | --- | --- | | Traffic or external links, and a direct replacement exists | Redirect (301) to the replacement | | Traffic, but no replacement exists | Keep the 404 and make the error page useful | | Nothing: no visits, no links, no value | Keep the 404 or return 410 and move on | ## Can I redirect every 404 to the homepage? No. Redirecting every dead URL to the homepage is the shortcut Google advises against in the video, because it confuses visitors and muddies your data. The person clicked looking for a specific product and lands on the homepage with no explanation. In analytics, the homepage starts collecting visits that never meant to be there. A redirect only makes sense when it leads somewhere genuinely relevant. Between the homepage and a well-built 404, the well-built 404 wins. ## What is Google's framework for dead URLs? Google's framework for dead URLs has three steps, laid out by Cherry Prommawin and Gary Illyes in the video: 1. **Redirect one to one, only if a true replacement exists.** An updated version of the same article, a newer model of the same product. The destination has to be specific and highly relevant. 2. **Let it die when there is no successor.** If the page is gone for good with no natural replacement, return a 404 or a 410 (the code that says "removed on purpose"). The crawler drops the URL from the index and the site's life cycle stays healthy. 3. **Make the 404 page helpful.** A friendly message, a search box, links to your main categories or latest content. No tricks: just a map so people can find their way. Step two is the one most people skip. There is a good reason not to: fewer useless pages in the index is often good business, as the case of the [site that removed 4 million pages and gained 600,000 clicks](/en/blog/diluicao-nao-autoridade-seo-local/) shows. ## Do 404 errors drain crawl budget? 404 errors do not drain crawl budget (the number of URLs Googlebot is willing to visit on your site). Illyes called it a common myth: hitting normal 404s does not make Googlebot lose interest in the rest of the site or crawl your good pages less. This is probably the most expensive myth on the list, because it is the one that justifies whole "404 cleanup" projects on large sites. If crawling is your problem, the cause is elsewhere, such as infinite filter URLs, badly built pagination, or a [sitemap](/en/blog/como-criar-um-sitemap-melhorar-indexacao-site/) full of URLs that should not be there. ## Which errors does Google actually worry about? 500 errors and the 429 status code are what Google actually worries about, because they affect crawling across the whole site. A 500 error (server failure) tells Google the server is struggling. Googlebot responds by slowing down so it does not make things worse, and that reduces crawling of every page, not just the one that failed. A 429 ("Too Many Requests") is technically a client-side error, but Google treats it like a 500: it is the server saying "you are asking too much right now". Illyes put it plainly: for Google, it is the same thing. Here is the irony: plenty of sites block AI bots and scrapers with rate-limiting rules, and the same rule ends up returning 429 to Googlebot. If you changed your firewall or WAF because of fake bots, and many people did once it became clear that [most of the "crawler" traffic in your logs is fake](/en/blog/o-experimento-que-provou-a-maior-parte-do-crawler-no-seu-log-e-fake/), check that the real Googlebot did not get caught in the same net. ## Summary: what Google says about 404 errors and SEO - **404 errors do not cause penalties** and do not affect your other pages; Google drops the URL from the index over time. - **Fixing 404s is optional.** It is a business and UX decision, worth it when the URL still gets traffic or links. - **Do not redirect everything to the homepage.** Redirect only to a direct, relevant replacement. - **Three-step framework:** redirect when there is a replacement, keep a 404 or 410 when there is not, and build a useful 404 page. - **404s do not drain crawl budget.** 500 and 429 errors are what make Googlebot slow down. What to do now: export the 404 list from Search Console, cross it with traffic and backlink data, and keep only the URLs that still get visitors. Those get a redirect or a better error page; the rest can stay as they are. Then open the Crawl Stats report and look for 5xx and 429 responses, which is usually where the real problem is.