404 errors: what they are, why they matter and how to handle them

You clicked a link, and instead of the page you were after, a screen appeared saying it does not exist. It is a 404 error, and you have met it hundreds of times. For someone with a website, though, it is not just other people's annoyance: it is a signal to understand, measure and manage.

In this guide I explain what it is, why it happens, how much it really matters for SEO and how to build a good 404 page. And, if you want to see how I did mine, my 404 page is waiting for you: there is a little game too.

What a 404 error is

When the browser asks for a page, the server answers with a status code, a number that says how it went. 404 means "not found": the server works, but that address matches no content.

Other codes worth knowing:

• 200: all good, the page exists.
• 301: the page has moved permanently, here is the new address.
• 302: temporary move.
• 404: not found.
• 410: permanently removed, and it will not come back.
• 500: server error.

Why a 404 happens

The most common causes are banal.

• A badly written link, with a typo, on your site or another.
• A deleted or renamed page with no redirect left.
• A change of addresses after a site rebuild, without a redirect map.
• Old external links: other sites, newsletters or profiles pointing to addresses that no longer exist.
• Removed products or articles, still linked or indexed.
• Content moved between sections or languages.

Do 404 errors hurt SEO?

The honest answer is: it depends on which pages are missing.

• Having a few 404s is normal. Every site has pages that do not exist, and Google itself explains that it is an ordinary situation that does not penalise the site as a whole.
• Real damage comes when important pages disappear: they brought visits, had links from other sites, were ranking. If they now give 404, that value is lost.
• Broken internal links waste search engines' work and confuse people navigating.
• Soft 404s (error pages that answer "200") create confusion in indexing.

For people visiting the site, the damage is immediate: someone who lands on a nonexistent page often leaves.

How to find 404 pages

• Search Console: in the pages report it lists the addresses Google tried to visit and found not to exist, and where they come from.
• A site crawling tool: it follows all the links and flags broken ones, internal and external.
• Server logs: they show which addresses are requested and do not exist, including those requested by other sites.
• Analytics: visits to the error page, and where they come from.

How to fix them

Each 404 has to be assessed; not all should be "solved" in the same way.

1. If the page has moved: a 301 redirect to the new address.
2. If the content exists in a similar form: a redirect to the most relevant page (not to the home page at random).
3. If it has been deleted and there is no replacement: leaving a 404 is fine, or better a 410 if it is permanent.
4. If the link is wrong: fix the link, if it is yours; if it belongs to another site, you can ask for it to be updated, if the site is important.
5. If it is an address that never existed (random attempts, typos): leave the 404.

A useful rule: never redirect everything to the home page. The visitor does not find what they were looking for, and Google often considers it an error.

The 404 page: an opportunity, not a dead end

There will always be pages not found. What you can decide is what whoever lands there sees. A good 404 page:

• says clearly what happened, without jargon ("this page no longer exists");
• keeps the look and menu of the site, so it does not feel like you have left;
• offers a way out: a button to go back to the home page, the main services, perhaps a search;
• really answers with code 404, not 200, otherwise it becomes a soft 404;
• is not indexed: search engines are told not to show it in results;
• has a tone in keeping with the rest of the site, even with a touch of lightness.

I chose to turn mine into a little game: instead of a mute page, there is a defence of the site against an invasion of bugs. You can try it here: my 404 page. Note: that address really returns a 404, as it should, so if you analyse it with a crawling tool it will flag it as a page not found: that is intended.

How to prevent 404s

• Before deleting or renaming a page, decide where whoever was looking for it should go.
• Keep a redirect map when you change structure or rebuild the site: I talk about it in Rebuilding your site without losing rankings.
• Check the site's links now and then, especially after big changes.
• Choose stable addresses, short and without dates or numbers that change.
• Check Search Console every month to see what breaks.

Common mistakes in handling 404s

• Redirecting everything to the home page.
• An error page that answers 200 (soft 404).
• A 404 page full of nothing, with no links or menu.
• Ignoring 404s of pages that brought traffic.
• Redirect chains (A to B to C).
• Never checking, until the day enquiries drop.

Why entrust it to me

I have been programming since 1999, with more than 300 projects delivered and about 30 agencies as an external developer. I work on technical SEO and speed: I identify the errors that lose visibility, prepare redirects, look after the site's structure and measure again before and after interventions, so you know exactly what changed. I work from Rome and remotely.

If you want to know how many pages not found your site has and which ones matter, write to me: a chat of about 20 minutes, no strings attached, then a written, free proposal. The service is in Technical SEO and site speed.