HTTP Status Codes and Their Meanings: Full List and SEO Impact

Short answer
What are HTTP status codes?
HTTP status codes are three digit numbers a server returns for every request. The first digit sets the class: informational, success, redirection, client error and server error. So the code tells the browser what to do next and tells a search engine whether to process the page.
HTTP status codes are the short answer a server gives your browser and a search engine when a page loads: the page exists, it moved, it is gone, or the server cannot answer right now. Those three digits decide what a visitor sees and also whether Google indexes the page. Below you will find a table by class, decision rules and the SEO impact of each group. When you test redirects, our redirect checker turns the job into one step.
Definitions live in one place. The current HTTP specification lists the classes and the name of each code in RFC 9110, and browser vendors and server software follow that text.
Search behavior forms the second half of the story. Google documents how it treats each class in its guide to HTTP status codes and network errors, so every judgment below rests on that page.
A code also works independently of what a visitor sees. You can get a friendly error page in the browser while the server returns something else entirely, and keeping that separation in mind shortens half of the technical arguments.
Because of that, this guide focuses on decision rules instead of listing every code. Knowing which code to return in which situation helps far more than memorizing the list.
Which classes do HTTP status codes fall into?
Five classes exist, and each one carries a fixed meaning, so the class alone narrows the search. Knowing the class tells you where the problem sits: in the browser, in the address, in permissions or on the server.
| Class | Meaning | Common codes | What it means for SEO |
|---|---|---|---|
| 1xx | Informational, work continues | 100, 101 | No direct impact in practice |
| 2xx | Successful response | 200, 204, 206 | Content moves to the processing step |
| 3xx | Redirection | 301, 302, 304, 307, 308 | Target gets processed; signal strength depends on the code |
| 4xx | Client side error | 401, 403, 404, 410, 429 | Content gets ignored; the URL drops over time |
| 5xx | Server side error | 500, 502, 503, 504 | Crawling slows down temporarily |
One practical shortcut helps here: a leading 4 points at the request, a leading 5 at the server. That split also tells you which team should pick the ticket up, since one works on content and URLs while the other works on hosting.
One row deserves a note. The 429 code sits in the 4xx class, yet Google treats it like a server error, so we gave it a section of its own below.
What does the 2xx class tell you?
The 2xx class says the request succeeded. Google passes that content to the next processing step, which means indexing becomes possible, although nothing guarantees it.
Code 200 is the familiar one, and also the one most people recognize. Alongside it, 204 returns no content, so a search engine finds nothing to process, while 206 carries part of a response and usually appears during large file transfers.
There is a second detail in this class. A response that actually returns content and a response with an empty body are not the same thing for a search engine, because the second leaves nothing to index.
So the practical lesson runs like this: a 200 tells you the page answered, yet an error page can also answer with 200. We cover that trap in the soft 404 section.
What separates the 3xx codes?
The 3xx class says the address points somewhere else. Google follows a limited number of redirect hops, so processing stops when a chain grows too long.
Chain length matters for people too. Every extra hop adds a little delay, and on mobile connections that delay becomes noticeable in page load time.
These codes also differ in strength. Both 301 and 308 announce a permanent move, so they pass a strong signal that the target is the real address, while 302, 303 and 307 describe a temporary state and carry a weaker signal.
Using a temporary code for a permanent move therefore loses the signal on the way. We compared the codes in detail in 302 vs 301.
What is the difference between 301 and 308?
Both codes describe a permanent move. The difference sits in the request method: 308 preserves the original method and body, while 301 carries older behavior that allows the method to change in some clients.
For page addresses either one works. In requests that carry data, such as a form submission, preserving the method matters, so 308 becomes the better choice on the application side.
Either way, a server configuration or a content management system produces the redirect. Knowing which layer creates it matters, because two layers can hold conflicting rules for the same address.
How long you keep a redirect in place is a separate question. We set out the criteria in how long 301 redirects should stay.
When are 302, 303 and 307 the right choice?
Temporary codes have their place; the problem is using them in the wrong place. For maintenance, a test or a promotional period that really ends, a temporary code sends the correct message.
- 302: the address sits elsewhere for now and will return.
- 303: after a submission, show the result at another address.
- 307: a temporary move; the request method stays the same.
- None of them fits a permanent move.
A temporary code also helps during testing. If you redirect temporarily while trying a new page, rolling the test back takes one line, and you never sent a permanent message to search engines.
The most common mistake we see on business sites is a temporary redirect that has been live for years. The site moved, the address changed, and the code still says temporary, so fixing it is usually a one line job.
How does 304 affect crawl budget?
A 304 says the content has not changed since the last crawl. The server skips the body, so it saves bandwidth and crawl resources at the same time.
On large sites the difference gets serious. Instead of downloading thousands of unchanged pages again, the server answers briefly, so the crawler spends more time on new pages.
On small sites the effect stays limited. A few hundred pages rarely strain crawling, so your energy goes further on content and internal links.
Getting this right depends on consistent date headers. We collected the details in 304 not modified, and for sites with crawl problems it is one of the cheapest wins available.
What do 4xx codes mean for SEO?
The 4xx class says something went wrong on the request side. Google ignores content at those URLs, drops previously indexed addresses over time, and gradually reduces crawling.
One detail matters here: every 4xx except 429 gets the same treatment. So the difference between 403 and 404 means something to a visitor and nothing to a search engine.
Visitor experience is another story. The same code can hide very different pages: a blank screen, a helpful page with a search box, or a list of related products. Codes stay equal while conversions do not.
Crawl rate forms the second detail. The same documentation says plainly that you should not use 401 or 403 to limit crawl rate, because the right lever sits in configuration. When that layer stays out of reach, we run the audit together as part of seo consulting.
Should you return 404 or 410?
A 404 says not found, while a 410 says permanently gone. Google handles both in a similar way, yet 410 states your intent more clearly.
Visitors see no difference between the two, because both lead to an error page. The difference shows up in the message you send search engines and in your own reports.
So the practical rule reads: if the page will not return and no equivalent exists, 410 fits. When you genuinely replace it with a similar page, a permanent redirect works better.
Check internal links before you decide as well. Closing an address that still gets links from a menu or a category page sends people into a dead end. Stores face this often, and we worked through the scenarios in out of stock product pages.
When do 401 and 403 appear?
A 401 says authentication is required, while a 403 says your identity is known and still not allowed. Two different problems, so you look for the fix in two different places.
Sessions, passwords and keys usually sit behind a 401. We collected the causes and the order of checks in 401 unauthorized error.
Firewall rules are a frequent source on the 403 side. A perfectly normal request gets blocked, and finding the rule takes a look at the hosting layer rather than the site code.
Neither code belongs in your toolkit for keeping search engines out. Index decisions live in tags and configuration instead, because access blocks do not do that job and create side effects.
Why does Google treat 429 as a server error?
The 429 code means too many requests and sits in the client class. Google, however, treats that response as an overloaded server, which counts as a server error.
So in practice a site returning 429 sees a temporary slowdown in crawling rather than an immediate drop from the index. During short bursts of load, that response stays safer than showing an error page.
You can also use the code deliberately on your side. If you rate limit during a campaign or a heavy crawl to protect the server, returning this response beats serving a broken page.
Interfaces behave similarly under rate limits. We walked through typical causes and the back off logic in rate limited interfaces.
How do 5xx codes change crawl rate?
The 5xx class says the problem lives on the server. Google temporarily slows crawling, keeps indexed URLs for a while, and eventually drops them if the errors persist.
For 500, 502 and 503 the same source notes that the slowdown stays proportional to the number of affected URLs. Once the server returns 2xx again, crawl rate recovers gradually.
During an outage, narrowing the source comes first. Does the problem sit in application code, the database, the cache layer or the gateway in front of the server? Ruling those out in order beats restarting things at random.
Causes differ across the three codes, so we wrote separate walkthroughs: 500 internal server error, 502 bad gateway and 503 service unavailable.
What is a soft 404 and why does it slip past you?
A soft 404 is a page returning 200 while showing an error in its content. Google reports an empty page, or a page displaying an error message, as a soft 404 in Search Console.
Template driven sites produce most of these. You delete a product, the template survives, and the page answers 200 with a line saying nothing was found. Technically everything looks fine, while a search engine sees thin content.
Internal search results form the second common source. Once empty result pages enter the index, a pile of thin pages builds up, so keeping them out of the index and returning the right code fixes the problem at the root.
We collected detection and repair steps in soft 404 errors. The fix usually means returning the correct code, because changing the wording alone does nothing.
How do you check HTTP status codes?
Checking codes looks technical, yet in practice a few minutes cover it. What matters is looking at the server response rather than the page in your browser.
For redirects a ready tool is fastest. Open our check redirects page, enter the address, and see how many hops the chain takes and where it ends. The same screen shows the order of the codes.
Test both the trailing slash and the plain version while you are there. On many sites one version works cleanly and the other adds an extra hop that nobody notices.
For site wide broken addresses our broken link checker does the job. When the server answers nothing at all, a quick look with is it down settles the first question.
Why do redirect chains cause problems?
Every redirect adds a hop, and every hop adds delay. Googlebot follows a limited number of hops, so the page at the end of a long chain may never get processed.
Measurement forms the second problem. Some clients drop parameters once a chain grows, so the source disappears from campaign reports and traffic looks direct.
Maintenance is the third. The longer the chain, the harder it gets to track which rule affects which address, so a new edit can quietly break an older rule.
Shortening a chain is simple: point every old address straight at the final target. We covered the method in redirect chains.
Which code should you return in each situation?
Decision rules beat memorized lists. The sequence below covers almost everything that comes up in daily work.
- Page is live and correct: return 200.
- Address changed permanently: connect it with 301 or 308.
- Something moved temporarily: use 302 or 307.
- Page is gone with no equivalent: return 410.
- Page does not exist and the case is unclear: 404 is enough.
- Planned maintenance: return 503.
- Request volume spiked: signal back off with 429.
Keep one exception in mind as well: the same address should not return different codes to different visitors. A setup answering 200 to a logged in user and 404 to a crawler ranks among the hardest problems to diagnose.
Writing each decision down makes life easier. Three months later nobody asks why a temporary code sits on a permanent move.
Which code fits a maintenance window?
During planned maintenance, 503 is the correct answer. The code describes a temporary state, so a search engine takes no rush to drop the page from the index.
Publishing the maintenance page with 200 is the wrong path. A search engine then reads the maintenance text as the page content and updates its record with that text.
You can also include a retry hint in the response when the work will take a while. Clients then know when to come back, which cuts needless repeat requests.
We described the visitor side and the page itself in maintenance mode page. Announcing the window in advance also reduces the number of support messages.
What does a migration plan look like?
Address mapping is what most teams lose during a relaunch, because nobody owns the old list. Without a list of old addresses, you end up guessing which page should point where.
So the order runs like this: export the old address list, assign a target to each one, connect them with a permanent code, then confirm that no chain remains.
Measurement belongs in the plan too. Launch without confirming that new addresses carry your tracking code and you spend days debating whether traffic fell or tracking broke.
Our redirect mapping tool speeds the matching work up. Pointing unmatched addresses at the nearest category beats sweeping them all to the homepage.
Which report in Search Console shows each code?
Search Console does not list raw codes; it groups them under status names instead. Pages with redirects, not found pages, server errors and soft 404 appear on separate rows.
Report labels change from time to time, so leaning on the logic beats memorizing screen text. When a row puzzles you, opening the official help page is the shortest route.
Reports also lag, which invites misreading. A code you fixed today does not vanish from the report immediately, so confirming with a URL level inspection answers faster.
One reading order works well: server errors first, soft 404 second, not found pages third. Server errors are both the most urgent and the most likely to distort the other rows.
Which ecommerce decisions depend on HTTP status codes?
In stores these decisions touch pages that carry revenue, so the stakes rise quickly. Promotional addresses, filtered links, sold out products and off season categories each ask for a different answer.
If you reopen a promotional page every season, keeping the address stable and changing the content beats opening a new address and discarding the old one. You also keep the signal that address accumulated.
Filtered addresses deserve their own rule. Generating a page for every filter combination burns crawl resources and creates thin content, so keeping those addresses out of the index is the healthier call for most stores.
On sold out products the decision follows stock policy. Keep the page live if the product returns; close it with a permanent answer if it does not. Checkout steps follow a different rule again: they need no index presence, yet they must work for visitors at all times.
Do HTTP status codes directly lower rankings?
A single wrong code does not lower rankings by itself. The effect is indirect, because the wrong code blocks indexing, crawling or the transfer of a signal to its target.
Using a temporary code for a permanent move illustrates that well. The page stays live at the new address, yet the accumulated signal never fully arrives, so visibility recovers more slowly than you expect.
Effects also take time to surface. Recrawling happens on its own schedule, so the consequence of a wrong code appears weeks later, and that delay sends people hunting for the cause in the wrong place.
With server errors the impact depends on duration. A short outage rarely leaves a trace, while days of errors raise the risk of pages dropping out, so monitoring matters more than repair speed.
Which mistakes show up most often?
Four patterns come up again and again in audits, and in practice they share one trait: nobody complains about them. Each one is easy to fix and stays unnoticed for a long time.
- Redirecting every missing address to the homepage.
- Leaving a temporary code on a permanent move.
- Serving a maintenance page with a 200 response.
- Blocking access to reduce crawl volume.
A fifth belongs on the list: skipping the post launch check. Teams that never confirm a new page returns 200 discover weeks later that it never reached the index.
We discussed the first item in redirecting all 404s to the homepage. In short, the visitor does not find what they wanted, and search engines rarely treat that redirect as meaningful.
How often should you audit status codes?
On small sites a monthly look suffices, so set the rhythm by how often you change the site. Sites publishing often, or carrying many products, benefit from a weekly check instead.
- Look for addresses returning server errors.
- List new not found addresses.
- Measure the length of your redirect chains.
- Check the soft 404 report for new rows.
- Confirm recently published pages return 200.
Running the audit in the same order every time also lets you compare results. Change the order and two weeks stop resembling each other, so you miss the trend.
Give the audit an owner as well. An unowned checklist disappears within two months, while a short calendar task costs almost nothing next to the price of an outage.
Should you fix these codes yourself?
On a single, simple site you can handle this alone. The decision list and the audit order above give you a workable framework for most situations.
Scale changes the picture. Once multilingual structures, filtered category URLs, product feeds, a cache layer and firewall rules meet in one place, finding which layer changes a code becomes its own expertise. Talha Aslan and our team step in at that point.
If you want to continue from here, our seo consulting services page describes the scope of a technical audit and how we work. We promise no rankings; what we promise is a setup that explains why each code answers the way it does.



