Are too many 301 redirects harmful?

[Erku]

Hi everyone.

I am working with a website that has 4.7 million valid URLs, but also 4.9 million page with redirect. Is this normal? Are these too many redirects normal?

Thank you.

[not2easy]

That depends on how and why these URLs redirect. Simply having a large number of redirects is not necessarily a bad thing. Given the information posted, either yes or no is a valid answer.

Are the redirects due to canonical or consolidation reasons? Is this due to having multiple URLs for basically the same content? This happens with many WP (or other CMS) and/or eCommerce platforms and is normal under some conditions.

It helps to understand where you are getting the information because some tools or services can count normal 301 rewrites as redirects(found in logs as duplicate requests) when non-https version is requested for https URLS for example. It helps to understand the environment or platform for a good response.

[lucy24]

In some situations, like a move to https, there can be as many redirects as there are URLs. This is as it should be. Worry less about the raw number of 301s, and more about making sure no single request garners more than one redirect.

[tangor]

Any 301 that ends in a single jump to a valid 200 is NEVER a problem. Nature of the beast and is to be expected.

[phranque]

but also 4.9 million page with redirect

where are you getting/how are you counting this number?

[tangor]

Yeah ... I started to ask that question, too ... then again dynamic sites are something I don't spend a lot of time with (too messy). If one actually DOES have 5M pages, it would take quite some time for the user to actually read them all. :)

Or is ordinary "housekeeping" not quite up to snuff?

Or the messy part of dynamic is generating a boatload of "false positives"?

To me 5ish million pages (I suspect an ecommerce) would rival Amazon in scope...

If true, GO FOR IT!

Meanwhile, 4.7M is the base and 4.9M are redirected there ... in that case there's a caution that some urls are getting redirected incorrectly. (Should be equal).

Just observing!

[SweetPotato]

Redirects from where to where?
This can't we answered without context.

[Erku]

Ok, for example; the way it has been set up in our organization (I am new here), is that when product titles are changed, it automatically changes the URL to reflect the same title. It's a huge retail organization and they make thousands of changes like this and therefore thousands of redirects like this. As a result, we have more redirects than actual pages. indexed. Is this normal? Doesn't seem normal to me. URLs shouldn't change if titles are edited.

[lucy24]

Is this normal?

“Normal” is a statistical word. It’s normal for new businesses to fail, for first novels to be rejected, for your first boy/girlfriend to turn out to be a loser, and so on.

Is it normal for a business to constantly change the names of their products? Maybe, maybe not. Last year’s purple widget is this year’s mauve widget because someone in Marketing decided it’s better. Are they really changing the names of existing products, or are they constantly discontinuing A in order to push B instead? (Like those idiotic commercials that were popular a few decades back: “Honey, where’s the {brand}?” “We’re out of {brand}.” “Wha--? But I neeeed it! We always buy it!” “We’re out of {brand}. We have New Improved {brand} instead.” Have fun imagining how this would play out in real life.)

To me it doesn’t seem as if a redirect is the right solution. When you pull up an ancient product page at {major online retailer} you don’t get redirected; you get a page at the original URL telling you this product is no longer sold, but here are some related products--some of which may well be the identical product under a new brand name. And this can all be done with approximately the same code that generated the redirect, since we’re obviously not talking about 4.7 million hand-rolled HTML pages.

[not2easy]

They need to look at those old URLs' traffic to decide whether they should redirect. At the same time, it is an error to redirect a product to a completely different product (usually but not always) depending on whether the result would be deceptive for the UX.

IF there is incoming traffic from those old product pages you should check to see if/why they are still indexed. Make sure that the pages don't exist somewhere (on a sitemap for example) that might lead a bot to 'think' they should still list it. Do those old URLs have inbound links?

[lucy24]

Come to think of it--a question that should have been clarified at the outset--who is receiving these redirects? Human visitors following bookmarks, human visitors sent by search engines ... or search-engine robots requesting old URLs (which they will continue doing forever, though the rate will drop off)?

When you say 4.9 million, do you mean that the site's configuration file contains code for 4.9 million redirects, or do you mean that 4.9 million 301 responses were issued in the past {some-time-period}?

[phranque]

but also 4.9 million page with redirect

where are you getting/how are you counting this number?

you haven't answered this question.

when product titles are changed, it automatically changes the URL to reflect the same title

are you stating that there are 4.9 million redirect rules in place as a result of these edits?

[tangor]

A widget is a widget ... Marketers (sales folks) think they are all unique when it is the same widget in 16 designer colors or forty-eleven different sizes. Can't tell you how many fights I have had over this absurdity with clients and their sales folks ... :)

HOWEVER, you do what the customer asks, even if you know better. :)

I doubt that even AMAZON has 4.9 million unique widgets ... and they are the world's largest these days!

Changing TITLES and not changing the page content does NOT deserve a redirect ... merely a reindex .. and g does that routinely. Marketers want to FORCE things instantly and by golly, the web, fast as it is, is NOT THAT FAST, and constant changing of URLS will ONLY CONFUSE G or put your site into the TCMUTM category (They Can't Make Up Their Mind!).

The product didn't change, only the title of the page did.

Time invested in getting it right BEFORE going live will have the best benefit and stronger results.

[Erku]

Thank you everyone

[browndog]

Can I jump in and ask a question?

I have 301's on an old url, so let's say:

widgets.com.au/green-widget
Redirects to: widgets.com.au/green-widget.html

But, I've since moved over to .com

So widgets.com.au/green-widget.html now redirects to widgets.com/green-widget.html

Will that cause problems? Should I go into the au domain and change all those old redirects? To stop it going from original > new > new domain?

[phranque]

widgets.com.au/green-widget should also now redirect to widgets.com/green-widget.html
all of your internal links should refer to the new widgets.com urls.

if you're changing all urls anyway, i would personally forgo the .html extension in the url at the same time, but that's a separate discussion.

[browndog]

All of the internal links have been sorted out already.

The old urls were a mix of site/category/widget or site/widget, and older ones had no html, newer ones had html.

[phranque]

The old urls were a mix of site/category/widget or site/widget, and older ones had no html, newer ones had html.

in the end, any previously used urls should be redirected to the current canonical urls in a single hop.

[lucy24]

To stop it going from original > new > new domain?

You mean, chained redirects? If it is in your power to do so, always edit the original redirect so it points straight to the current URL in a single step.

But why are you redirecting from /pagename TO /pagename.html ? As we all know, I do not care for extensionless URLs at all (“Go back in the server and put some clothes on!”) but as long as the URL already exists and has been used, why not stick with it?

True story. Some years back I separated my “real” site from my personal site. It was a clean split: the entire contents of six directories moved; everything else stayed the same. In addition to the predictable redirects, I made one that said
^(dir1|dir2|dir3)(/index.html)?$
>>
example.new/$1/
just to prevent incorrect URLs from being sent to the new site, where they would have to be redirected once more. I did the same thing recently when moving to https, except that this time I didn’t need the “index.html” component, just the directory slash. (Search engines will only request index.html if it has at some time been part of a visible URL.)

And then ... over the next five and a half years, every time I continued tweaking URLs at example.new, I went back into example.old and duplicated the rule so, again, all redirects would happen in a single step.

Postscript: Half a year ago I finally said the ### with it and replaced ALL example.old >> example.new redirects with a 410 response, reducing the size of htaccess by fully two-thirds. But that’s an option for the distant future.

[phranque]

As we all know, I do not care for extensionless URLs at all (“Go back in the server and put some clothes on!”)

everything written about url style and usability says otherwise.
from using shorter urls to not locking yourself into or even exposing the technology used are the more important reasons i can think of.
in other words, nobody cares if you are serving .html flat files, and if you aren't why bother pretending that you are?

Half a year ago I finally said the ### with it and replaced ALL example.old >> example.new redirects with a 410 response

i would only do this with the knowledge that there are no important inbound links to the legacy domain and no significant logged requests by valuable human visitors or crawlers.

[browndog]

When the site started out in 2002, it was widget.com.au/GreenWidget.htm
In 2008 I switch to Joomla, and the example url changed to widget.com.au/green-widget
At some point in the future, I think when I switched to WordPress, the url changed to widget.com.au/green-widget.html
I personally don't really have an opion on html vs no html, when I switched from .com.au to .com, the new urls had no .html and everything was bouncing, so I changed the setting in WP to add the html. It was for no other reason than that.

The urls from the old site to the new are all complete and functioning, but there's all those old urls from previous versions of the site which redirect to the new url on the old site, and then the old site to the new site. I didn't redirect for Google (because they already have the new urls), but for all of those old links to my content on other websites.

Anyway, after I asked the question, I decided I'd make the change, so am half way through editing to get rid of those 'chains' (good word), so all url versions go straight to the new domain/url.

[lucy24]

i would only do this with the knowledge that there are no important inbound links to the legacy domain and no significant logged requests by valuable human visitors or crawlers.

After five and a half years, there were definitely NO humans getting redirected--and the only links from that long ago happened to point to one of the directories that didn’t move. It just got to where it seemed pointless to redirect when the only, only, only requests were from search-engine crawlers. Stop asking, bingbot: it isn’t coming back.

and if you aren't why bother pretending that you are?

That’s exactly why ;) Let the robots think I’ve got hand-rolled html; they don’t need to suspect that certain pages are really php-or-equivalent. Besides, I did say it’s only a personal preference.

I didn't redirect for Google (because they already have the new urls)

That won’t stop them requesting those old URLs. Search engines never forget.

[blend27]

-- Search engines never forget. --

That is why i had to put several 'AbortRequest' on IIS level in web.config(drop it on LOGS Level). I mean common, the URIs do not have any inbounds for the past 12 years already, yet being hit by Goog and Bing 120 times a month including a nice CSS file that contained references to old formatting images (pixels of all diff sizes).

That Burns Solar power too.

[brighteryeg]

If you have redirect chains then definitely a problem. If you have a 301's getting you from A to B in the most efficient and relevant way possible, that's fine.

[lucy24]

the URIs do not have any inbounds for the past 12 years already

A depressing thought. The oldest redirects that I have a record of are from /index.html, instituted in September 2012, i.e. over 7 years ago. Search-engine spiders still request URLs in /index.html periodically--all of them, not just the specific one that may still have an ancient link in /index.html form.

Sigh.

[blend27]

@Lucy24, I have BingBot hitting Pre Florida URIs on monthly bases like there is no tomorrow(well, next month) on one site. Some had links from AngelFire(gee), those links were gone over a decade ago.

Have to get to that site to do a cleanup, no reason to see these requests in Logs anymore.

Anyway, this is a really interesting thread, I will open one of my own soon with bunch of questions that have not yet been asked here yet.

[lucy24]

Some had links from AngelFire

That made me laugh out loud.

Happily, most of my non-search-engine referrals these days come from an actively maintained directory which was able to globally change all its existing links from http to https. Search engines will still keep requesting the http URLs periodically, but at least I know they're not getting continuously reinforced by no-longer-valid links.