Managing Redirects and 404's

[shaunm]

Hi All,

How do you all manage the redirects and 404 on your Website?

Is there a ratio like this many no of redirects is good? Also, should have to redirect the 404 pages to their most recent version? Leaving them there will affect negatively?

It looks like I should redirect them so that Google will not continue to report about them through the Webmaster tools?


Thanks for your help!

[goodroi]

It depends on the size of the website. Large websites will always have some 404 errors. It doesn't make sense to chase after every single 404 especially when it is caused by a weird one-time typo or one-time crawling error.

I would suggest you go after the more common 404 errors that appear month after month.

You should be careful that your redirects do not grow out of control. Having an extremly high number of redirects can cause trouble.

[lucy24]

If the 404 is due to a typo of your own, redirect to the correctly spelled form. (Truism: If you notice and fix a typo five minutes after coding it, some major search engine will have crawled during those five minutes-- even if it normally only visits the page once in three days.)

Consolidate your redirects.
If you have
firstURL >> secondURL
and then later change again to
secondURL >> thirdURL
go back and change the earlier redirect to say
firstURL >> thirdURL

[shaunm]

Thanks @goodroi & lucy24 :-)

I usually run those 404 pages through Moz tools to see the no of internal links/external links they have as well as the page authority and prioritize them based on this. Is this good?

I also see that Google suggests only to redirect to the most relevant pages. Not from the pages to the home page, saying that it will result in soft 404 errors. Is that true? What if the page that returns 404 has high page authority and high no of links pointing to it and doesn't have a relevant page that I can redirect it to?

You should be careful that your redirects do not grow out of control. Having an extremly high number of redirects can cause trouble.

This is what I'm concerned about. How can I know that I'm exceeding the limit? Say my website has 1 million pages, at what ration that I can have redirects?

(Truism: If you notice and fix a typo five minutes after coding it, some major search engine will have crawled during those five minutes-- even if it normally only visits the page once in three days.)

That's interesting find, does that mean they are on the websites all the time even though the last crawled date was one week before?!

firstURL >> thirdURL

That's I need to work on as it will eventually lead to the chained redirects right?

Lastly,

1. 404 >> New Target Or
2. Page A >> Page B

In either instances, the 404 and Page A will stay in Google's index until the next crawl. Until then, will these old pages be reported as referrals in the analytic?! Or they will never be reported? Sorry if it doesn't make any sense.

How would I know that that particular 404 page, and Page A is getting some hits in terms of organic traffic or clicked from an external website and then lead to these redirects? Is there a way to monitor the performance of redirected pages?

Thank a lot!

[lucy24]

Do you have analytics code on the 404 page itself? Also put it on the 403 page. In each case, the URL reported by analytics should come through as the originally requested page, not as the URL of the 404/403 page. (This works in piwik so I can't imagine it not working in GA. Considering how, by amazing coincidence, they each added a pape-speed feature earlier this year, I'm sure they all crib from each other.)

[goodroi]

If you have really strong links pointing to non-existing 404 page and no relevant content to redirect it to, do not redirect to the homepage. Create a new content page that is relevant and then redirect it to that new page or just name the new page the same url as the destination of those links.

[shaunm]

Thanks @Lucy24,

Do you have analytics code on the 404 page itself?

Yes, the old page have the analytic code.

In each case, the URL reported by analytics should come through as the originally requested page, not as the URL of the 404/403 page.

I am confused, could you please explain? Also what does a 403 page has to do with my post? I'm curious, isn't it the forbidden pages?

Thanks @Goodroi

just name the new page the same url as the destination of those links.

That sounds like tricky. Sure, will try that :-)

[lucy24]

Yes, the old page have the analytic code.

Not the requested page. The 404 page itself. I recommend putting analytics code on the 403 page too, because it's another way of tracking human behavior.

When a human gets a 404-- or a 403-- there are two pages involved. On one side is the originally requested URL. On the other side is the physical page that the user sees. Analytics code is called by the physical page, which may or may not correspond to a requested URL.

:: detour to piwik ::

Don't know about GA, but piwik lists pages by title or by directory. And I remembered wrong: they give the name of the page that holds the analytics code. So the part of the report covering the /boilerplate/ directory doesn't only name things like "Contact" or "Legal" but also "Forbidden" and "Missing". The URL must be hiding in there somewhere-- I know it's part of the query string sent to piwik.php-- but I can't find it.

[shaunm]

@lucy 24

Thank you so much for info :-)