Return 410-Gone or 404-Not Found for removed pages?

Effect on a Google overall site quality score

[1script]

Hello everyone,

It's been repeatedly mentioned in these forums that overall site quality score that Google allegedly keeps for a site has some affect on rankings, including the notorious "minus 30" penalty.

Several of my sites now have been hit with this penalty, so I have an unfortunate "luxury" of looking at some of the commonalities using my own stats and observations.

Long story short - one of the common features of the sites hit by "minus 30" is the fact that they are pretty busy forums and I do delete quite a few articles to keep it free of spam. Couple months ago I implemented a script that returns 410(Gone) HTTP error when a deleted page is requested. I guess 404(Not Found) is a more common implementation. But my line of thought was that these pages are gone forever and there is no need to come back and check on them later. Yahoo Slurp, for example, is notorious for coming back for pages missing for many months and sometimes years. Googlebot exhibits similar behavior.

Anyways, I have just noticed that in your Sitemaps (W.M. Tools) account Google separates "404"s from the rest of the HTTP errors. So, with these 410s that I send out my site show ridiculous amounts of HTTP errors (in 5000 to 10000 range) and something like two 404 errors.

Do you think this many HTTP errors may lower the quality score for the site and "help" the site trip some extra filters on the way to the "minus 30" or any other rating degradation for that matter?

[tedster]

I know that Vanessa Fox confirmed months ago that 410 and 404 were handled identically -- at that time. This reports sounds like it may not be the case anymore. My own preference has always been to serve a 404. Seems to be so universal, and such a "default" approach, whereas the 410 often takes some intentional effort.

I would be astounded if too many 410 urls alone could generate a minus 30 penalty. But then again, I obviously don't know for certain.

[1script]

@ tedster:

I know that Vanessa Fox confirmed months ago that 410 and 404 were handled identically -- at that time.

It's a shame I cannot find it now - searched for 410 on their blog - because I've done the whole thing exactly because I thouht they are treated diferently. That is: 410 being premanent and 404 temporary. It would be great to see what exactly did she mean by that. They are, after all, different errors. Why treat them identically?

[tedster]

What I thought I remembered was a comment by Matt Cutts quoting her -- this was back last spring around the time of PubCon Boston. Maybe that wasn't ever in print, but I thought it was. At any rate, clearly there's something different now. And I have searched for the comment with no result.

Also, I note that we had a similar report about 410 issues:

[webmasterworld.com...]

[Shurik]

For at least a year now I have been serving a 410 error to googlebot whenever it requests a URL that I don't want to get indexed. And based of my logs that happens a lot. So far I'm in the clear.

[g1smd]

I remember Matt Cutts confirming that 404 and 410 are treated the same.

I would say that post was sometime in the range 6 to 10 months ago.