What should I do with all these 404 pages?

[linklander]

I have a website that Im currently working on that has been fairly dormant for a while and has just been given a face lift and brought back to life. I have some questions below about dealing with 404 pages.

In Google WMT/search console there are reports of thousands of 404 pages going back some years. It says there are over 5k in total but I am only able to download 1k or so from WMT it seems.

I ran a crawl test with Moz and the report it sent back only had a few hundred 404s in, why is that?

Im not sure what to do with all the 404 pages also, I know that both Google and Moz recommend a mixture of leaving some as 404s and redirect others and Id like to know what the community here suggests.

The 404s are a mix of the following:

Blog posts and articles that have disappeared (some of these have good back-links too)

Urls that look like they used to belong to users (the site used to have a forum) which where deleted when the forum was removed, some of them look like they were removed for spam reasons too eg /user/buy-cheap-meds-online and others like that

Other urls like this /node/4455 (or some other random number)

Im thinking I should permanently redirect the blog posts to the homepage or the blog but Im not sure what to do about all the others? Surely having so many 404s like this is hurting my crawl rate?

[Igor.Go]

Remove all internal links leading to 404 pages. If you're using a redirect, your internal links shouldn't link to 404+302->new page either, link straight to the new source.

[Linda-Vassily]

The posts and articles with good backlinks, does that content still make sense in your renewed site? If so, I'd bring them back. If you don't have the content, you can try the Wayback Machine. The same goes for any old posts you think would be useful to your new readers.

The problem with redirecting a bunch of 404s to the same page (like the homepage) is that you end up with soft 404s and not a very good user experience. Pick the ones that correspond to specific pages that you have on the updated site and redirect those to the equivalent page.

Anything else, I'd let 404. A bunch of old posts, with no good links, the content of which you no longer have a use for on the site don't represent value to searchers—those pages will just drop out of Googles index (and crawl attempts) over time.

[This isn't just theoretical. We changed domains back in November and we had lots of old content—going back 10+ years, which is ancient history for a financial publisher. I ended up with about 6,000 404s. We are now down to about 4,000 404s as pages drop off. Google crawls us quickly and regularly and our organic traffic is up 86.49% .]

[linklander]

thanks, I have planned to do that, there's so many of them though

[linklander]

OK will try that thanks