Migrated website: PageRank and 404's

Question

We have recently migrated our website and a lot of pages that existed on the old website no longer exist on the new website. We did create redirect rules to match the URLs from the old and the new website with a 301 but only for 50% of our content. We have around 195K pages which give a 404, and 180K pages properly redirected. We know serving a 404 to search engine bots is bad for PageRank, but are wondering whether it's bad for the overall ranking of the website. Here are our questions:

Q1: Is a 404 on a particular page affecting the PageRank of other pages and/or overall ranking of a website?

We know we can remove the 404 pages from the Google index via the webmaster tools but given the number of URLs that were and were not properly rewritten, we don't think we will be able to remove only the 404. We also know that eventually Google will remove those 404 pages of its index.

Q2: Is it better to remove everything and let Google re-index the website from scratch (and only non-404 will remain), or leave everything and have the 404 pages removed over time by themselves?

Answer 1

404s have no effect on PageRank per se, as PR is just a calculation of how many links are incoming to that page. However, 404s will have a small effect on the PageRank of the other pages on your site, as they will now have fewer links pointing to them. Since you have thousands of pages and sound like an established site, this effect will be negligible.

Google will remove those 404 pages from search results fairly quickly, but if there are external links pointing to the page, Googlebot will continue checking those pages occasionally.

IMO there are two feasible solutions:

1. Do a 301 redirect from the deleted pages to the homepage on the new site. This would maintain any "link juice" incoming to those 404 pages.
2. Keep the 404s and do what's better for users - show a message explaining that the content has gone, link your new homepage plus any content related to this page.

Answer 2

Q1: No I don't think it is, unless your own site links to pages that 404, so I would check with a scanner like Xenu's link sleuth to ensure that there are no broken links on active pages.

Q2: I would have the site return a 410 (Gone) status code for content that isn't going to return - google will understand and act accordingly. For simplicity you could replace the 404 with the 410 for a while allowing google to adjust.

It's a bad idea to remove everything - I know it's a lot of work to redirect everything but it will be worth it in the long run, google does take into account the length of time a site has been up.

@DisgruntledGoat suggests redirecting to home as a way of preserving link juice, but I would have some reservations about doing that. Firstly I'm not completely convinced 'juice transference' works that way, it has more to do with end content's relavence. Secondly this is a sure fire way to increase your bounce rate, because a lot of these queries will come in from 'the long tail' of results. If you do redirect in this way I would create a special site map landing page which specifically directs users around the site and advises them about the site redesign.

Otherwise send a 410.

Answer 3

404's are not bad for search engines please read this article directly from Google http://googlewebmastercentral.blogspot.com/2011/05/do-404s-hurt-my-site.html

You should have already re-submitted a new sitemap to Google and Bing after the change. Redirect the old top content pages. Check Analytics to see what were the most visited pages and redirect those if possible.