These projects are *very* time-consuming, and getting all of the details right before pushing the ‘deploy’ button is absolutely crucial.

One very important point about the 301 redirects, is that any request for any of the old URLs, both canonical and non-canonical, must pass through a single 301 redirect to arrive at the final, correct, canonical new URL for that content.

That is, canonicalisation rules for old URLs, those that are correcting non-canonical requests within the old site, also need to be modified to instead redirect to the new canonical URL on the new site. Any and all ‘redirection chains’ must be avoided where at all possible.

That is (for example) the non-www to www redirect on the old site has to be modified to redirect non-www requests on the old site directly to www on the new site. Additionally, you might need either new rules for redirecting www on the old site to www on the new site or you will modify the existing non-www to www rule on the old site to cope with both non-www and www requests on the old site. Finally, there are the essential new rules for redirecting non-www on the new site to www on the new site.

That’s a simple example. The complexity can become almost mind-boggling, and making lists of URLs and looking for patterns that can be redirected in bulk using a low number of rules is crucial. Rule ordering is also important, both in preventing conflicts and redirection chains, as well as increasing the efficiency of operation.

Great article that will hopefully help alleviate some of the fears people have with this. It will always be a daunting task but your experience is definitely one I’ve encountered, thankfully.

I don’t think you mentioned 301′s enough though…

Just one thought that might be relevant: You did this change slap bang in the middle of the Google caffeine roll-out. Aside from you being braver than me in that respect (needs must), it may have had a bearing on the improved rankings. Have you had any -ve results since?

Again, thanks for sharing.

James.

Thanks again, your blog is now bookmarked :)

I mean will google penalize our site rankings because of email spamming
My ip is 88.208.216.153 and i check on many sites like http://www.mxtoolbox.com
please help

- Testing 301s on a test server – use Xenu to crawl all existing URLs on old site, export to a page map and open up in Excel. Then do a find and replace to your test server domain and then reimport all the URLs back into Xenu – this should give you an idea if you’ve missed any 301s that should be in place – without waiting for the site to go live before you notice

- Set up a 404 notification system like http://www.linkpatch.com (free for a single domain) – by adding a bit of code to your 404 page this will ping you an email every time an actual user hits a 404 page and tell you where they came from – bloody useful at quickly picking up cracks that you missed out

- One of the issues that comes from a domain / URL migration is waiting for search engines to actually pick up the redirects – for lower priority pages this can take a while for the 301 to be picked up and updated in the SERPs. One technique I like the idea of is <a href="http://www.thegooglecache.com/white-hat-seo/xml-sitemap-assisted-redirects-advanced-white-hat-seo/"XML Sitemap Assisted Redirects but recently WMT has been rejecting XML sitemaps containing redirected URLs – but I think you could create a temporary HTML sitemap with historical URLs so that these are recrawled within the new site. Be warned that Yahoo is rubbish at picking up 301s – I’ve seen old domains showing up literally years after 301s have been put in place.

- Have patience… I’ve seen dropouts occur as a result of broad sweep changes such as these but if you’re doing this right the results will come back, and often better than before. But the worst thing you can do is panic if this temporary sandbox effect happens.

Hope those are helpful =)