Losing pages on www but not non-www

[HRoth]

Recently I noticed a growing difference between www and non-www number of pages in the G index for my site. Previously they were identical. Lately the www has gradually lost 30% of what the non-www has. I thought it might be the canonical thing. In the past I tried but could never figure out how to get the htaccess redirect thing to work, but when I saw this happening, I managed to figure out how to do it. I also read the articles on how pages that didn't have any internal links to them could/would end up in supplemental and I figured that this was happening to my site, since I knew I had a lot of pages of content about widgets I no longer sold but kept up as helpful info and to hopefully draw in the curious. So when I saw this happening, I created links to the orphan pages off some broad informational pages I have. The drop stopped for a day or so and some pages were regained, but now it appears to be dropping again. I don't know what the problem is.

I just now checked the two groups against each other. I can't see any commonality in the pages of the non-www that are dropped in www. I thought there was a relationship with whether they are shown as having external links in Google Webmaster Tools, but no, I was wrong. So what gives?

[edited by: HRoth at 2:16 am (utc) on May 14, 2009]

[jdMorgan]

To clarify, which did you pick as your canonical hostname, www.example.com, or example.com? Which did you redirect to the other?

The hostname you redirected should disappear completely from search results, leaving only URLs referring to the canonical hostname that you redirected to. The "link juice" from the links to the non-canonical hostname will be transferred over (albeit slowly) to the canonical-hostname URLs.

Be sure you are linking only to your canonical hostname within your own site.

I would suspect that the "sluggishness" you see in getting all this cleaned up (assuming that your redirect is implemented correctly) is due to very slow spidering of your more-obscure pages. Any remaining orphan pages may not ever be spidered again, and so may never get cleaned up; more likely they'll just disappear after awhile if there are no links to them on the Web whatsoever.

Jim

[AnkitMaheshwari]

Do you also have sub-domains, as that might be the reason for more indexed pages for non-www version.

[HRoth]

I don't have any subdomains.

I redirected from example.com to www.example.com. Perhaps I should have done it the other way around, as the non has more pages. But I thought that was the way I should go. ?

Google turns up every few days. It last visited on the 12th and crawled about half the site, so my efforts to link orphans in hasn't apparently been even noticed yet. Some pages that are definitely linked in to the site and that are for fairly popular products have always been disdained by Google, and I have no idea why. It took perhaps a year to get some of them in the index. I even did things like completely redo the page(s) from scratch.

I do see that according to G Webmaster Tools, the amount of time the server took recently was three times longer than normal. Would that have any influence on this?