How do I fix apparent duplicates

[NicDale]

I'm auditing a site and would appreciate your help with possible explanations and solutions as to why Google Analytics in the Content Drilldown page is showing what appears to be duplicate pages. (Refer image)

I'm wondering if I have got my head around the rel=canonical tag because the page I'd consider a duplicate "page/" has a Canonical tag pointing to "~/page.html"

This is the tag from the page Locations/

rel="canonical" href="http://www.domain.com/Locations.html" /> so am unsure why both versions of the page are generating views.  Shouldn't the Canonical tag work like a 301 redirect?

I'm unsure how the pages using the path page/ are generating so many views because I have not been able to find them and they are not indexed by Google.

Unfortunately the site is built using a Propriety CMS I'm not familiar with.

exK4EqrU25

[NicDale]

Hi Paul

I appreciate your explanation of when to use Canonical tags.  I had previously thought they were limited to redirecting www.domain.com to domain.com.

I understand your solution to the Dupes problem and will be searching SEOMoz's resources for how to write rewrites and Search & Replace filters using RegEx in Analytics for that matter.

It's not the first time you've provided an high quality answer to a question of mine.  I very much appreciate your contribution to my growing knowledge and the SEOMoz community.

Best

Nic

[ThompsonPaul]

A canonical tag is fundamentally different from a 301-redirect, Nic. There's nothing about a canonical tag that stops a visitor from being able to visit that URL. A 301-redirect actually forwards the visitor to the target page as if the initial page doesn't even exist so there's no physical way for a visitor to land on it.

Put another way, the source page of a 301-redirected URL doesn't even exist as far as the search engines are concerned (and eventually the'll actually drop the original URL altogether).

The canonical tag serves a very specific purpose. When two pages must continue to be reachable by 2 different URLs but the page content is essentially identical (e.g. a product page sorted by size or colour), then a canonical tag suggests that the search engines should consolidate the ranking value in the primary URL. That's it.

In the case of the /contact+us.html and /contact+us/ pages - that page should only be reachable at one or the other URL. There's no reason or value to the user for the page to be reachable at the second address. The correct way to deal with this is to use a rewrite rule to 301-redirect all the page/ versions of the site's pages to the page.html (assuming that's what you've decided should be the canonical.

The only time to use canonical tags instead of redirects in a case like this is if it is technically impossible to implement the rewrites (a shared server that doesn't allow access to the .htaccess file for example). But this is sub-optimal and would still leave you with the same Analytics dupe page problem you're currently running into.

So what to do about the dupes in Analytics, given the site wasn't configured with the rewrites? You can write a custom Search and Replace filter for the site's profile that uses regex to merge both versions of each page into a single line. You'll absolutely want to do this in a new profile created just for this purpose though, keeping the original unfiltered profile for reference and historical data.

Note that this will only affect data collected from the date of creation of the new profile/filter. It's not retroactive. If you want to combine results for these pages for the existing data, you'll need to dump it to Excel and use a formula to combine the dupes.

Hope that all makes sense?

Paul