Does anyone have any tips for SEO in WebSphere Portal with Lotus WCM?

[jcmoreno]

Hi I'm working on performing SEO on a portal that's implemented on WebSphere Portal with Lotus WCM.

Now WebSphere Portal comes with some limitations such as non-pretty urls among other things. Portal also generates non-normalized urls, so you have to turn on a process for detecting and normalizing urls for the Googlebot, etc.

Does anyone have any experience with SEO in this platform, and could offers me a few tips for this specific platform?

Thank you

[john4math]

Let me preface this by saying I don't have any experience with WebSphere Portal with Lotus WCM.  If you could describe more about what the platform outputs, or know of a site using it, I could take a look.

Non-pretty URLs aren't great, but they're not the worst either.  If you're stuck with 'em, you're stuck with 'em.  According to the SEOMOZ 2011 rankings report, having keywords in the URLs are still pretty important (69.9/100).

One way to work around the non-canonicalization of the URLs is including rel=canonical tags to all of your pages.  That way, when Google comes across a non-normalized URL, you'll have the canonical tag to tell it the right page to pass the Pagerank to.  Provided you always know what the normalized URL is supposed to be for non-normalized pages, this makes it easy to always have the Pagerank going to the right page.  Even if the normalized version rel=canonicals to itself, Matt Cutts gives that the aok (video here).  If you have a procedure that's working to normalize URLs, that's best, since we never know what's going to happen next.

More sources about rel=canonical: Google Webmaster Tools help and Webmaster Central Blog.

[jcmoreno]

Hi and thank you for your help. The site in question is the following: http://www.universalpr.com/

In this platform the URL holds session information such as the language, navigational state, among other things. Now this platform has a normalization process that detects the user-agent and looks for bots such as the Googlebot. Once portal detects a crawler bot it strips out most of the session information out of the URL. The end result is still not pretty, but it aims for consistency otherwise you could get a huge number of URLs that reference the same content.

When using WebSphere Portal with Lotus WCM (the CMS) content is displayed as modules or sections of the page that are called portlets; such portlets don't have direct access to the HEAD. This makes using canonical URLs is a bit challenge. However we're working on finding a way to write to the HEAD for updating the TITLE and adding canonical elements to the pages.

The following are two other examples that use this platform:

• http://www.ncaa.org/ - This site uses WCM stand alone and it redirects to WebSphere Portal only whenever necessary (i.e. the login). The benefits of this approach is that they can use canonical elements and that the can update their titles depending on which content is being displayed.
• http://miamidade.gov/ - Miami Dade is a lot similar to my example; which is www.universalpr.com. In Miami Dade's case the use the Lotus WCM portlet. If you click around this site you'll see that their Title is always the same and that they don't use canonical elements in their page either.

It goes without saying that I could benefit from using the NCAA's approach, however this would require quite a bit of re-work.

These are some of the shortcomings that I can identify with my limited experience in SEO. If you need any more details at all please let me know.

Thank you,

[john4math]

The NCAA site does seem to be pretty well put together.  Like you said, their TITLE changes from page to page, which you need for decent SEO.  Also, once you figure that out, you should also be changing your META description tag within the HEAD on each page on your site, so it describes that page you're on.  In the meantime, I would recommend removing your meta description altogether from all of your pages if you can't set it on a page by page basis.  That way, the search engines will try to pick out the most relevant text from your pages (and relating to the user's search query) instead of always using "Auto Insurance from Universal Insurance Group – Trusted. Fast. Service." (also you're missing some punctuation here between Group and Trusted).

Working within this platform, one thing which the NCAA site is doing well,which I think you could try to do is to get that nasty session parameter out of your URLs.  I think you only need that if you're tracking the user's session?  The NCAA site doesn't include it until I click the member login link.  Since the Googlebot won't be signing in, if you can manage to get rid the these parameters when people aren't signed in, that would improve the SEO of your URLs.

The more of these random directories you can get out of the URLs, the better.  So it looks like the way it's configured right now, is the URL doesn't change, except for the WCM_GLOBAL_CONTEXT parameter at the end, which is setting the content of the page?  The NCAA site is getting the WCM_GLOBAL_CONTEXT parameter to appear at the end of the URL, like it should be, rather than as a parameter.  That is much more natural, and I would imagine Google rather see keywords in the URLs rather than in parameters in URLs.

[jcmoreno]

Thanks for your input. It's just as I though apparently we should re-think our strategy into using the WCM servlet directly instead of through Portal. When WCM is used through Portal via the WCM portlet the content item is fetched through that WCM_GLOBAL_CONTEXT parameter, but the parameter itself seems less relevant to Google.

I'll look into this. Thank you for all your help.