WebSphere Commerce SEO: URL Flattening Lite

Out of the box, WebSphere Commerce has some poor SEO. This is a shame, since pretty URL’s aren’t all that difficult. WebSphere Commerce does have it’s own way to do “SEO URL’s”, but it’s pretty lacking. Basically, this will transform a URL like:
http://www.mysite.com/webapp/wcs/stores/servlet/TopCategoriesDisplay?storeId=10551&catalogId=10001&langId=-1
To:
http://www.mysite.com/webapp/wcs/stores/servlet/home_10551_10001_-1
So, it’s almost worthless. This kind of “flattening” has a number of issues:

1. It doesn’t add any additional keywords to the URL–in the above example, the word “home” is the only thing of SEO value, and it’s pretty deep in the URL.
2. It makes URL generation extremely difficult, as you can no longer use <c:url>, and have to literally build URL’s by hand.
3. You can no longer just omit certain parameters, even if a default is set in the SEOUrlMapper.xml–”/webapp/wcs/stores/servlet/home” won’t load, but “/webapp/wcs/stores/servlet/home___” might.

So what’s the answer? There are a number of things you can do, but there are two that are extremely easy to implement, yet will make a huge difference in your URL’s.

Remove “/webapp/wcs/stores/servlet” at the webserver level

This is pretty easy if you know how to use mod_rewrites–but even if you don’t, stay tuned, I’ll drop some wisdom on you.

“mod_rewrite” is an Apache module available by default in IHS, and allows some pretty advanced URL rewriting capabilities. Basically, it lets you transform URL’s on the fly, redirecting them to other URL’s, or in this case, rewriting them on the backend.

Just go ahead and drop this into the “80″ and “443″ <virtualhost> elements of your httpd.conf:

RewriteCond %{REQUEST_URI} !^/webapp.*$
RewriteCond %{REQUEST_URI} !^/wcsstore.*$
RewriteCond %{REQUEST_URI} !^/js/.*$
RewriteCond %{REQUEST_URI} !^/robots.txt$
RewriteRule /(.*) /webapp/wcs/stores/servlet/$1

So what does this do? Simply, it just means that you can cut off that ugly “webapp/wcs/stores/servlet” anytime you want.

Let me demonstrate this through a sample request:

User: “Hey, I want to take a look at ‘http://www.mysite.com/TopCategoriesDisplay?storeId=10551&catalogId=10001&langId=-1‘, can you get that for me, Mr. Webserver?”

Web server: “Sure. Well, it looks like the URI doesn’t start with “webapp“, “wcsstore” or “js/“, and isn’t “robots.txt“, so I think the user really wants ‘http://www.mysite.com/webapp/wcs/stores/servlet/TopCategoriesDisplay?storeId=10551&catalogId=10001&langId=-1‘. Hey Mr. App Server, get a copy of that will ya?”

App server: “Here’s a copy of ‘http://www.mysite.com/webapp/wcs/stores/servlet/TopCategoriesDisplay?storeId=10551&catalogId=10001&langId=-1′, just like you wanted!

Web server: “Hey user, here’s “http://www.mysite.com/TopCategoriesDisplay?storeId=10551&catalogId=10001&langId=-1‘ just like you wanted!”

So basically, the web server lies, making the user thinking he’s getting one URL, and making the app server think that another thing entirely is being requested. “So”, you’re probably thinking “I get the last line, but why the first four lines?”

RewriteCond %{REQUEST_URI} !^/webapp.*$
RewriteCond %{REQUEST_URI} !^/wcsstore.*$
RewriteCond %{REQUEST_URI} !^/js/.*$
RewriteCond %{REQUEST_URI} !^/robots.txt$

Pretty simple: we want to add “/webapp/wcs/stores/servlet/” automatically to the front of a whole bunch of URL’s, but not all of them. Specifically, we don’t want to add it to anything that:

Anything that already starts with “webapp” : “RewriteCond %{REQUEST_URI} !^/webapp.*$”

Anything that starts with “wcsstore” : “RewriteCond %{REQUEST_URI} !^/wcsstore.*$”

Anything in the “/js/” folder: “RewriteCond %{REQUEST_URI} !^/js/.*$”

The “robots.txt” file : “RewriteCond %{REQUEST_URI} !^/robots.txt$”

You’ll probably have to add some more rules, but you can figure them out.

“Hold up a minute!” you’re probably thinking “But now I have two URL’s that load up the same page! Isn’t that terrible SEO?”. Why yes it is. Let fix it, and just add something like this:

RewriteRule /webapp/wcs/stores/servlet/(.*) /$1 [R=301, L]

Pretty simple. This just redirects anything that starts with “/webapp/wcs/stores/servlet/” to the new, flattened version, using a 301 redirect.

So, there you go, we got rid of the worst part of the WebSphere Commerce URL’s, and it only took about 10 minutes. Next time, we’ll fix up that home page, since it really shouldn’t be at “TopCategoriesDisplay”.