Best way to handle page filters and sorts

[jcgoodrich]

Hello Mozzers, I have a question that has to do with the best way to handle filters and sorts with Googlebot.

I have a page that returns a list of widgets. I have a "root" page about widgets and then filter and sort functionality that shows basically the same content but adds parameters to the URL. For example, if you filter the page of 10 widgets by color, the page returns 3 red widgets on the top, and 7 non-red widgets on the bottom. If you sort by size, the page shows the same 10 widgets sorted by size. We use traditional php url parameters to pass filters and sorts, so obviously google views this as a separate URL.

Right now we really don't do anything special in Google, but I have noticed in the SERPs sometimes if I search for "Widgets" my "Widgets" and "Widgets - Blue" both rank close to each other, which tells me Google basically (rightly) thinks these are all just pages about Widgets. Ideally though I'd just want to rank for my "Widgets" root page.

What is the best way to structure this setup for googlebot? I think it's maybe one or many of the following, but I'd love any advice:

• put rel canonical tag on all of the pages with parameters and point to "root"
• use the google parameter tool and have it not crawl any urls with my parameters
• put meta no robots on the parameter pages

Thanks!

[OlegKorneitchouk]

I'd go with the parameter option:

1. Go to Webmaster tools > Crawl > URL Parameters > Configure URL Parameters and enter all of the sorting/filtering parameters there.

2A) If all of your items are on one page, you can set up a canonical URL for that page (which would ignore all sorting parameters)

2B) If your categories have multiple pages, be sure to use rel=next/prev for pagination

[ThompsonPaul]

I'd definitely go with option 1 - to canonicalise all the parameter variations to the root page. This is a textbook example of what the canonical meta-tag is designed for.

In addition, because you say that many of the variations are also ranking, this will pass that ranking to the root page, instead of throwing it away as would happen if you used the GWT to ignore the parameters.

Lastly, the canonical will be understood by most engines and only needs implementing once. If you go the GWT route, you'll also have to do it manually in Bing Webmaster Tools as well, and then you'll have to remember to update both each time new parameters are implemented. And this still won't work for secondary search engines, assuming they have any importance to your site.

I always think of the Webmaster Tools solution as the method of last resort if for some technical reason I am unable to implement correct canonicalisation/redirects. Consistency and lack of manual intervention are paramount for me in these situations.

Hope that helps?

Paul

[Everett]

The only thing I might add is that, depending on the business, it might be worth building a "Red Widgets" category (as an example). However, you would treat this like a sub-category and write its own category description. You would give it its own rel canonical tag, treating it as the root of the "Red Widgets" category root.

Nine times out of ten it isn't necessary to give sorting and filtering options their own category page though, and a rel canonical tag to the canonical version of that page is the second best option. The first best option would be to not change the URL at all, only re-order the items, hiding some and featuring others. Most eCommerce platforms don't have this functionality at present, however. Rel Canonical was made to span the gap until they do.