how to structure a list filtering clean url for usability and seo?

Question

I am a big proponent of using super clean urls for all pages and lists. Generally my pagination urls are just example.com/section/page/2 and tags are example.com/tags/tagname. I generally even try to leave the row id out of the url.

But how would you guys suggest doing a filter list?

Say you have a list of cars and you want to sort by a type, color, price or a combination of those. Say you want to filter the list to get all sedans that are green.

It makes the most sense to me just to:

example.com/cars/?color=green&type=sedan&order=price

It doesn't look very nice at all... I can read it just fine though.

But..

example.com/cars/green/sedan/price

doesn't make any sense. Also it would be a mofo to try to figure out a routing scheme for this.

Also how does this work with SEO? Will google crawl after the ?. Is that good if it does or doesn't crawl the params? Would google indexing endless permutations of the same data have ill effects?

Answer 1

http://httpd.apache.org/docs/2.0/mod/core.html#acceptpathinfo

Answer 2

I've seen a couple of schemes that use named parameters in the clean URL. For example:

example.com/cars/color:green/type:sedan/order:price
example.com/cars/page:2

The code to implement it can be a little tricky, but it's easy to understand for the user.

Answer 3

Google have long started crawling dynamic pages with query strings. So a '?' in an URL is not a problem for Google.

the reason we use clean urls is not because search engines don't like it. Clean urls are meant to be for the users. So don't sweat if all you are concerned about is SEO, don't sweat it.

Answer 4

Also how does this work with SEO? Will google crawl after the ?.

Normally, it will.

Is that good if it does or doesn't crawl the params?

That's up to you to decide, isn't it? I could imagine a number of general queries that would be good to have in the index, and a great number of very detailed ones that would not. Maybe you want to prepare a few queries to be indexed and block out all the rest.

Answer 5

when you have a single list it should only be index by google once. remember "duplicate content"?

so generally i would not suggest to have the filtered list in google. (but maybe it fits to your service)

i avoid to use clean urls for filterparams and in my robots.txt i block all urls with queryparams.

Answer 6

If you object is purely SEO and indexing then do not alter your URI addresses. Instead use the canonical link tag and like to the address of your page that you wish to be indexed. The canonical link tag is supported by all, or soon will be, major search engines and prevents redundant indexing of the same page with different URI addresses.

<link rel="canonical" href="http://example.com/cars/green/sedan/price" type="text/html"/>

The type attribute is not required, but I find it is generally good practice to indicate mime types on all requested or directed media.

Answer 7

The main thing to think about is what happens when the user wants to find all sedans, but doesn't care about the colour?

Personally I would order the fields by how "high level" they are, as if they are categories. For example, put the type first, then maybe the colour. Any variables that only change the way you look at the data, like ordering, keep in the query string.

So you'd end up with these:

All sedans: example.com/cars/sedan
All green sedans: example.com/cars/sedan/green
Sedans, ordered: example.com/cars/sedan?order=price
Green sedans, ordered: example.com/cars/sedan/green?order=price