Pagination & Canonicalization for the Pros
This fairly technical panel was pretty fascinating because I’ve always been interested in these concepts. This is something that I find so many clients forget about entirely that I wanted to hear more info from experts on the subject, so I could share it with my clients and anyone else who would listen. This panel was moderated by Vanessa Fox of Search Engine Land, and the speakers included Adam Audette of RKG, Jeff Carpenter of Petco and Maile Ohye of Google. Let’s take a look at some of the major talking points:
There’s No One Proven Pagination Method – In this panel, there were several methods given to try and achieve crawlable, scalable pagination on eCommerce sites with a lot of pages, as well as a lot of categories and series within categories. The first mentioned is the “noindex, follow” method. This means that page two onward in a series uses a robots meta tag with “noindex, follow” so that Google will follow all the pages in a series, but only index the first one on its own. Each following page after the first needs a self-references canonical tag as well as unique title and meta tags to avoid severe duplication problems that may arise. This seems to be the preferred method, since Bing will also recognize it, as they don’t recognize rel next/prev tags.
The second method that some prefer is the “rel next/prev” method, which uses “rel next” and “rel prev” tags to denote the next and previous pages in a series, respectfully. These are also used in page two onward in a series, and should also be used with a self-referencing canonical meta tag on each page. In this instance, Google will try to index this series rather than individual pages, so unique titles and metas aren’t a requirement, but they do help. The individual pages will get indexed by Google, but should only show up in specific searches, such as “site:”. This method can cause problems if the self-referring rel canonical tag and the next/prev tags aren’t set up properly, causing confusion to the engines because they’re too similar, for example, so these must be set up carefully.
The final method mentioned is “view all”, which right off the bat seems problematic. Rather than the canonical tag being self-referential, or using next/prev, the canonical tag refers to the “view all” page for that series. This method needs to be set up so that the pages load fast, or else you’re more likely to get people bouncing away. This also seems like it would prevent searching granularly if one wishes, as the results would lead to ALL of the data in a series, rather than specific data. Thankfully this method isn’t as popular as the others.
While these methods work, it’s up to the organization to figure out which method works best for them. These methods each have their benefits and issues, but if done correctly, can help both search engines and users have a better experience with a website.
Faceted Navigation Can Be Problematic – Mr. Audette went into some detail about why faceted navigation – which allows users to search by one or more filters – can be problematic for both SEO and users. For example, since filters can be applied in any order the user wishes, how should the resulting URL be rendered? Where do you point the canonical tag to? The home page? The product page? Do you build the URLs based on categories regardless of the ways in which the facets were selected?
These are big problems that would need to be solved before the filtering system was implemented. Even if you force canonical paths and stick to overhead facets for more structured URLs, these URL strings can get very long, causing crawling problems. Therefore, while users really like faceted navigation for the choice it offers them, it has to be implemented properly.
Mr. Carpenter also found in his work on the Petco home page that offering users LESS options for filters actually increased retention and conversion. So even if you stick with filtered search/faceted navigation, it sounds like simplicity is key. Overall, this searching method sounds like a big problem from an SEO perspective, even though from a UX (User Experience) perspective, they may love it. Tread cautiously if you’re considering this for your own site.
Utilize Google Webmaster Tools – Ms. Ohye of Google went into detail as to how one should use Google Webmaster Tools to help with canonicalization, especially for sites that use parameters in their URLs (i.e. most eCommerce sites). Webmaster Tools allow you to specify parameters in order to help crawl your site more efficiently.
Parameters you can specify that help with crawling include things such as those that don’t change a page’s content (i.e. session ID), parameters that do change content, parameters that specify Googlebot’s default behavior, as well as sorting behaviors. These can be used in specific instances to help Google better understand your URLs. You can also set Webmaster Tools not to crawl specific pages, such as translation pages for example. Utilizing Webmaster Tools to help Google better understand your parameters, in conjunction with proper canonicalization tags, can really help Google know how to index your pages overall.
Some of the simple best practices that all of the speakers outlined included:
• Internal links should only include canonical URLs
• List canonical URLs in sitemaps
• On-page indexing makeup is still helpful, and canonical or next/prev tags can be used in tandem
• Canonical tags need to be used as preferred URLs, not indexing shortcuts
There was a lot of information in this panel, most of which was fairly technical. I learned a great deal, and will feel better next time I come upon a client with canonicalization or pagination issues.