Pagination Strategies In The Real World
One of the hot topics at SMX Advanced in Seattle this June was the best way to handle paginated sites. It seemed like the topic that would not go away, as it came up in panel after panel. The reason this happened is that it is a complicated topic.
There are two major scenarios that we will examine, including a look at the potential solutions and the choices that publishers can make.
No Place For Rel=Canonical
The “The Really Complicated Technical SEO Infrastructure Issues” panel at SMX Advanced started with controversy when REI’s Jonathan Colman said that REI.com benefitted from using rel=canonical on the product pages of its catalog.
For example, if there were 10 pages of tent products, pages 2 through 10 all implemented a canonical tag which pointed back to page 1.
This is when Google’s Maile Ohye, who was also on the panel, piped up and said that this was not a proper implementation of the canonical tag. So ideally, do not use this approach. Even though REI thinks it is working for them currently, and it might be, there is simply no assurance that it will work that way in the future.
Search engines implement their algorithms and update them from time to time based on the way they believe things are supposed to work, and any time you use a feature such as rel=canonical in a way other than intended, you face material risk of a problem at some point.
The rel=canonical tag should only be used when the target page (the one that the tag points to) has substantially all of the content on the source page (the one implementing the tag).
Scenarios where this works can occur when you have different sort orders for products, or pages that show subsets of the products on the target page.
The Scenario: Article Pagination
This scenario arises when a magazine has a very long article. One possible approach to this is simply to have a “more” button which populates the content on the existing URL if the user wants to see it (Salon Magazine uses this approach).
However, there are some publishers that choose to continue the content on one or more additional pages as you can see with this example from Discover Magazine:
In this particular example, the content for the article lives on these two different URLs:
This scenario could lead to potential SEO issues. There are two major ways to approach this situation:
1. Do Nothing
In other words, let the search engines find and index both pages. There is a strong argument for this, as the pages have different content and some searchers may only be interested in the content on the 2nd page of the article. To get the most of this strategy, consider some of the tips by Tedster on Webmaster World for optimizing paginated articles.
2. NoIndex the additional pages
You should consider this option if you believe that having visitors land on the 2nd page of the article (and other pages other than page 1) would be a bad user experience. Note that if third party publishers link to your 2nd article page, that it can still pass link juice back into the rest of your site, and that is a good thing.
The Scenario: E-Commerce Catalog Pagination
Anyone who has done shopping for popular products online has seen this scenario.
Here is an example from the Men’s Boots page on Zappos.com:
The major choices remain the same, but the arguments about them are a bit different:
1. Do Nothing
On the SEO Infrastructure panel, Maile Ohye argued that the search engines still see value on those pages because they list different products than the first page of products. The basic argument for leaving the pages alone is that they are not duplicate content and they may pick up their own search traffic.
2. NoIndex the additional pages
The counter argument to the do nothing scenario is that the unique text content on the pages is likely to be quite low. For that reason, it is not unreasonable to be concerned that the search engines will perceive these as low quality pages, and that they may therefore have a negative impact on search traffic.
Obviously, the latter would not be the goal of the search engines, but there are software algorithms involved, and these algorithms deal with a nearly limitless number of scenarios.
In spite of their best efforts, sometimes there are sites that get hurt for unintended reasons. Given that pages that simply list a bunch of products are not that likely to garner much direct search traffic it may make sense to avoid that risk.
No Place For NoFollow
If you do decide to NoIndex pages on your site, I do not advise that you NoFollow the links to those pages. NoFollow does not conserve any link juice when you implement it. It just does not pass the link juice to the page receiving the link, and that link juice is thrown away (it is not redistributed to other pages on your site).
Let the juice flow to your pages with NoIndex tags, because those pages can accumulate and pass PageRank, and they can then vote a portion of that PageRank back into the rest of your site.
Key Considerations Regarding Pagination
My inclination in the Article Pagination scenario is to simply do nothing, and let the search engines discover and index pages in the article beyond the first page.
There is little downside, and with a little page design effort, you can pretty much eliminate any concerns that users will have a poor experience landing on the 2nd page of an article.
Whether or not to place NoIndex tags on additional pages on e-commerce sites is a judgment call. Just be aware that the search engine’s preference is to discover and index that content. They want to be aware of it, and they want to handle it properly.
That said, it is not a webmaster guidelines violation to NoIndex the pages, nor is it an improper use of the NoIndex tag. In addition, there is some risk that the search engines will make an unintended mistake in understanding the value and uniqueness of the additional product pages. You need to decide which path makes you more comfortable.
Everything you need to know about SEO, delivered every Thursday.