Faceted indexing

Kevin says: “Faceted indexing has been a controversial topic in SEO, especially over the last few years in eCommerce. If large eCommerce sites index targeted facets, under the right circumstances, they can significantly increase their organic traffic and revenue from it.”

Is there not a danger that these pages could be perceived as thin content that doesn’t offer much additional value?

“That’s the crunch point. When we talk about facets, there’s a high overlap between the facet and the root or canonical URL. However, there are instances where long-tail versions of the short head keyword have lots of search volume. It’s often the case that there’s not a lot of competition for the long-tail versions. There’ll be an opportunity to go after them with facets of a category or subcategory.

There are two ways to make content on the facets relevant. One is by product selection, where you ensure the narrower categories filter to create that facet. You’ll want to ensure the product section really reflects the filters that are relevant for the facet.

The second way you can do this is through content on the facets. Some of the best eCommerce companies use this strategy and add a little bit of unique content to the faceted category page in the form of text. The aim is to make sure that Google understands that a page is relevant for a slightly different keyword, or variation of the keyword, for the canonical category.”

How do you identify the faceted pages that you want to index?

“Firstly, you’ll want to drag all of the keywords that are important to you. Secondly, you should group them by category. Thirdly, look at the long-tail and short head keywords for that specific category.

Look at the signals so that you can decide whether you want to create a faceted category and let Google index that to see who is ranking for what. If you see the same competitors ranking for short head keywords in the same category, and for long-tail keywords with the same page, that’ll be a signal to not create a faceted category version. However, if you notice that the same sites that rank for the short head keywords of a category are not ranking for the long-tail version of that keyword, that would be the signal to create a facet and index it to Google.

The meta point, or the key to being successful, is to have the technical ability to create facets and decide when Google indexes them or not. In many cases, this is a manual effort but there are ways to automate that. Companies that automate have a huge competitive advantage over others.”

Is this not something that can be completely automated or are you always going to be able to do it to a higher standard manually?

“It’s possible to automate this with the right set of tools. You’ll probably have to build a custom logic, but you can monitor short and long-tail keywords for your most important categories and who ranks for them regularly. If you can monitor this on a weekly cadence, if not daily, this type of infrastructure will tie that to the creation indexation of facets.

This is a powerful way to automate things, but it will probably take some sort of data leak or intelligence that unites these strings of data and is attached to the CMS. It should create a facet or let Google index that facet. There are companies that can help you accomplish this.”

What are the key style elements commonly associated with category pages that you should use with these faceted pages?

“You should have a selection of products related to the category/subcategory and filter. For example, if your facet is for ‘blue t-shirts’ you’ll want to make sure that the t-shirts are really blue. This is an important criterion to provide more filter and search functionality for the users.

What often happens is that users come through facets on a category page and still browse around or filter until they have the right product. Don’t just make this a landing page with no filter options; you should make sure it’s a category page where users can keep exploring to find the right product for them.”

How do you ensure that Google finds them?

“Internal linking. Ideally, you’ll have some form of internal linking for several facets. It could be something like ‘other users also bought’, or complementary products around what users are already looking at. You can also use HTML sitemaps. Ideally you’ll create a unique HTML sitemap specifically for facets. This is an old-school internal linking trick that still works very well. If you can get indexed or open facets into an HTML site quickly, that can help Google find the pages that might not have been accessible before.”

What’s the best practice for content on these pages? Is it all about having a good UX or do you need a reasonable amount of relatively unique content on these pages as well?

“Reality shows that unique text on category pages still helps Google better understand the relevance of a category. It’s still deemed necessary, but the question is: what type of content is important and how much content is important?

With facets, you won’t need as much content, but a couple of sentences could be helpful. In the case of eCommerce, you should make it very buyer guide-focused and really hone in on what additional information you can provide to help users make a better choice. For example, you could talk about the material of the product and how there are differences, providing specific context around it.

There are lots of ways to provide helpful information, and that doesn’t always have to end in a wall of text. It can sometimes just be an elegant paragraph that gets users the right amount of information to make the right choice.”

Is it completely satisfactory to have the text underneath the items you’re selling on the page or is it more effective for search engines to have the text above them?

“The verdict is out on this one. Sometimes it’s fine to have the text at the bottom of the page, but more and more users are getting used to scrolling and exploring a page. If the text is really helpful and provides extra information, it should live at the top of or on the side of the page.

The challenge is that we’re yet to see enough A/B tests of different placements of text to determine the perfect way to do it. In reality, different placements work and there are also versions where you can have a bit of text at the top and then a bit of text at the bottom. This will allow users to click a button that leads them to the bottom of the page. Be creative and mindful that we’re still waiting on the body of research to inform us on the best way to do it.”

What are your thoughts on the creation of content using AI?

“You have to have really good inputs. It’s possible, but the quality has to be high. AI content is generally considered to be of lower quality. The reality is that there are lots of companies or people who use AI to generate content, it’s just that they need to spend the necessary time to vet and edit it to make sure it’s a high standard, usable, and easy to read.

If you have these resources then there’s not necessarily a problem with automating content generation, especially when you have several inputs for every category and want to create lots of content for lots of different categories. This is totally possible, but the inputs have to be right and the outputs have to be of high quality. It all depends on where you start, and what kind of information you have available to create that content.”

As Google gets better at determining which section of a website a user should be driven to, will you still need to create very small niche faceted search pages? Could you just create big pages with different sections on them?

“Yes. There are trade-offs in terms of performance, though. This has much higher importance in eCommerce than in SaaS and other types of SEO. There are trade-offs in terms of internal linking, performance, Core Web Vitals, and user experience. It’s interesting to see how some of the big players do that. For example, Amazon indexes lots of search pages that sometimes seem to be overlapping in user intent or product.

Google can understand that there are fine differences that lead to a much better experience for users. There could be a world where very large categories rank for all sorts of long-tail queries. However, it seems like this is getting harder and harder because there are lots of players that successfully create very targeted, narrow filter pages to hit precisely what the users are looking for. It’s a bit of a trade-off, but the best in the game test one against the other and make a data-driven decision.”

Is it important to create friendly URLs for search engines, even when there are so many sub-sections that pages can be built out of?

“In the context of the content, the user experience, and things like Core Web Vitals, price, delivery, etc. URLs are less important than they were five or ten years ago. However, you still want to aim for creating a taxonomy that makes sense and is short and to the point. Working with subdirectories or scanning with subdirectories is a great approach because they translate into breadcrumbs which can be shown in rich snippets and give users a feeling of whether they’re at the right address or not.

It depends on how large the site is, what the scale mechanisms are, and how many facets you want to index. It’s all about taking an iterative approach - not indexing thousands of pages overnight, but undertaking small experiments and seeing how they go. If these are successful, you can proceed.

You can take a similar approach with site structure and friendly URLs. Avoid building a huge construct of millions of pages and start with categories then build subcategories and facets. Take a slow, iterative approach to rein in problems like URLs being too long.”

What shouldn’t SEOs be doing in 2023? What’s seductive in terms of time, but ultimately counterproductive?

“SEOs shouldn’t be going after overly competitive keywords that they have no chance of ranking for. It’s juicy to see some of these short head keywords that have lots of search volume and potential traffic but are highly competitive. The problem is that going after these types of terms and not ranking well for them can mean that you lose a lot of trust and credibility, especially in-house but also with a consultancy or agency. You could create all these expectations and high hopes with customers or stakeholders but then not deliver on the promise. You typically need lots of firepower to compete with these keywords. In my mind, what SEOs shouldn’t do is jump the gun and go after super competitive keywords.

What you should do is develop a very good understanding of the keywords you’re eligible for and where they can rank. Maximise that potential, and then build on top of that like a pyramid and iterate toward those competitive keywords, rather than going after them right away. This would be a rookie mistake that’s worth avoiding because it could cost a lot of credibility and trustworthiness, both internally and externally. It could even go as far as causing you to lose funding or lose a client.”

Kevin Indig is a Growth Advisor and former Director of SEO at Shopify and you can find him at kevin-indig.com.