Faceted navigation is awesome. Filtering items by their exact details is a great way to find what you’re looking for, and the ability to get as deep and granular as possible is a huge boon to any user. If your ecommerce website has a large variety of products, faceted navigation is the way to go.
But faceted navigation has a dark side. All those extra pages it creates can get indexed by Google, and that leads to internal competition, infinite page combinations, and very confused Google robots. There’s a lot↝ to it.
A lot of enterprise search vendors↝ (I personally am most experienced with Endeca↝) include basic SEO considerations, and some include additional “SEO add-ons” in case it’s something a webmaster is concerned about. While these can be helpful, you’re gonna need more personalization than that. And don’t trust anything that treats SEO as some extraneous “add-on,” either.
So how can we best use faceted navigation for SEO? And how can we make sure it doesn’t suck for those same users we’re trying to please?
There’s a 2014 post↝ on the Google Webmaster Blog about this that I’m going to assume you’ve read through first and understand completely. I’ll try not to duplicate too much of that information. However, it focuses more on indexation, rather than building more SEO benefit on top of that.
Hell, Just Block It!
Your first instinct, especially if you’re easily defeated, may be to simply throw your hands up in the air in surrender and say, “this is gosh dang difficult! I’ll just make sure every faceted search page has a certain URL parameter and block it entirely in robots.”
That might be a good idea if you just got this account and it’s a total disaster. But you can do better than that. These pages are likely some of your core categories, and you need those to rank. Blocking all of the filtered search content is like pouring sand on a chlorine trifluoride fire↝: you think it will help, but it’s only gonna make it worse.
So we don’t want to block all these filtered pages. We want to use this opportunity to get even more SEO benefit than we had before. How do we mine this potential?
Hell, Just Block (Some Of) It!
That’s the spirit!
SEO-friendly URLs are great to have, but for facets, you’ll have to sacrifice a little bit for the sake of practicality.
For the sake of example, let’s look at the Corporate Charm of a parallel dimension, where instead of a marketing blog, it’s a shopping website for magical business charms and talismans. (Not that I don’t do a bit of that myself.)
Pictured: SEO Strategist (casual day).
Top Level Main Category Pages
This alternate Corporate Charm has five top-level categories of products: Talismans, Idols, Gems, Machina, and Miscellaneous.
These categories are all up in the website’s main, horizontal top navigation, along with a blog link.
The categories are also the first section of filters on the side of any category page.
These pages all have nice, friendly URLs, titles, meta descriptions, and H1s:
Title: Machina | Magical Engines | Corporate Charm
Desc: Looking for the best machina? Find powerful magical engines and magic constructs – shop Corporate Charm today!
You’ll also find a couple paragraphs of good, unique copy about machina here too, above and maybe below the products. All standard stuff for a product category page so far.
And on the left, we have our facets.
Facets To Completely Block
The following will apply to most online stores, but make sure you use your own data and research to back up these assumptions for your own situation.
In most shopping areas, there aren’t a lot of searches for ratings. Ratings are important for the user, and we don’t want to take them out, but we don’t need a “Machina Rated 4 Stars And Up” page (along with the 5 star page, the 3 star, 2 star, 1 star, and 0 star pages) getting indexed. This kind of thing has internal competition written all over it.
Any page modified by a rating facet should have a different parameter included in the URL. That Machina 4 Stars And Up page will be located at corporatecharm.com/category/machina?rating=4, for example.
We will never need any rating category pages indexed. No search opportunity will ever exist regarding ratings (on this, and most, sites). So, we can block the entire rating parameter in robots right off the bat.
If you’re working with IT, it’s even better to set the page up to automatically generate a <meta name=”robots” content=”noindex, follow” /> tag in the head of any page with a “rating” parameter. This is more beneficial than using robots, as it will better prevent the page from ever showing up. Plus, the “follow” bit ensures the products linked to from the category pages still get picked up on.
There’s also an alternative approach using canonical links. It’s a little riskier, but could be more beneficial depending on your unique situation. Have IT change the canonical tag of any page with a “ratings” parameter to point at the version of the page without that parameter. So /category/machina?rating=4 would canonical to /category/machina and all would be well.
Again, there are benefits to each approach, and you may be able to implement one far easier than another. Choose wisely.
This will likely apply to more than one set of facets. For instance, price is another good example. No one is searching for “machina between $51 and $100.” Price is vital to a user, but useless to a search engine. You can probably block it.
“But what about that enormous search volume for ‘cheap machina‘? Can’t I finagle my lowest price facet to target that?”
You can, but I would recommend against it. First, labeling any actual category page with “cheap” reeks of desperation. Second, there’s a better place to do that. We’ll get to it later.
The unique nature of your products will also suggest some facets that can be blocked, but be careful. For instance, the facets for the Gems category may be individual stone types, such as garnet, onyx, and citrine. These terms are searched on their own and should not be universally blocked. However, the facets for Idols may be based on height ranges – under 5 inches, 5-10 inches, 11-23 inches, 24 inches and up. No one is searching for “idols 11-23 inches” or any other version here, so this sort of product-specific facet should be blocked.
Clicking One Facet
Let’s see what should happen when you click one facet. From the nicely-optimized Idols page, for instance, we click to sort by color. What does the Red Idols page look like?
Title: Red Idols & Crimson Idols | Corporate Charm
Desc: For blood red idols, check out Corporate Charm’s stunning collection today. From crimson to maroon, we have the one. Shop now!
H1: Red Idols
This page will also have custom, unique body copy geared towards the main search volume around “red idols”, and you can see we’re going after the “crimson idols” terms with this page too.
This page is obviously index, follow. All is well so far.
“Whoa, you want me to do custom work for all these pages? My site has 2749 categories with 23849 facets, that would take all day. Can’t these fields be filled up automatically?”
Yes and no. Yes in the sense that it should automatically populate the basic parameters. So this Red Idols page, before we filled in the custom information, would’ve probably looked something like this:
Title: Red – Idols | Corporate Charm
Desc: Shop Red Idols at Corporate Charm, for all your magical needs.
H1: Red – Idols
So this isn’t terrible, and if you have hundreds and can’t get to it right away, it’s okay. It’s still pulling the facets programmatically for these fields, which allows them to at least be unique. And, yes, they’re targeted in a vague sort of way. They’re not optimized, but it’s better than not having any Red Idols page at all.
Once this kind of thing is in place across your site, you’re in a good place, but that’s just the first step. You’ll then want to prioritize by brand importance and search volume, and go through and manually optimize these pages individually to get their full value.
Selecting Three Facets
For most websites, three facets will get too niche to be useful. (For your site, maybe it’s four, maybe it’s five, maybe it’s nine; this principle still stands, just shift the number in your head.) For the sake of this post, I’m not counting the top-level categories as facets themselves. Pay attention in case your solution does count them as facets, in which case you’ll have to adjust the number here.
Take, for instance, the Purple Granite Dragon Talisman page. Wow! That’s specific. Three facet modifiers on the Talisman category. How many purple granite dragon talismans are there? Not a lot.
Two facets are still usually useful. Purple Granite Talismans? Okay. Granite Dragon Talismans? Sure. Two facets is the level where most will have low search volume, but when all taken together, it will add up to quite a bit of long tail mass.
Pictured: the long tail.
Or, while most two-facet pages may have low search volume, there will be a couple that will inexplicably have a ton and make it worth it. You know what I mean; there always are a couple. For instance, while Blue Shale Talismans, and Red Shale Talismans, and Green Shale Talismans, etc. all have negligible search volume, the immense volume for Black Shale Talismans more than makes up for it.
A certain number of facets, though – in my experience, usually three – have no use to search engines. So have IT set up the same <meta name=”robots” content=”noindex, follow” /> tag on every category page with three facets, no matter what.
By process of elimination, the pages we’re letting enter the index are now down to top level categories, categories with one facet, and categories with two facets, and only useful facets at that. That’s pretty good, but there’s more.
Two Facets In The Same Group
One thing to look out for is when a user selects two facets underneath the same facet group. You may or may not want to block this, or parts of it, depending on your website and situation.
For instance, our alternate Corporate Charm sells pure gemstones, and never sells combinations of gems, or mixed-gem jewelry. However, given our logic so far, it’s totally possible to go to the Gems page and select the Emerald facet, and then select the Opal facet on top of that.
Opal Emerald Gems? That doesn’t make any sense.
However, in another category, it may make total sense to search for two facets in the same group. For instance, Talismans often get their power through a combination of animals. So a Monkey Dragon Talisman page would be completely appropriate.
Again, only your website/unique situation can determine whether you block all instances of same-group facets, only block some of them, or don’t need to worry about the issue at all.
Selecting Multiple Top-Level Categories
Near the beginning of this article, I mentioned that the top level categories were also on the side like the facets. I don’t count them as facets technically, but it’s possible to select more than one. What happens when you do that?
Well, you get a sort of useless page, like for instance, “Gems and Idols“. Just in case you wanted to look through both at the same time.
Depending on your enterprise search solution, this URL could be handled in a few ways. But in most situations, these pages aren’t necessary in the first place. They can also likely be noindexed.
In fact, if these pages are just as useless to the user, have IT remove them completely and don’t even waste the space.
Handling Alternate Clickpaths
From the Talisman page, User A clicks on the boar facet to get to the Boar Talisman page. Then User A clicks on the granite facet to get to the Granite Boar Talisman page. This page is indexed, located at /category/talisman?animal=boar&stone=granite.
Meanwhile, User B starts at the Talisman page but clicks on the granite facet first, going to the Granite Talisman page, then clicks on the boar facet to get to the Boar Granite Talisman page. This page is also indexed, located at /category/talisman?stone=granite&animal=boar.
Obviously, this type of thing will create a lot of duplicate content very quickly. But there’s a simple solution; have IT make a preferred order for the facets. So in User B’s case, when they clicked the boar facet, the system would insert that parameter in the URL before the stone parameter.
Should a user try to manually go to a URL with incorrect parameter ordering, it should be 301 redirected appropriately, or at least include a canonical link to the correct version. Even a 404 page is better than having infinite page combinations though.
Even a big puddle of mud is better.
So that’s that for blocking things. That should take care of every situation. Of course, your website may have some other weird issues that come from whatever niche it’s in. But hopefully this guide has given you some idea how to handle them.
Should I NoFollow Facet Links?
This is a practice I’ve seen recommended many times. If you have links to trigger facets, should you rel=nofollow the ones you don’t want indexed?
You certainly can, and it definitely doesn’t hurt. However, if you’ve done everything else I’ve said so far, you shouldn’t need to. All Google would do is follow a link to a noindexed page. I’m not much of a believer in nofollow sculpting, but if you still swear by it, or are extra paranoid, feel free!
Why Use Noindex, Follow? Why Not Nofollow?
Talking about the robots meta tag in the head here – you always want to make sure that any page featuring your products has its links followed by default. If you nofollow all these pages, you run the risk that Google has a more difficult time finding certain products on your site.
With noindex, follow, even if Google stumbles onto the page by accident, it still won’t index it, but it will pick up on all the product links.
How Should I Use An XML Sitemap?
Make an XML sitemap how you usually would; use all the canonical links. The canonical links to indexed faceted pages that you want targeting keywords should be pointing at themselves. They should be self-referential to their own parameters.
For instance, the Black Monkey Talisman page, located at /category/talisman?animal=monkey&color=black, would indeed point to that same URL in its canonical link. And that’s again the same URL that should be included in the XML sitemap.
Put them in the HTML sitemap, too, while you’re at it. They should all be there.
Aren’t Automatic Titles/H1s/Etc. Good Enough?
No. And you should be ashamed of yourself for asking.
Sure, “Brown – Dragon – Talismans | Corporate Charm” is still pretty decent for going after “brown dragon talismans”. And that’s sorta the point of this whole thing; by weeding out the duplicate and unnecessary stuff and making pages like this the only ones available to index, this page may do okay without any manual optimization. This is how blocking some parameters does most of this work for you.
But your goal should still be to optimize every page, at least eventually. Perhaps “brown dragon talismans” has a dismal search volume – fine, then, put it lower on your priority list, save it for a rainy day next quarter. But you need the ability to customize all these fields on every page, and you need to do it for every indexed page eventually.
What Facets Should I Use?
This is a bigger conversation you should be having with your client or your merchandising and broader marketing departments.
However, have facets that are appropriate for each type of product. You shouldn’t display “Gem Type” on the Talisman page, but on the Gem page, it’s vital!
Look at your internal traffic flow and see what filters are being used the most (maybe expand them) and what filters are being used the least (maybe change or remove them).
Use facets based on search volume. If there are huge volumes for terms like “circular talisman”, “hexagonal talisman”, “triangular talisman”, etc., and you don’t have a “Shapes” facet on your Talismans page, get one!
Check out your internal site search logs, too. If you find a lot of users searching for something, it probably means they can’t find it easily enough via your facets. Make the necessary changes or additions.
Clean Up With Your Blog
There are some common problems you’ll have after doing all this. Let’s take a look at some, because they can all be solved with one solution.
The solution is to use your mystical powers.
- “I blocked all combinations of three facets and more. But there’s one term – ‘black & red steel rat machina’ with a HUGE search volume that would go perfectly for this page with four facets. But I can’t just unblock all of that, and IT literally growled at me when I asked them to programmatically unblock a page this complicated. How do I get that term?”
- “Actually, a lot of our shoppers search for ‘highly rated talismans’. But I already blocked the ratings facet.”
- “People are looking for ‘cheap gems’, but you said not to label a category page as cheap, and I noindexed the whole price facet. What do I do?”
- “Going after both ‘red idols’ and ‘crimson idols’ with the same Red Idols page doesn’t work – these terms are too competitive, and we need separate pages. But splitting off crimson from red under ‘colors’ doesn’t make sense. How do we target ‘crimson’?”
These issues can all be solved with the blog. There’s no need to make your IT department work a week to programmatically unblock a single page; instead, keep that page blocked, then write a comprehensive blog post that matches the desired keyphrase, link to the (noindexed) faceted page for users, and try to even pull in products from that page to display on the blog post, if you can.
It’s the same for the rest. Write a thorough piece on your highly-rated talismans. Link and feature products.
Avoid calling your own gems “cheap” – instead, while the title may be “Cheap Gems,” the post is about why some of the competitions’ cheap gems are far weaker in magic than yours, then link to your own (noindexed) low-price faceted gems page.
Post about the sudden take-off of crimson idols and how they differ from red idols. Feature your crimson idol products here. Even link to an internal search results page for the word “crimson”.
With programmatic, logical solutions like we’ve used to block various facets, there will always be some imperfections that you just can’t address. The solution is not to continue to make the system more and more complicated. The solution is to fit your logic to your website as best you can, then make sure you have a space to catch the couple terms you just couldn’t reach.
Faceted navigation is a great thing for users, but it’s clear how damaging the potential internal competition can be for your organic presence. Instead of blocking it all entirely, which is nearly just as detrimental, a strategic plan of selective blocking can actually do additional wonders for SEO.
Remembering that the exact details may change for you based on your own website, the basic points are:
- Completely block parameters that have no use in search, like ratings or price.
- Block pages with too many facets selected.
- Block pages with facets from the same group selected, when it makes sense.
- Block multiple top-level categories from being selected at once.
- Correct duplicate content from alternate clickpaths.
- Customize the optimizations on the left-over indexed pages to fully assume your potential.
- Use a blog to cover the few exceptional pages that this logic can’t work with.
Doing all this will let you turn a massive SEO headache nightmare doomsday situation into a triumphant victorious conquest you can brag about when you ask for your next promotion.
Thanks for reading Corporate Charm. Do you have any other advice?
Enjoy this post?
Put your email below to hear about the next one!