This week SEOMOZ put up a post on crawling and indexing issues with Faceted Search. They also made reference to an interview with Matt Cutts on Faceted Search and SEO.
Normally I’d join the conversation on there and pop in a comment or two, but in this instance I want to save myself the flames as I know there’s a high likelyhood lots of people will disagree.
First up, when the topic of facted search comes up, Matt firstly tries to suggest it’s a bad thing that users don’t “get”. Personally I love faceted search – and I think most users do too. I think what Matt really means here is that GoogleBot has some trouble with it. For all it’s cleverness, GoogleBot simply cannot mimick human behaviour that well, so a site using a navigation system like this presents it with a problem. How can it tell which content is good, if it doesn’t understand how to get to it?
For instance, if you and I were to arrive on a site with a good facted system (I’ve always liked the way dabs.com implemented theirs) and assuming we’re going to try and find a new laptop, we’d click the laptop category link, then use the facteted search to drill right down to get a list of the products that have the features that are important to us. GoogleBot’s problem is that it doesn’t arrive on a site with a view to finding something specific – it arrives at a site with a view to finding as much as it can.
When Matt is told by the interviewer about a site that has great faceted search that users love, he’s got nowhere to go. He absolutely can’t speak against anything that users say they like because (and quite right too) Google’s mantra has always been headed up by “build sites for users, not for search engines”. So he simply says “absolutely” in response to the statement “It’s actually a good user experience…resulted in much better revenue per visitor which is a good signal”.
The trouble is, the interviewer has also mentioned that this website with facteted search, has seen it’s rankings drop alongside it’s conversions increasing.
My problem here (if it’s not clear already) is that this idea of building sites for visitors and not for search engines sounds like it’s breaking down. Google is now such a huge source of traffic in percentage terms for almost every website it’s becoming the victim of it’s own success. If you make a site and only think about users, you’re unlikely to have any users because no-one will find you. If you build it for a search engine everyone will find you, but they’ll also leave just as quickly.
I was looking into facted search for a site, and have decided to abandon it having read what I’ve read this weekend. It’s going to be simple category structures going no further than 2 levels simply because I have to find a balance between what will work for a user, and what will work to get traffic – a website is worthless unless you have users and search engines interested.
@Andy
There is a way to make faceted navigation/search play nice with SEO. It’s just tricky. The real issue is that bots don’t have intent, and if the bot doesn’t detect the trap, well, it’s over.
Thing is, humans don’t really like drill-fatigue, either. This is what I’ve been working on for awhile now. See for the missive if you’re interested.
Saying humans don’t like faceted search is silly. It must be that big box stores pay Endeca just to be nice. Likewise, saying robots don’t need to see the navigation is also silly. There are many ‘gray areas’ where some sort of classification could be done with the category tree or broken off as a facet. That just about proves it. If you hide facets, then that particular nugget of spider food just went poof.
There are approaches to fixing it, and a few are doing so, but few are paying attention.