There’s been a lot of talk of late about the interaction of design and search. And with good reason. Much of what Panda was designed to look at includes design-centric elements.
Mobile SEO features design as one of it’s main components, and going forward in to the future, there’ll be more emphasis (one would suspect) on Google pushing up sites that deliver great user experience on a broad range of form factors over those which don’t.
However, if we consider design as the presentation, interface functionality and ordering of the elements of a user interface, then we’re potentially missing something.
While historically we’ve simply concerned ourselves with designing for users, and letting spiders pick up the content that we generate from the HTML on the page, we’re now entering an age where we’ll be creating and curating two sets of content: one visible to humans, and one marked up for machines.
Schema.org and the rise of the machines
(inside the tower – if you think this is complex, you don’t want to think how hard Google’s problems are.)
Schema, and microformatted data more generally, offer a very interesting glimpse of the future. By their nature, they force us to ensure that we’ve got access to data that previously we may have ignored, and to present that data to search engines (and anything else crawling the web).
For example, take just the elements that start with the letter “a” for product reviews markup – we’ve got the following elements:
That’s one small part of one schema, for one item of a page. If you’ve got a full product page to play with, it’s not hard to imagine then that you could end up with just as much content for search engines as there is for users. Which presents an interesting problem.
Business implications of a microformatted world
(C-3PO vs. Data (137/365) – data can cause all sorts of problems.)
There’s a lot of this information that site owners aren’t going to be able to, or even want to expose.
For example, who do you put down for “accountablePerson”? The site owner? An editor/CMS user? The person who wrote the review? Moreover, is that legally binding? What will a court do if you’re asked to take down a review by a manufacturer that’s considered defamatory, and they want to prosecute the person responsible under liable laws?
Equally, whilst some of this can be worked out and displayed easily, like aggregateRating, headline, itemReviewed and so on, some of it’s more tricky. inLanguage for example, or isFamilyFriendly will almost certainly require human reviewing.
So what I suspect we’ll see here is sites picking and choosing which elements to use. This could potentially lead to greater disparity in the future, with larger companies with more budget being able to throw more resource at ensuring data completeness and correctness, and if that’s then used to augment or influence search results, that’s going to further favour larger brands, at the detriment of smaller publishers.
Design considerations for microformats
(Violet the Pug – it’s hard to look this good.)
There’s another issue though. Previously we’ve not had to worry about exposing this data to search engines, because we’ve either relied on them to find it themselves, or sent it to them via XML files (such as video, image or site XML sitemap files).
However, if we’re now embedding this data in the page, we need to find ways of doing that without damaging the user experience.
Now, while we could do that by simply creating lots of hidden elements and dumping all that data in them, it’s not what you could call an elegant solution.
So how else can we do it? Should we only mark up data that we’re exposing to users? Or should we strive for completeness, ignoring that much of the data that we can show to search engines then either isn’t going to be of interest to (and thus won’t be shown to) users, or expose it using hidden elements, so as not to clutter up the UI? An interesting question, with no obvious answers.
Room for improvement
(Pinky and The Brain – Google just wants to know everything.)
With the recent introduction of the GoodRelations data formats into Schema, it seems that the Powers That Be are aware that Schema in it’s current form isn’t perfect. So hopefully, in time, we’ll see further guidance being given as to what should be included, and what’s more optional.
We’ll also see the fleshing out of categories to allow more granular guidance to be given to crawlers about how content should be defined, indexed, filtered and sorted.
In the meantime though, that’s no reason not to mark up what you can, where you can. Because one thing is for sure, even though it’s not much of an issue now if your site isn’t machine-friendly, it will be in the future!