I have always been pretty dismissive of Schema. To me, it was and is an exercise in futility. The number of webmasters with the time, knowledge and inclination to enact Schema tags is a tiny fraction of the publishing audience, and thus its impact was never going to change the face of the world.
In addition, Schema itself is pitifully incomplete and actually retrograde, in that it attempts to force responsibility for telling search engines what a page is about onto site owners, rather than forcing the search engines to get better at what they’re supposed to be getting better at.
There’s also the notable potential side effect of Google slowing taking “ownership” of information away from sites. Google’s “Knowledge box” has, for some time now, slowly been taking traffic away from Wikipedia. Once Google “knows” something (the classic example being the height of the Eiffel Tower) it increasingly displays that information itself up front and centre rather than merely giving you a bunch of links to parse for yourself.
And actually, if you are sat in Mountain View that makes sense. The height of the Eiffel tower is known and is exactly the kind of thing that people could try to “spam” to get some AdSense revenue. Once you’re confident you have the right answer in your “knowledge box” database why run the risk of polluting your own reputation by sending people to potentially disreputable sources through the SERPs?
Everyone has experienced that moment when you’ve Googled something, clicked the first link and taken the answer as true, only to find later on that it was actually just rubbish. Not for nothing is Yahoo! Answers a poisoned chalice.
All of which brings us down to the subject of Schema. Schema is sold as a way for you to add a structure to your site to allow Google to get that knowledge directly, thus contributing to their knowledge graph.
As an SEO by trade, it would be remiss of me not to be dabbling with it to see what benefits and potential pitfalls it could have.
Firstly, it is worth pointing out that Schema isn’t brilliantly documented. To say it is backed by some of the biggest names in tech, it is (ironically) presented in such a way as if the internet hasn’t changed since 2004. It is text-heavy. There are no walk-though videos explaining the potential benefits. In this sense, it reminds me a lot of the W3C website – probably appealing to geeks, but lacking the sense of ‘fizz’ that is necessary to draw in the casual users who will absolutely define the success or failure of Schema or web standards (I once wrote at length about why there is no such thing as “web standards” but my ex-employer has deleted the post – I will revisit the subject in the future).
As such, the technically-minded can pick there way through to discover what it is you have to do to enact Schema. It’s basically adding a load of additional attributes to HTML elements and (disappointingly) adding additional <span> tags around things to fulfill Schema’s structure.
To see some examples, check out the <a href=”www.weirdisland.co.uk/people/murders/peter-sutcliffe-the-yorkshire-ripper.html”>source code of this page</a>. Apologies for my poor coding standards overall – it’s been many years since I considered myself to be a developer. As you can see, there are various additions to the code like:
<span itemscope itemtype="http://schema.org/Answer" itemprop="suggestedAnswer"> and <div itemprop="aggregateRating" itemscope itemtype="http://schema.org/AggregateRating" id="rating">
That sort of thing.
Obviously, it’s fairly trivial to add additional bits of code, but it’s another layer of work to add to your to-do list and does add to nesting and tag redundancy, which runs counter to everything we’ve been told to strive for for the last decade.
And as such, it must compete with other less trivial matters like writing content, maintaining a database, running plug-ins, refreshing the design, promotion etc. So I imagine that unless there is a strong imperative, deploying Schema is going to be well down the list.
Secondly, while deploying the bits of code necessary to enact Schema is fairly trivial, understanding the way that a Schema ‘object’ is constructed is often very frustrating. In my original critique of Schema, I harangued the void about the limited range of things available. The Schema for ‘person‘ for example barely touches on what a person can be and is heavily skewed towards the professions.
And just try understanding why “diet” is a property of “person”.
The best way to actually test your Schema code is basically trial and error and constant testing through Google Webmasters Tools structured data tester. Some of the “errors” are baffling – for example when you are told that an “event” object must be in the future (a proviso I got over by simply ignoring it).
So. Having “done” Schema for my site, what are my findings? Honestly, it’s hard to tell. Positive aggregate star ratings always look pretty attractive in the results, so I have no doubt that my primitive voting system allied to Schema is helping to improve my clickthrough rates. Aside from that though? I can’t say there’s been any notable benefit. I am deliberately not promoting the site as part of my experimentation, beyond automatically tweeting each new post and contributing a couple of comments to threads on other sites. As such, there is little wonder that my traffic remains below 150 visits a day (itself a riposte to those who would have you believe that active SEO promotion is a waste of time.) Since enacting Schema there has been no noticeable leap in the gentle upward slope of traffic, so claims that “doing Schema” is going to transform your website’s performance in itself are probably misplaced.
Nonetheless, enacting Schema has made me think more deeply about the way that data is structured and how I build content. I couldn’t recommend it in all good conscience, but as part of a broad effort to give Google what it claims to want, it is probably a tick worth having if your data in any way fits into any of the available schemas.
And, at the back of it all lurks the suspicion about what happens if all your markup and the trust it helps to build leads to your site getting highjacked by Google itself. For argument’s sake, let’s pretend that my article on the Yorkshire Ripper becomes the definite oversight. With so much Schema data in there – geo co-ordinates and dates of his attacks, properly attributed images, factually correct dates etc – Google could, unilaterally, decide to take that info as gospel and simply pull it through into their knowledge box. And what then for my traffic…?
In conclusion. As an exercise, Schema is worth thinking about and experimenting with, but as a long-term venture it comes with risks that probably at least equal the potential benefits.