My Changed Perspective on Structured Data

Up to quite recently I was a bit of a structured data skeptic. I thought schema.org markup was just an admission of defeat from Google, an implication that they’d given up the fight to try and understand web content properly and instead left it up to webmasters to do Google’s hard work for them.

However, my opinion on structured data has changed over the past few months, to the point where now I’m an active proponent of its implementation.

I can’t quite pin my conversion from a schema-hater to a schema-rater to a specific moment in time. I think it was a gradual shift in perspective as I was made aware of many other different uses that structured data might have in the future (aside from making Google’s job easier).

I’ve now come to believe that structured data is, in fact, the future of the internet. At some stage all websites will be made up of structured data, and this will make every bit of information published online an active node that can be tapped in to by other applications.

For example, right now an ecommerce site needs to supply product feeds to third party sites such as shopping comparison engines. With widespread adoption of structured data, the burden of effort can be shifted and shopping comparison systems can crawl ecommerce sites and retrieve all the product information they need. (Or, more likely, a third party product indexer will sit in the middle and do the product discovery & indexing for the shopping sites.)

In his comment on a Search Engine Land post, Aaron Bradley explains it perfectly:

“One of the principles of the semantic web – of which structured data is the, um, structural foundation – is the separation of the data layer from the presentation layer. The benefits of this are that data can be provided for machine consumption without impacting the display of information to humans, and that richer and more precise data can be offered for machine consumption.

This is the reason why the HTML standard was actually NOT sufficient, and why XML has become utterly indispensable in the web ecosystem.

HTML5 notwithstanding (which itself recognizes the insufficiency of earlier HTML), structured data markup leverages the power of HTML to enrich web documents without the need to come up with new HTML tags; from a data standpoint, it provides HTML with almost limitless extensibility.

[…] People have been labeling things from the beginning of recorded history, and computer technology doesn’t magically free us from the necessity of classification, though it does enhance and extend the benefits of doing so.

A public library card catalog in 1970 was produced by people who classified books by the use of formal data fields and controlled vocabularies, exactly as catalogers of books continue to do so today using computers, and this procedure is fully analogous to publishers marking up web pages with structured data today. Organized information is more useful information. [emphasis added]

Google and other data consumers do not seek structured data solely because it “will make things so much easier for us.” In numerous realms structured data makes the discovery, classification and linkage of information not only “easier” but possible in the first place.”

I genuinely believe that making information widely available online in a structured format that will allow third party systems to engage with it has more benefits than drawbacks. Yes, you will see some scrapers find ways to make malicious use of structured data.

But on the whole I believe it will be a good thing, as it will facilitate exchange of information across many different online platforms. And that, in turn, allows for all kinds of cool technology to be developed that will make all our lives a lot easier.