An Open Letter to Bing Regarding JSON-LD

An Open Letter to Bing Regarding JSON-LD Support

Update, 31 July 2018

On 30 July 2018 (in a post dated 2 Aug. 2018) Bing announced JSON-LD support in Bing Webmaster Tools. Bing had previously announced it was supporting JSON-LD at SMX Advanced in June, but this is the first published confirmation that Bing supports JSON-LD. As of 31 July 2018 the Bing structured data page, though, makes no reference to JSON-LD.

Update, 20 June 2018

On 12 June 2018 Bing announced, at SMX Advanced, that it would be adding JSON-LD support for its markup validator. This support was confirmed that same day.

Update, 6 March 2018

As per this discussion, on 2 March 2018 Bing indicated that they now support JSON-LD. However, this has yet to be reflected in their documentation or reflected in their Markup Validator. I’ll update this post when and if these changes surface in official Bing documentation.

Hi Bing, how’s it going?

Pretty well, I think.

I was really happy to learn in October 2015 that Bing had started turning a profit. Competition in the search engine arena is vital, as it provides search engines with an incentive to innovate and, more importantly, ensures that search engines will endeavor to win over users by returning the most useful and relevant search results they possibly can.

But Bing also understands that sometimes it’s in their best interests and that of their users to work cooperatively with their competitors to align on standards and best practices.

This was the case in November 2006, when Microsoft, Google and Yahoo! agreed to support a common sitemaps protocol.

And, more recently, you, Google and Yahoo! joined forces in June 2011 (joined later by Yandex) “to create and support a common set of schemas for structured data markup on web pages”: schema.org. The initiative has been an enormous success: millions of domains now use this standard to provide information about the objects referenced in their web pages.

A structured data vocabulary of course requires a method of encoding it that data consumers can understand, and when schema.org was released the sponsors recommended microdata as the preferred syntax, and soon (after some missteps) embraced RDFa as well.

Microdata and RDFa are both excellent methods of adding schema.org annotations to HTML documents, and few would argue that there was a better standard available for this purpose when schema.org was announced.

In January 2014, however, JSON-LD became a W3C Recommendation.

As you’re doubtlessly aware, JSON-LD offers some significant advantages for webmasters over microdata or RDFa. Since it does not rely upon HTML elements to make declarations it is far less error-prone than microdata or RDFa, and, unlike those syntaxes, does not require the use of tags to make statements about information that is not visibly present in web documents.

Perhaps most importantly, as it is “100% compatible with JSON” it provides a method of deploying linked data that the vast majority of web developers are already familiar with and use on a daily basis in their work.

In recognition of JSON-LD’s benefits, Google has incrementally increased its support for it as a mechanism for providing structured data markup, and it is now its recommended method for doing so. And, at least as evidenced by what’s returned by their validation tools, Yandex, Apple and even Pinterest are now capable of understanding JSON-LD-provided data.

As a result, in particular, of Google’s support for the syntax, JSON-LD has seen significant and rapid adoption on the web. While looking at changes to Web Data Commons data extracted from Common Crawl over time is not a strict apples-to-apples comparison (as Common Crawl does not use the same corpus for each crawl), a comparison of their November 2015 extraction to that of October 2016 nonetheless suggests JSON-LD is quickly gaining in popularity.

Web Data Commons domains with triples by data type, November 2015 and October 2016 Common Crawl extracts

In this environment webmasters, developers and – especially – search marketers are finding themselves having to make an increasingly difficult choice: continue to use a cumbersome inline markup syntax in order to maintain or gain search visibility in Bing, or use JSON-LD with the realization that this will result in the loss of structured data-powered features in Bing search results.

Bing’s failure to support JSON-LD means, too, that their users are increasingly ill-served by that lack of support, insofar as more and more webmasters are choosing the latter of those two options.

That is, as Bing has long generated featured snippets for things like products and recipes based on structured data markup, it must consider these snippets to be of value to searchers.

Bing product rich snippet

Bing recipe rich snippet

But as more and more webmasters use JSON-LD exclusively to provide structured data, Bing users will see fewer and fewer rich snippets, from fewer and fewer publishers, populating their search results.

I’m confident that you’re aware of many, if not all, of the points I’ve made here in order to try and persuade you to publicly and demonstrably support JSON-LD.

And I’m actually fairly hopeful that you’re already making efforts to do so. Nothing could make me happier than to find out that writing this letter was unnecessary by virtue of a soon-to-be made announcement that you have approved JSON-LD as a method of providing Bing with structured data.

But if you’re still equivocating on this course of action, let me make this plea on behalf of developers, search marketers, linked data enthusiasts and Bing users everywhere: please join the party and throw your weight behind JSON-LD.