In case you ever wanted to prompt Google to yield eye-catching rich results for your web page, you would have been requested to submit a number of structured data.
Whether your site is a blog or an e-commerce, Google has been recommending the use of structured data to let Google better understand your page, increase the chance of organic traffic and ultimately help the search engine speed up the crawling and indexing procedures.
Since the rise of Rankbrain in 2015, structured data played a crucial role in allowing Google to define entity relationships. With the search engine honing skills in predicting the search intent and parsing information from semantic attributes, structured data might soon be taken over by the looming NLP and NLU machine learning models, with the automated extraction of unstructured data sources.
In this post, I am going to walk you through structured data optimization in the context of the rising semantic search and provide a few tips on how to buckle up for the optimization of semantic schema markup.
Best Structured Data Format
Before kicking off structured data optimization for semantic search, it is imperative to clear a few points on the existing structured data formats.
Whether Schema.Org officially recognizes a ton of schema markup types, you need to make sure to implement a single structured data format across your website’s pages. Unlikely, stuffing your website with different formats may genuinely confuse web crawlers and curb crawling frequency on your pages.
The structured data formats available to your convenience are as follows:
JSON-LD is Google’s preferred structured data format, as confirmed by John Mueller in a hangout in 2019. In fact, it represents the easiest format that the search engine is able to parse.
It is not by chance that reading, updating and troubleshooting a script such as JSON-LD that is completely separated from the HTML is so much easier than trying to code the structured data within the HTML.
RDFA is an HTML5 extension making largely use of HTML attributes such as “Type of“. Whether it is not considered the worst to devote to in your SEO efforts, it is certainly not even the best as it is bloated with HTML code.
Microdata is a form of structured data which is supported by the Schema.org project as it belongs to the same family. Where fully supported, it makes large use of Itemprop/Itemtype attributes. It is considered one of the most daunting and chaotic to handle with respect to web crawlers
How to optimize Semantic Schema Markup
Semantic search engines have been using knowledge bases such as Wikipedia as sources of data to describe data.
In addition to structured data, Google has also started retrieving pieces of information from unstructured data sources such as authoritative and enterprise sites.
Hence, there are a few concepts to keep in mind before optimizing schema markup in a context where the lines between parsable and un-parsable information are drastically blurring.
🚨 Make it simple
To optimize your schema markup for semantic search, you want to write your copy in ways that translate easily to structured data. Write your content and title tags in Triples or :
subject —> verb —> object
🚨 Markup relevant content only
Against the odds of what has been injected into SEOs’ minds in recent years, not all the pages need a schema mark-up if they are not supposed to ultimately bring additional value to the search
🚨 Mark up existing content on your page
Sure enough, you don’t want to highlight a piece of information that is even not on your page. Because of that, you should recall that Google may penalize your rankings.
🚨 Avoid Schema Stuffing
Other than including multiple formats on your website’s pages, do avoid including multiple schemas in your content as well. This may do nothing but jeopardize crawl performance and affect your organic performance due to a series of misleading content suggestions to search engines and your public.
🚨 Define your primary entity using mainEntity
This might look like a nitpicking tip but it covers much more as the mainEntity attribute is pivotal in describing the page’s relations to other existing entities across your website. This can be done by using
🚨 Link your entities to authority websites to provide context
There is no reason to fear linking to external properties when they are authoritative and fit with the gists of your website. As confirmed by Tim Berners-Lee, it is considered “bad etiquette” not to link to related external material, as it would fall short of the ultimate goal of the Internet as it was devised.
🚨 Nest your entities within other relevant entities
You have to act like Google on a smaller scale. As Google eagerly collects entities and thoroughly devises a huge entity empire, you need to nest and subset your website’s entities to create your own network.
(Un)Structured Data is the Future
Structured data serve the purposes of NLP as they help search engine algorithms better understand the content of your web pages and trigger rich results, thereby prompting potential uplifts in CTR.
Once NLP machine learning algorithms will have eventually geared up the search engine, structured data are doomed to become obsolete. In turn, this might prompt Google to discourage SEOs from using them.
Structured data is code bloat – we shouldn’t need to say the same thing twice (which ironically, I just did).— AlexHarford-TechSEO (@AlexHarfordSEO) July 7, 2022
I’ve also thought about structured data and when it’ll become unnecessary. Perhaps one of the reasons I wasn’t surprised to see these pros and cons in Google’s SERP.
Whether this might happen or not, finding our feet in structured data for semantic research it’s crucial to wrapping our heads around Google’s ever-changing developments.