The Role of Schemas in AI Driven Search

Why Are Schemas So Complicated?

"We need schemas on our website!" "Schemas bolster SEO!" "I think it has something to do with Google's AI Overview, right?"

Sort of, sometimes, kind of.

Let's begin by defining these terms/concepts.

Schemas are created by Schema.org which is a consortium describing themselves as "...a collaborative, community activity with a mission to create, maintain, and promote schemas for structured data on the Internet, on web pages, in email messages, and beyond." 

The best way to think about these schemas, which are constantly being refined and expanded, is as a way to structure and label certain types of content in a consistent, universally acceptable way. Leave SEO or AI out of it for a moment. Schemas are an agreed upon hierarchy of things and attributes of things with common labeling protocols. There is actually a schema for Thing which is the most generic type. 

Example, the schema properties for "Event" are things like :

• "actor" the performing group or person for the event, if applicable
• "aggregate rating" which is only applicable if the event has reviews or ratings
• "door time" just like it sounds, this is the start time

As you can see from this very simple example, if there is data/content related to an event, you would use the schema to understand all of the particulars of an event and how to structure (name) them.

Are Schemas Helpful for SEO?

Yes, some are, but not all. If we think about Google's traditional search results listings, fast being overtaken by AI Overview, only some schemas will be helpful because Google only uses a subset of schemas for SEO purposes. Further, unlike schema.org schemas, where nothing is "required" they are just organizational guidelines, Google has some strict rules on how they expect schema markup in your website code. 

For instance, Google guidelines state that any structured data must also be visible on the page. So, no stuffing things in a schema that are intended to trick search engines.

Google, currently, supports only 3 schema formats; JSON-LD, Microdata and RDFa. JSON-LD is the preferred format 

Some of the schemas supported by Google include:

• Article
• Local Business
• Organization
• Product

However, if required fields are missing from these schemas, they will not be honored, according to Googles' documentation.

Here is a full list of Google Supported Schemas.

If Google does not recognize a schema that fits your content, schema's will not be helpful for SEO. But, that does not mean they may be helpful for other purposes.

Are Schemas Good for AI Overviews in Search?

Yes, schemas do play a role in the AI overviews many web users are seeing at the top of the traditional search results, but much like SEO, they only play one part. Note: this does not mean that all LLMs look at schemas, they do not. 

According to Search Engine Journal, traditional, on-page SEO is a big driver of inclusion in Google's AI Overview. The article sites a study that found 52% of sources mentioned in the AI overview rank in the top ten traditional search results. You can read the full study findings here.

As we know, traditional on-page SEO involves more than just schemas. The tried and true concepts persist such as:

• Attention to your meta data, particularly creating a click worth meta-description
• E-E-A-T guidelines still hold a ton of sway with the search engines.
• Use H tags correctly and use words that specifically speak to the purpose of the page. In other words, having a H1 that reads "Continued Improvement Tips & Tricks" is missing out on important context because it does not speak to what sort of Improvements will be discussed. 
• Create excellent content that your intended audience will love. Click though engagement gives Google a clue as to whether they served up a great search result or a mediocre one.
• Page format & clarity - is the content arranged is a way making is easy to use in AIO? Lists, bullets, Q&A's etc...
• Unique, non-spammy content

So while schemas play a role, the existing set of principles are still relevant. 

What You Need To Know

If we sound like a broken record, that is because this advice rarely changes. Concentrate first on creating insightful, unique content that speaks to your ideal customer. Focus on their needs and goals.

Once you have done that, see if any of this valuable content would lend itself to a schema that is recognized for SEO. Maybe you see that all of your amazing content would make sense presented to your users as FAQs. Great, there is a schema for that and it would likely be valuable to present your content in a way so that the schema can be applied.

Bottom line - lead with rich, meaningful content then organize content that makes sense so a schema can be applied. Do not try to fool the search engines. You may find those pages being deindexed.