Schema Markup and Knowledge Graphs

1. Google’s New Recommendations for Schema Markup

This year, Google introduced a new guideline recommending the addition of Organization schema markup to the homepage. As a result, the homepage should now include two types of markup:

• Organization

• Website

These help Google differentiate your company from competitors, but this differentiation is not always advantageous. It’s crucial to implement these carefully.

Security Risks to Avoid

Google suggests including various details in the markup, such as VAT ID. However, including sensitive data like VAT ID is a significant security risk and must be avoided. Carefully review Google’s guidelines, as some suggestions may do more harm than good.

Trends from the Web Almanac

An increasing number of websites now use similar markups on their homepage, often implemented through plugins or modules. Unfortunately, many plugins create verbose markup, adding unnecessary elements such as the WebPage type and grouping everything into a graph. This practice can lead to:

• Inefficient or non-functional markup.

• Google ignoring the markup entirely.

2. Best Practices for Homepage Markup

To ensure effective and safe implementation, follow these simple rules:

• 🟢 Avoid verbose markup.

• 🟢 Exclude unnecessary information.

• 🟢 Do not mark up everything indiscriminately.

• 🟢 Avoid including sensitive data that could harm your business.

• 🟢 Represent data accurately.

• 🟢 Limit hierarchical markup to two levels.

• 🟢 Align OG (Open Graph) markup with schema and the visible page content.

💡 Pro Tip: If you develop a highly effective markup configuration, consider hiding it from competitors.

3. The Role of Schema.org Vocabulary in AI Systems

In 2024, schema.org vocabulary will play a pivotal role in powering AI chatbots and assistants. These systems rely on schema to verify or retrieve data for generating responses. Notably, they use the general schema.org vocabulary rather than versions tailored for Google.

AI Data Retrieval Methods

AI systems employ two main approaches to answer user queries:

1. RIG (Retrieve-then-Improve Generation): Generate a response first, then verify facts in a knowledge graph.

2. RAG (Retrieve-Augmented Generation): Search the knowledge base first and then generate a response.

Regardless of the method, schema vocabulary remains crucial for:

• Speed

• Decentralized data access

If your schema is incorrect, untrustworthy, or lacks valuable information, AI systems may annotate your content independently, ignoring your provided markup.

4. Differences Between Knowledge Graphs for Search and Chatbots

• Google Search: Focuses on users who intend to or can make purchases.

• Chatbots: Cater to students, researchers, or those seeking information for study or research purposes.

5. Example Knowledge Graph for Chatbots: Data Commons

Data Commons is an example of a knowledge graph that supports chatbots and can also enhance your SEO efforts. Here’s how you can leverage it:

• Access it via API or NLP-enabled interface.

• Integrate it into your website for enhanced research capabilities.

Data Commons is not just for chatbot development—it can also aid in content creation, link-building strategies, and gaining insights into structured data use.