Why Traditional SEO is NOT Dead

TL;DR

Classic SEO remains crucial for optimizing generative models.

Transcript

So today I'm going to be showing you why classic SEO is critical for geo or generative engine optimization. So let's dive right into it. So first we need to talk about what actually happens when a large language model is trained. Okay. So there's two things that happen you know when you search inside of chat GBT or any large language model. either ... Read More

Key Insights

• Classic SEO principles still apply to generative engine optimization, emphasizing the need for high-quality and unique content to pass filtering processes.
• Large language models (LLMs) use a rigorous filtering process to select training data, emphasizing text quality, domain whitelisting, and topic authority.
• Domain whitelisting involves using open-source lists of top websites to determine which domains make it into the training data, mirroring traditional page rank methods.
• Topic weighting focuses on the depth of content coverage, encouraging websites to specialize in specific topics to build authority and reduce ambiguity.
• The training process for LLMs involves multiple stages of filtering, with only a small percentage of documents making it to the final training set.
• Retrieval Augmented Generation (RAG) enhances LLMs by retrieving real-time data from search engines, emphasizing the importance of traditional SEO for real-time optimization.
• Synthetic queries generated by LLMs may not always reflect actual demand, highlighting the need for SEO strategies to adapt to these new query patterns.
• Citation analysis reveals that domains with strong link profiles and high authority scores are more likely to be cited by LLMs, reinforcing the importance of backlink strategies.

Questions & Answers

Q: Why is classic SEO still important in the context of generative engine optimization?

Classic SEO remains crucial because it ensures that content is of high quality, unique, and relevant, which are key factors in passing through the filtering processes of large language models. These principles help in optimizing content for both static and enhanced responses, allowing it to be included in training data and cited in real-time search results.

Q: How do large language models select their training data?

Large language models use a rigorous filtering process to select their training data. They evaluate documents based on text quality, domain whitelisting, and topic weighting. Only a small percentage of documents make it through these filters, ensuring that the final training set consists of high-quality, relevant, and authoritative content.

Q: What is domain whitelisting, and why is it important?

Domain whitelisting involves using open-source lists of top websites to determine which domains are included in the training data for large language models. This method mirrors traditional page rank methods and emphasizes the importance of having a strong link profile and being listed among the top domains to influence LLMs and improve SEO outcomes.

Q: What role does Retrieval Augmented Generation (RAG) play in LLMs?

Retrieval Augmented Generation (RAG) enhances large language models by retrieving real-time data from search engines, allowing them to provide more accurate and up-to-date responses. This process underscores the importance of traditional SEO in ensuring that content is optimized for real-time search results, as it affects how LLMs enhance their static models.

Q: How can synthetic queries impact SEO strategies?

Synthetic queries generated by large language models may not always reflect actual search demand, posing challenges for traditional keyword research. SEO strategies need to adapt by targeting these queries despite the lack of search volume data, as they can still influence LLM responses and citations, potentially driving traffic and authority.

Q: What insights can be gained from citation analysis in LLMs?

Citation analysis reveals that domains with strong link profiles and high authority scores are more likely to be cited by large language models. This finding reinforces the importance of backlink strategies and being listed among the top domains to improve the chances of being included in LLM responses, thereby enhancing SEO effectiveness.

Q: Why is topic weighting significant in the training process of LLMs?

Topic weighting is significant because it evaluates the depth of content coverage on specific topics, encouraging websites to specialize and build authority in particular areas. This approach reduces ambiguity and improves the chances of being included in the training data of large language models, aligning with traditional SEO practices of topic specialization.

Q: How does the training process of LLMs affect SEO practices?

The training process of large language models involves multiple stages of filtering, with only a small percentage of documents making it to the final training set. This process affects SEO practices by emphasizing the need for high-quality, unique, and authoritative content that aligns with traditional SEO principles, ensuring better optimization for both static and enhanced LLM responses.

Summary & Key Takeaways

• Classic SEO remains essential for optimizing generative models, as it ensures content quality and relevance, helping it pass through LLM filtering processes. The focus is on creating unique and high-quality content that aligns with traditional SEO principles.

• Large language models use a rigorous filtering process involving text quality, domain whitelisting, and topic weighting to select training data. This process highlights the importance of building authority and specialization in specific topics for better optimization.

• Retrieval Augmented Generation (RAG) enhances LLMs by retrieving real-time data from search engines, making traditional SEO crucial for real-time optimization. Understanding query patterns and citation analysis can guide effective SEO strategies in the evolving landscape.