GEO Research Suite - The thought leading podcast for Generative Engine Optimization and SEO

In this episode a new Microsoft patent is in focus. The patent provides information on how Copilot selects sources in the RAG process. It is important for LLMO / GEO to know which criteria are used in order to get clues for optimization. It also shows how important classic SEO or optimization for search engine rankings is for the output of generative AI.The Microsoft patent titled "Informational g...

The podcast examines how Google extracts information from unstructured data sources and integrates it into its Knowledge Graph. He highlights the challenges of ensuring completeness, validity and timeliness of the data, especially when dealing with tail entities. Various methods such as named entity recognition, event extraction and the use of machine learning are explained. Finally, the Knowledge...

The podcast discusses the importance of E-E-A-T (expertise, experience, authority, trust) for Google rankings. E-E-A-T goes beyond simple SEO and requires a holistic approach that includes content marketing, PR and branding. The author argues that Google ranks websites based on entities and their authority, with on-page and off-page signals playing a role. Successful ranking therefore requires col...

In this podcast episode are discussed three patents by Google researcher Dan Popovici about the Click and Skip Re-ranking concept. Google's three patents describe methods for improving the relevance of search results by analyzing user interactions (clicks and skips). One patent evaluates the proximity of search terms in the results, another the optionality of search terms and the last the reorderi...

The podcast deals with Google's development towards a semantic search engine, the individual steps and innovations of Google since 2010, in particular the role of the Knowledge Graph, machine learning and natural language processing. It also focuses on a deeper understanding of semantic search, the importance of entities and their impact on ranking, and how the Knowledge Graph works. https://www.k...

The podcast explores the concept of “Digital Authority Management”, a new marketing discipline that highlights the growing importance of E-E-A-T (Experience, Expertise, Authoritativeness, Trust) in the context of search engine optimization and generative AI. E-E-A-T is described as an important quality classifier that influences visibility in search results and guides the selection of trustworthy...

The Google patent US10055467B1 discussed in the podcast describes a system for improving the relevance of search results. It uses group-based modification factors, calculated from metrics such as repeat clicks and direct visits, to adjust the initial scores of search results. Thresholds control the strength of the adjustment, treating navigation and non-navigation queries differently. The goal is...

The podcast discusses how Google uses entities (people, organizations, products, etc.) and their digital representations (websites, social media profiles, etc.) to assess the relevance, authority and trustworthiness of content. This is done in the context of semantic search and the E-E-A-T principle (expertise, experience, authority, trust). Google links entities via relationships and attributes i...

The podcast explains how search engines such as Google use Natural Language Processing (NLP) to build semantic databases such as the Google Knowledge Graph. He describes the structure of the Knowledge Graph as an entity-based index system, the different data sources (structured, semi-structured and unstructured), and the role of NLP technologies such as BERT and MUM in the extraction and validatio...

The Google paper “Knowledge-Based Trust: Estimating the Trustworthiness of Web Sources” presents an algorithm for evaluating the credibility of websites. Instead of conventional methods based on links or user behavior, the algorithm focuses on the factual accuracy of published information. An important aspect is the automated extraction and verification of facts using natural language processing (...

The first source is an excerpt from the Google Search Central documentation and describes in detail how Google evaluates helpful and reliable content. It provides checklists and guidelines for creating user-centered content and explains concepts such as E-E-A-T and the role of user signals in ranking. The second source, a blog post by Olaf Kopp, also analyzes Google's evaluation of “helpful conten...

The podcast discusses a research paper from Google for improving zero-shot LLM rankers by incorporating finely granular relevance labels into the prompts. Instead of binary labels (“Yes”/“No”), multi-level labels (e.g. “Highly relevant”, “Somewhat relevant”, “Not relevant”) are used to increase the accuracy of the ranking. Experiments on eight BEIR datasets show significant performance improvement...

The podcast discusses an extended model for analyzing search intentions, the so-called “micro intents”. In contrast to classic categories such as “informational” or “transactional”, it breaks down search intentions into more detailed subcategories in order to create more relevant content. This is illustrated using examples and a visualization of the customer journey that shows different phases and...

The article discussed in the podcast explains the BM25 algorithm, which is used by search engines to determine the relevance of documents for a specific search query. It explains the components of the BM25 formula - term frequency, inverse document frequency and document length normalization - and compares BM25 with the TF-IDF algorithm. https://www.kopp-online-marketing.com/what-is-bm25

The podcast discusses the development of a “Prompt Canvas”, a structured framework for the effective design of prompts for large language models (LLMs). Based on a systematic literature review, various prompting techniques are identified and visualized in the canvas to provide practitioners with a clear guide for application. The canvas includes categories such as persona/role, task/intention, con...

Olaf Kopp presents the “SEO Research Suite”, a database with analyses of SEO-relevant patents and research papers. The suite offers members AI-supported research assistants and thus accelerates access to in-depth knowledge. It aims to turn SEO experts into thought leaders through exclusive insights and time-saving tools. Positive user reviews underline the added value of efficient research and com...

This podcast discusses the article Information gain score: How it is calculated? Which factors are crucial? by Olaf Kopp. This episode explains information gain as a ranking factor of modern search engines. It is described the calculation of information gain using entropy and machine learning models, its application in various areas of search engines (ranking, filtering, personalization) and the c...

The podcast deals with the personalization of search results at Google. Google's approach is discussed based on official documents, patents and data leaks, including metrics used such as user interactions and location data. The impact of this personalization on SEO, in particular the increasing variability of rankings and the importance of user engagement, will be discussed. Finally, strategies fo...

The podcast deals with the Google Shopping Graph, a semantic database for products that is closely related to Google's Knowledge Graph. The Shopping Graph is fed by various data sources such as manufacturer websites, online stores and YouTube videos and is central to the new search technology with generative AI (e.g. SGE). The texts explain how the Shopping Graph works, its importance for e-commer...

This podcast discusses the new article by Olaf Kopp on Search Engine Land. The article describes the new challenge of Generative Engine Optimization (GEO), i.e. the optimization of content for generative AI applications such as ChatGPT or Perplexity. It explains how these applications work, in particular the process of Retrieval Augmented Generation (RAG), which uses external sources to answer que...

The Google patent of the week is about the patent “Ranking Search Results”. The patent describes methods and systems for improving the relevance of search results. It concerns the adjustment of initial search result scores based on group membership of the resources and other factors such as click rates and dwell time. The modification of the scores is intended to improve the quality of the display...

The article discussed in the podcast the factors that Google considers when evaluating E-E-A-T (expertise, experience, authority, trustworthiness). The author, Olaf Kopp, has compiled over 80 different signals from Google patents and other sources that reflect these factors. He divides these signals into three levels: Document level, Domain level and Source level. The signals at each level provide...

The podcast deals with an article by Olaf Kopp, who describes three important ranking methods of modern search engines: BM25, Vector Ranking and Semantic Ranking. BM25 is based on keyword matching, vector ranking uses the geometric relationship between words, and semantic ranking focuses on the meaning of search queries. The text explains in detail how each method works and when they are used most...

The Google patent describes a system and methods for document evaluation for search engines. The evaluation is based on historical data , such as link development, content update frequency, user behavior and traffic. The goal is to improve search results by evaluating and ranking documents based on various factors, including the detection of spam activity. The system uses various weighting factors...

The podcast accompanying the article examines the influence of generative AI systems on companies' SEO strategies. It sheds light on how Large Language Models (LLMs) work and the importance of “LLMO” (Large Language Model Optimization). LLMO aims to place companies prominently in the results of AI chatbots such as ChatGPT or Gemini. The text presents various strategies for LLMO, including the opti...