Nik Ranger.pdf
本地来源：seo-llm/raw/seo知识库/TXT整理/seo方法论/ppt/2025 Shenzhen SEO Conference Speaker Slides/Nik-Ranger.pdf.txt
A Technical SEO's Adaptation For AI Visibility Advanced model interpretability and brand perception analysis powered by machine learning and natural language processing. Nik Ranger: Senior Tech SEO Consultant Dejan Marketing, Dejan AI ➢ ➢ ➢ ➢ ➢ Strategy Lead Technical SEO Specialist Developer Liaison Machine Learning Google Women Techmakers Ambassador ➢ SEO Collective Australia ➢ Published Author Nik Ranger 2025 2000s 2025 G G #1 #2 #3 #4 #5 #6 The Search Landscape Has Fundamentally Changed #1 #7 #8 #2 #9 #3 #10 #4 Nik Ranger 2025 2000s 2025 G G #1 #2 #3 #4 #5 #6 #7 Modern Search Engines Are Still Fundamentally Based On Information Retrieval #1 #8 #2 #9 #3 #10 #4 Nik Ranger 2025 2000s 2025 G G #1 #2 #3 #4 #5 #6 But They’re Now Powered By Two Distinct Layers Of AI Augmentation #1 #7 #8 #2 #9 #3 #10 #4 Nik Ranger 2025 G #1 #2 #3 #4 The Interpretative Layer: This is the part everyone sees. The AI overview, the chat response. It's the final output. The Agentic Layer: This is the backend, where the AI makes its decisions: what queries to expand on, whether to search for fresh information, and which sources to trust. This is what we need to influence. Our job as SEOs is evolving. It’s no longer just about ranking, it's about a two step mission: understanding and control. Part 1: Currently, how is your brand being impacted by AI? Your AI Visibility Part 2: How can we influence the AI models to recommend your brand in their results? Nik Ranger 2025 Part 1 Before you can strategise, you need an accurate diagnosis. You need to understand the 'as-is' reality of your AI visibility. Nik Ranger 2025 Part 1 in three layers: #1 - Isolate Google algorithmic impact & technical discovery #2 - Map GSC query landscape to understand grounding. #3 - Audit SEO foundational health. Nik Ranger 2025 #1 The Diagnostic Framework: How Are You Being Impacted? Modern sophisticated Algorithm updates have necessitated intelligent means of detecting patterns from traffic change data to answer questions like: ➢ What happened? ➢ Was it something we did? ➢ Or was it a Google update? Google Algorithmic Events Nik Ranger 2025 Common Traffic Impact Reasons Keyword Intent Shift Spam Low Quality Content AI Results Competition Data Accessibility & Architecture Issues Localisation Nik Ranger 2025 Diagnose Traffic Impact The principle is simple: you have to establish a performance baseline. You can do this by analysing at historic GSC data to understand your site's unique seasonality and trends. By creating a forecast of what your traffic should have been, you can measure the difference (pDelta) between expected CTR vs reality. Google Algorithmic Events Traffic Gain +pDelta Baseline Traffic Loss -pDelta Nik Ranger 2025 #2 Map GSC Query Landscape: Query Intent Classification Identify Patterns. Remove Noise. Scope Size Of Issue. Workshop Solutions. Search Console Data Query Intent Classification Branded Traffic Non-Branded Traffic Customer Journey Funnel QDG Classifier Traffic Impact Hypothesis AI Influenced Queries Nik Ranger 2025 Why QDG Classifier? Understand the two minds of an AI. Ungrounded Queries Grounded Queries ➢ Results are generated from the AI's core memory, everything it learned during its initial training. ➢ Results are generated by leveraging RAG (retrieval augmented generation) using real time live organic search results and the AI’s core memory. Nik Ranger 2025 Why QDG Classifier? Understand the two minds of an AI. Grounded Queries As an SEO, we can influence the grounded query results. ➢ Results are generated by leveraging RAG (retrieval augmented generation) using real time live organic search results and the AI’s core memory. Nik Ranger 2025 #3 Audit SEO Foundational Health Identify Patterns. Remove Noise. Scope Size Of Issue. Workshop Solutions. Query Classification Branded Traffic Non-Branded Traffic Tech Audit Customer Journey Funnel Content Audit Backlink Audit Priority Matrix Traffic Impact Hypothesis AI Influenced Queries PPC Audit Nik Ranger 2025 Data Accessibility & Architecture Issues AI Results Keyword Intent Shift Spam Localisation Part 2 Now that we have a diagnosis, we can build a strategy to influence AI models. This is about proactively shaping the AI's perception and becoming its preferred source. Nik Ranger 2025 Part 2 in three layers: #4 - Measure what you want to influence - AI Rank. #5 - Build topical authority with query fan-out research. #6 - Tree Walker analysis: unpack the AI’s decision making process and understand low confidence/ uncertainty. Nik Ranger 2025 Let's do this together (Workshop Exercise): Everyone, pull out your phone and open your preferred chatbot: Gemini or ChatGPT. Nik Ranger 2025 I want you to run this prompt for your own brand: Prompt: ➔ List ten things you associate with [Your Brand]. Example: List ten things you associate with Dejan AI. Nik Ranger 2025 I want you to run this prompt for your own brand: Prompt: ➔ List ten brands you associate with [Your Core Service]. Example: List ten brands you associate with Technical SEO experiments. Nik Ranger 2025 The lists you're seeing right now are your baseline. The fundamental principle is this simple, bi-directional survey. It's the first step to turning an abstract goal into a measurable one. Nik Ranger 2025 #4 Measure what you want to influence - AI Rank We track the frequency and rank of your brand's perception inside the models for Gemini, ChatGPT and Claude using a bi-directional probing methodology combining these into a Weighted Score, which becomes our primary KPI for AI visibility: Brand To Entity Entity To Brand What are the top 10 things you associate with my brand? Who are the top 10 brands for [Your Products / Services]? Nik Ranger 2025 To support the learning and development of our global SEO community, we’ve made the full model of AIRank FREE for everyone https://airank.dejan.ai/ Nik Ranger 2025 #5 Build Topical Authority for an AI Audience AI Overviews cite pages that are the definitive, most comprehensive answer for an entire topic. To become that source, we have to shift our thinking from keyword research to topical modeling. We call this the Query Fan-Out. The idea is to take a core parent topic and expand it into its full constellation of intent. You need to anticipate and answer every related question a user / AI agent would have on that subject. Brand Organic Queries Nik Ranger 2025 Query Demand Estimator Dan Petrovic 2025 https://dejan.ai/blog/fan-out-query-search-volume-prediction-using-deep-learning/ Let's do this together (Workshop Exercise): Look at the [Core Service] you selected to search for with either Gemini or ChatGPT. On a notepad, brainstorm five different sub-topics or questions related to it. Think about: Nik Ranger 2025 ● Definitional questions: What is...? ● Comparative questions: [Your Product/Service] vs. [Competitor] ● Problem solving questions: How do I solve X with [Your Service]? This exercise is a manual version of what we do with our own trained models. We take a client's core topics and generate these entire semantic clusters. The output is a strategic blueprint for a single, authoritative page. This makes your content the most efficient and logical source for an AI to use, dramatically increasing your chances of being cited. Nik Ranger 2025 #6 Tree Walker analysis: Deconstruct the AI's Thinking For true competitive advantage, you need to go deeper than the final answer to understand the answers the AI almost gave. Tree Walker maps these decision paths. Nik Ranger 2025 When an AI writes a sentence, it's making a statistical choice at every single word: BRAND 1/30 unique insights uncovering tree walker analysis in depth. Nik Ranger 2025 Example: Brand association to tours and activities BRAND BRAND BRAND BRAND Graph: The graph plots the model's confidence for each word/token it generates. The higher the line, the more certain it was about choosing that word. 2/30 unique insights uncovering tree walker analysis in depth. Nik Ranger 2025 Example: Brand association to tours and activities BRAND BRAND BRAND BRAND The model is highly confident (nearly 100%) for BRAND, primarily, promotes, and tours. This is fantastic news. The model has a very strong, established association in its core memory between your brand and the concept of tours. 2/30 unique insights uncovering tree walker analysis in depth. Nik Ranger 2025 Example: Brand association to tours and activities BRAND BRAND BRAND BRAND Token Position 4, for the word and, the AI has low confidence signalling semantic uncertainty. The AI sees BRAND as an authority on tours, but the conceptual pairing of tours and activities is weak in its mind. 2/30 unique insights uncovering tree walker analysis in depth. Nik Ranger 2025 Example: Brand association to tours and activities BRAND BRAND BRAND BRAND It could have ended the sentence at tours. It could have chosen a different concept entirely. The fact that it chose activities but with such low confidence tells us something strategically vital. Content and PR strategy needs to work on strengthening the association between those two terms. 2/30 unique insights uncovering tree walker analysis in depth. Nik Ranger 2025 #1 - Isolate Google algorithmic impact & technical discovery. Part 1 #2 - Map GSC query landscape to understand grounding. #3 - Audit SEO foundational health. #4 - Measure what you want to influence - AI Rank. Part 2 #5 - Build topical authority with query fan-out research. #6 - Tree Walker analysis: unpack the AI’s decision making process and understand low confidence/ uncertainty. The future of SEO isn't about choosing between traditional SEO and AI. It's about integrating them. It’s about building genuine, measurable authority in the mind of the machine. Thank you to Dan Petrovic & our DEJAN team. Blake Walsh Martin Reed Mike Jolly Dan Petrovic Senior SEO Strategist Senior Technical SEO Specialist Senior Link Building & Digital PR SEO, AI, ML Senior Specialist ● ● ● ● ● ● ● ● Audits and Analysis Competitive Analysis Content & Link Earning Project Management Audits and Analysis Technical Marketing Opportunity Detection Developer Guidance & Support ● ● ● ● Strategy Design Audits and Analysis Content & Link Earning Opportunity Detection ● ● ● ● Strategy Contributor Machine Learning Content & Link Earning Opportunity Detection Thank you! Thank You For Inviting Me, Organising This Conference & Supporting Our Search Community
本文档为站内渲染。原始文件本地路径：saas/source/seo-llm/raw-seo知识库-TXT整理-seo方法论-ppt-2025-Shenzhen-SEO-Conference-Spe-571139.txt（仅本地保留，不入库不部署）