VZ editorial frame
Read this piece through one operating lens: AI does not automate first, it amplifies first. If the underlying decision architecture is clear, AI scales clarity. If it is noisy, AI scales noise and cost.
VZ Lens
From a VZ lens, this piece is not for passive trend tracking - it is a strategic decision input. The Evolution of Search Engine Optimization from 1998 to 2026: Techniques, Data, and Practical Implications of Traditional SEO, GEO, AEO, and the Future LLMO Era. Its advantage appears only when converted into concrete operating choices.
Summary: This document takes you through the entire history of search engine optimization—from the invention of Google PageRank in 1998 all the way to the era of LLMO (Large Language Model Optimization) in 2026. We examine four major eras: traditional SEO, GEO, AEO, and LLMO. For each, we present the essentials, techniques, data, and practical implications.
The Smell of Coffee and the Future of Search
I warm my palms around my mug on the windowsill of the café. Outside on the street along the Danube, people hurrying by cut through the November fog, each with a screen as their destination. The bitter scent of steaming espresso mingles with the sweet fragrance of fresh pastries. My finger traces the cold glass as I watch a person stop, take out their phone, and type something—perhaps an address, a question, a desire. The movement is as natural as breathing. I think about how much this moment, this search, has changed since I first sat down at a computer and tried to figure out how to get myself on the screen. The world of keywords, which I knew so well, is dissolving into the steam, just like my tea. Now it’s no longer about what we type in, but what we want to say.
Table of Contents
- The SEO Era (1997–2020)
- The GEO Era (2020–2024)
- The AEO Era (2023–2025)
- The LLMO Era (2025–2026)
- Connections and Transitions
- Practical Guide
- Resources
1. The SEO Era (1997–2020)
What is SEO?
SEO (Search Engine Optimization) is the practice of designing websites so that search engines—primarily Google—display them as high as possible in the search results (SERP = Search Engine Results Page).
Think of it like a library: SEO is the method you use to ensure that the librarian (Google) recommends your book to the reader first.
Major milestones in chronological order
| Year | Event | Impact |
|---|---|---|
| 1997 | First use of the term “search engine optimization” (Webstep Marketing Agency) | The profession is born |
| 1998 | Google launches with the PageRank algorithm | The number and quality of links become the basis for ranking |
| 2000-2010 | The “content farm” era | Mass-produced, low-quality content dominates search results |
| 2003 | Google AdSense launches | Financial incentive for content creation |
| 2005 | Introduction of the nofollow tag | Protection against spam links |
| 2011 | Google Panda update | Penalizes thin, duplicated content — demotes entire pages |
| 2012 | Google Penguin update | Penalizes artificial link building (link schemes) |
| 2013 | Google Hummingbird | Semantic search — Google begins to understand the meaning of sentences, not just keywords |
| 2015 | ”Mobilegeddon” — mobile-friendly update | Mobile-optimized pages receive an advantage |
| 2018 | Mobile-First Indexing | Google considers the mobile version to be primary |
| 2019 | BERT update | Natural language interpretation improves for 10% of searches |
| Aug. 2022 | Helpful Content Update | The entire website may be penalized if it contains a lot of content “written for search engines” |
| Dec. 2022 | Introduction of E-E-A-T (Experience, Expertise, Authoritativeness, Trust) | The previous E-A-T is expanded to include Experience |
Explanation of PageRank
PageRank was invented by Larry Page and Sergey Brin (Stanford) in 1996–97. The basic idea is simple: if many high-quality websites link to you, then you are important too. It’s like the academic world: if many researchers cite your study, it means you’ve done important work.
The logic of PageRank (simplified):
[Page A] ──link──> [Page B] ──link──> [Your page]
[Page C] ──link──────────────────────> [Your page]
[Page D] ──link──────────────────────> [Your page]
The more links you have, and the better the sources, the higher your rankingThe Impact of Panda and Penguin
The 2011 Panda and 2012 Penguin updates brought about a revolution. Panda penalized “content farms”—sites that mass-produced low-quality articles to “game” search results. Penguin targeted artificial link building: if someone acquired links from a thousand irrelevant websites, Google would penalize them.
The E-E-A-T Framework
In December 2022, Google introduced the E-E-A-T (formerly E-A-T) quality framework:
| Letter | English | Hungarian | What does it mean? |
|---|---|---|---|
| E | Experience | Experience | The content creator has their own, firsthand experience with the topic |
| E | Expertise | Expertise | The author possesses the necessary knowledge on the topic |
| A | Authoritativeness | Authoritativeness | The website and the author are well-known, reliable sources |
| T | Trust | Trust | The site is accurate, honest, safe, and trustworthy (the most important factor) |
Important: E-E-A-T is not a direct ranking factor in the algorithm, but rather an evaluation framework used by human quality raters. However, Google’s algorithm attempts to model these signals.
The End of the SEO Era
By around 2020, the world of SEO had become highly sophisticated: technical optimization, content strategy, user experience (UX), mobile-friendliness, Core Web Vitals—all of these factors together determined search rankings. But search basically still worked like this: you type in a keyword → you get a list of links → you click on one of them.
This model began to change after 2020.
2. The GEO Era (2020–2024)
What is GEO?
GEO (Generative Engine Optimization) is an extension of traditional SEO: instead of just optimizing for search engine rankings, the goal is for AI-based answer-generating systems (ChatGPT, Google AI Overviews, Perplexity, Claude) to link to us, quote us, and mention us in their responses.
Think of it this way: in traditional SEO, you’re fighting to be number one on Google’s list. In GEO, you’re fighting so that when ChatGPT or Google AI answers a question, it uses your content as a source.
Who came up with the concept of GEO?
The term GEO and its scientific framework were created by a research group led by Princeton University in November 2023. The authors of the study are:
- Pranjal Aggarwal (Princeton)
- Vishvak Murahari (Princeton)
- Tanmay Rajpurohit (Georgia Tech)
- Ashwin Kalyan (Allen Institute for AI)
- Karthik R. Narasimhan (Princeton)
- Ameet Deshpande (Princeton)
The title of the paper is: “GEO: Generative Engine Optimization” (arXiv:2311.09735, November 16, 2023). It was also presented at the prestigious ACM SIGKDD conference in 2024.
Key findings of the Princeton study
The researchers demonstrated that certain optimization strategies can increase content visibility by up to 40% in generative engine responses. This is a huge number—it means that what AI systems cite can be actively influenced.
GEO strategies (based on the study)
| Strategy | Description | Impact |
|---|---|---|
| Including statistics | Adding specific numbers, percentages, and research data | +30–40% visibility |
| Quotes and citations | Citing sources, researcher/study names | Strong positive impact |
| Simplification | Easier-to-understand, clear wording | Moderate positive impact |
| Technical terminology | Use of precise, professional language | Domain-dependent — helps in some areas, not in others |
| Narrative structure | Storytelling, context-providing structure | Weak impact |
How does GEO differ from traditional SEO?
TRADITIONAL SEO GEO
=================== ===================
Goal: Ranking on SERPs Goal: Citation in AI responses
Metrics: Position, CTR, traffic Metrics: Mentions, citations, "citation share"
Channel: Google/Bing search results Channel: ChatGPT, Perplexity, AI Overviews
Technique: Keywords, links Technique: Structure, data, authority
User: Clicks → visits User: Reads the AI’s response (may
never click through to the page)The Context of GEO: Why Was It Needed?
Between 2020 and 2024, several parallel developments occurred:
- Nov. 2022: ChatGPT launches — people start asking AI the questions they used to search for on Google
- May 2023: Google announces Search Generative Experience (SGE, later renamed AI Overviews) — Google itself begins providing AI-generated answers
- Nov. 2023: The Princeton GEO study is published
- 2024: The popularity of Perplexity AI skyrockets (370% annual growth)
3. The AEO Era (2023–2025)
What is AEO?
AEO (Answer Engine Optimization) refers to optimizing content so that AI-based “answer engines”—such as Google AI Overviews, ChatGPT, Perplexity, or voice assistants (Siri, Alexa, Google Assistant) — select and display your content directly as the answer.
Think of it this way: if you ask Google, “At what temperature does water boil?”, you don’t want 10 links—you want a single, accurate answer. AEO optimizes so that your answer is the one the AI selects.
The History of AEO: Featured Snippets and Voice Search
AEO didn’t come out of nowhere. It had two important precursors:
1. Featured Snippets — since 2014 For certain searches, Google began displaying an answer in a highlighted box at the top of the results page. This was the first sign that Google wanted to provide “answers,” not just links.
2. Voice Search — since 2016 With the spread of smart speakers (Amazon Alexa, Google Home) and voice assistants, the format of search queries changed: people no longer typed “Budapest weather,” but asked, “What will the weather be like tomorrow in Budapest?” And the voice assistant gives a single answer—not 10 links.
Data: According to Google, 41% of voice search results come from Featured Snippets.
AEO Techniques
| Technique | What does it mean? |
|---|---|
| Question-answer structure | Questions in content headings, concise answers below |
| FAQ Schema | Structured data (schema.org) that signals to the search engine: “there are question-answer pairs here” |
| HowTo Schema | Step-by-step guides in a structured format |
| Passage-level optimization | Individual paragraphs on the page are meaningful and quotable on their own |
| Concise, accurate answers | The first 1–2 sentences directly answer the question |
AEO by the Numbers
| Metric | Data |
|---|---|
| Percentage of zero-click searches (with AI Overviews) | ~43% |
| Zero-click rate in Google AI Mode | ~93% |
| CTR drop when AI Overviews appear | -61% (1.76% → 0.61%) |
| CTR increase when AI Overviews quote your page | +35% |
AEO and the “zero-click” revolution
“Zero-click search” means that the user receives the answer directly on the search results page without clicking through to any website. This trend has been growing since 2019, but it has accelerated dramatically with the introduction of AI Overviews (2024–2025).
Changes in search behavior:
2010: Search → 10 blue links → Click → Read
▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ (80%+ clicked)
2020: Search → Featured Snippet + 10 links → Maybe a click
▓▓▓▓▓▓▓▓▓▓▓▓▓▓░░░░░░ (~65% clicked)
2025: Search → AI answer + a few links → Rarely clicked
▓▓▓▓▓▓░░░░░░░░░░░░░░ (~39% clicked for AI Overviews)
2026: AI Mode → Full AI response → Almost never a click
▓░░░░░░░░░░░░░░░░░░░ (~7% clicked)4. The LLMO Era (2025–2026)
What is LLMO?
LLMO (Large Language Model Optimization) means writing and structuring our content so that large language models — such as ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), and Perplexity — understand, utilize, and reference it in their responses.
If SEO was about convincing the librarian, then LLMO is about convincing artificial intelligence: you want the AI to use your content as its answer when someone asks it a question.
Who coined the term LLMO?
The term LLMO does not originate from a single academic study (unlike GEO). Rather, it is the result of collective language creation within the digital marketing industry during 2024–2025. Several players used it simultaneously:
- Semrush — one of the largest SEO tool companies — launched its “LLM Optimization” guide in 2025
- Neil Patel — the marketing influencer — wrote a comprehensive comparison in 2025 (AEO vs GEO vs LLMO)
- Search Engine Land — the industry media outlet — published an LLMO guide in 2025
- Ahrefs — published the article “LLMO: 10 Ways to Work Your Brand Into AI Answers” in 2025
- Adobe — launched the “LLM Optimizer” product in 2025
The term began to spread in late 2024 and became one of the industry’s most important buzzwords by 2025–2026.
Why was a new concept needed?
Because LLMO is different from traditional SEO:
| Aspect | SEO | LLMO |
|---|---|---|
| Target system | Google/Bing search engine | ChatGPT, Claude, Gemini, Perplexity |
| Operation | Crawler indexes the page | LLM reads it during training or via RAG |
| Result | Ranking position | Mention/citation in the AI’s response |
| Ranking factors | Links, keywords, technical signals | Brand authority, data richness, structure |
| User behavior | Click → visits the page | Reads the AI’s response → may never click |
| Metrics | Position, CTR, organic traffic | AI mentions, citation share, brand visibility score |
Key Techniques for LLMs
1. Structured Content
LLMs process well-structured content more effectively. This means:
- Clear headings (H1, H2, H3)
- Tables with data
- Lists (bullet points)
- Question-answer pairs
- Schema.org structured data (in JSON-LD format)
Data: Pages with structured data are cited 2.5 times more often by AI systems than unstructured ones.
2. Passage-level optimization
LLMs do not cite entire web pages, but rather individual paragraphs (passages). Therefore, each paragraph must make sense on its own—it should include context, the answer, and, if possible, data.
3. Strengthening E-E-A-T Signals
AI systems look for E-E-A-T signals 100% of the time before citing a source. Specifically:
- Author name and expertise
- Institutional background
- Indication of firsthand experience
- Original research data
4. Ensuring Crawler Access
AI systems must have access to the content. It is important that AI crawlers are not blocked in the robots.txt file:
| Crawler | Which AI does it belong to? |
|---|---|
| GPTBot | OpenAI (ChatGPT) |
| PerplexityBot | Perplexity AI |
| ClaudeBot | Anthropic (Claude) |
| Google-Extended | Google (Gemini, AI training) |
| Bingbot | Microsoft Copilot |
5. Original Data and Research (Information Gain)
LLMs prefer original data — proprietary research, statistics, and surveys. This is known as “information gain”: providing content that cannot be found elsewhere.
Data: Content containing original statistics achieves 30–40% higher visibility in LLM responses.
6. Brand Strategy and Omnipresence
In LLMO, brand awareness is one of the strongest ranking factors. It’s not backlinks that matter most, but brand search volume — a correlation of 0.334, which is the strongest predictor.
This means: if people search for your brand frequently, the AI “knows” it better and cites it more often.
LLMO by the Numbers (2026)
| Metric | Data |
|---|---|
| AI Chatbot Market Leader | ChatGPT: 64.5% [NOT VALIDATED — sources show varying data: 64.5%–80%] |
| Google Gemini market share | ~21.5% (450M monthly users) |
| Perplexity annual growth | 370% |
| CTR decline in top search results due to AI | -34.5% (Ahrefs, Feb. 2026) |
| Gartner forecast: organic traffic shift to AI | 25% (by 2026) |
| Listicle format in AI citations | #1 — 50% of AI citations |
5. Connections and Transitions
The Four Eras Building Upon One Another
SEO, GEO, AEO, and LLMO are not systems that replace one another, but rather build upon one another. The literature clearly shows that LLMO did not “kill” SEO, but rather expanded it.
EVOLUTIONARY LAYERS (as of 2026):
┌─────────────────────────────────────────────┐
│ LLMO (2025–2026) │
│ ┌─────────────────────────────────────┐ │
│ │ GEO (2023–2024) │ │
│ │ ┌─────────────────────────────┐ │ │
│ │ │ AEO (2023-2025) │ │ │
│ │ │ ┌─────────────────────┐ │ │ │
│ │ │ │ SEO (1997–2020) │ │ │ │
│ │ │ │ Foundation: technical, │ │ │ │
│ │ │ │ content, links │ │ │ │
│ │ │ └─────────────────────┘ │ │ │
│ │ │ + Answer optimization │ │ │
│ │ │ + Featured Snippets │ │ │
│ │ └─────────────────────────────┘ │ │
│ │ + Generative Engine Optimization │ │
│ │ + AI citation / citation share │ │
│ └─────────────────────────────────────┘ │
│ + LLM-specific optimization │
│ + Crawler access, brand, data │
└─────────────────────────────────────────────┘What remains of traditional SEO?
| SEO element | Staying? | Why? |
|---|---|---|
| Technical SEO (speed, mobile, security) | YES | AI crawlers also prefer fast, accessible pages |
| Content quality | YES | E-E-A-T is essential at every level |
| Structured data (Schema) | YES, and it will become even more important | LLMs interpret it 2.5x better |
| Keyword-focused optimization | DECLINING | AI understands semantics; it doesn’t search for keywords |
| Backlink building | DECLINING | Brand awareness is more important than the number of links |
| Meta description, title tag | STAYING, but evolving | AI responses present content in different ways |
| Local SEO | STAYING | Local search remains essential, especially in voice search |
What is disappearing?
| SEO practice | Why is it disappearing? |
|---|---|
| Keyword stuffing | LLMs understand text semantically |
| Artificial link building | Brand trust is more important than link quantity |
| Thin content | AI systems look for “information gain” |
| Clickbait headlines | AI doesn’t click—content quality matters |
| “Rank #1” as a goal | There’s no “rank” in AI responses—they either cite you or they don’t |
Ahrefs’ stance: “GEO, LLMO, AEO—it’s all just SEO”
It’s worth noting that Ahrefs (one of the largest SEO software providers) openly states that “GEO, LLMO, AEO… it’s all just SEO”—meaning they are different names for the same thing. Their argument is that good content, technical optimization, and authority building have always worked, and the same applies in the AI era.
This is partly true. But there are real differences:
- Blocking AI crawlers (robots.txt) was not relevant in SEO
- Passage-level optimization was less important in SEO
- “Citation share” as a metric is completely new
- Brand visibility in AI responses is of a completely different nature than SERP position
Comparison Table: SEO vs AEO vs GEO vs LLMO
| Aspect | SEO | AEO | GEO | LLMO |
|---|---|---|---|---|
| Focus | Ranking in search results | Appearing as a direct answer | Being cited in an AI-generated answer | Being used and referenced by LLMs |
| Target platform | Google, Bing SERP | Featured Snippets, voice assistants | AI Overviews, Perplexity | ChatGPT, Claude, Gemini, any LLM |
| Main technique | Keywords, links, technical SEO | Q&A structure, FAQ Schema | Structure, quotes, statistics | Original data, E-E-A-T, crawler access, brand |
| Metrics | Position, CTR, organic traffic | Position Zero, voice search appearance | Citation share, impressions | AI mentions, brand visibility score |
| ”Wins” if | The page ranks #1 | The page appears in the “featured answer” box | The AI cites the page in its response | The LLM mentions the brand accurately and positively |
| Published | 1997 | ~2019 (Featured Snippets + Voice Search) | 2023 (Princeton study) | 2024-2025 (industry consensus) |
6. Practical Guide
The 2026 Optimization Pyramid
If someone were to start optimizing now, they should follow this order of priorities:
/\
/ \
/ LLM \ ← 4. Brand strategy + AI monitoring
/ mention-\
/ mention \
/────────────\
/ GEO: citation \ ← 3. Statistics, references, data
/ in AI responses \
/────────────────────\
/ AEO: answer optimization\← 2. Question-answer structure, Schema
/────────────────────────\
/ SEO basics: technical + \← 1. Speed, mobile-friendliness, structure,
/ content + E-E-A-T \ quality content
/──────────────────────────────\10-Step Action Plan
| # | Step | Category |
|---|---|---|
| 1 | Technical audit: speed, mobile-friendliness, Core Web Vitals | SEO basics |
| 2 | robots.txt check: Enable GPTBot, PerplexityBot, ClaudeBot | LLMO |
| 3 | Implement Schema.org structured data (JSON-LD): FAQ, HowTo, Article, Product | AEO + GEO |
| 4 | Rewrite content into Q&A format | AEO |
| 5 | Incorporate original data, statistics, and research into every article | GEO |
| 6 | Passage-level optimization: ensure every paragraph makes sense on its own | LLMO |
| 7 | Strengthen E-E-A-T indicators: author name, expertise, institutional background | At all levels |
| 8 | Introduce brand monitoring on AI platforms (e.g., Semrush AI Visibility Toolkit) | LLMO |
| 9 | Omnipresence strategy: Be present in many authoritative sources, forums, and databases | LLMO |
| 10 | Regular measurement: AI mentions, citation share, brand visibility score | LLMO |
Measurement Tools in 2026
| Tool | What does it measure? | Price |
|---|---|---|
| Semrush AI Visibility Toolkit | Visibility on AI platforms | Paid |
| Otterly.ai | Monitoring of AI mentions and links | Paid |
| LLMrefs.com | Generative AI search analytics | Paid |
| Ahrefs | Traditional SEO + AI mentions | Paid |
| Google Search Console | Traditional search performance | Free |
| Am I Cited? | LLM citation checker | Paid |
7. Resources
Academic Resources
- Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K.R., & Deshpande, A. (2023). “GEO: Generative Engine Optimization.” arXiv:2311.09735. Princeton University. — https://arxiv.org/abs/2311.09735
- ACM SIGKDD 2024 proceedings — https://dl.acm.org/doi/10.1145/3637528.3671900
- GEO-optim official website — https://generative-engines.com/
Industry resources
- Neil Patel: “AEO vs GEO vs LLMO” — https://neilpatel.com/blog/aeo-vs-geo-vs-llmo/
- Search Engine Land: “What is LLMO?” — https://searchengineland.com/guides/large-language-model-optimization-llmo
- Ahrefs: “GEO is just SEO” — https://ahrefs.com/blog/geo-is-just-seo/
- Semrush: “LLM Optimization” — https://www.semrush.com/blog/llm-optimization/
- Kevin Indig: “State of AI Search Optimization 2026” — https://www.growth-memo.com/p/state-of-ai-search-optimization-2026
- Clearscope: “The 2026 SEO Playbook” — https://www.clearscope.io/blog/2026-seo-aeo-playbook
- Backlinko: “SEO vs. GEO, AEO, LLMO” — https://backlinko.com/seo-vs-geo
- Onely: “GEO vs. AEO vs. AI SEO vs. LLMO” — https://www.onely.com/blog/geo-aeo-aiseo-llmo/
- Indegene: “GEO vs AEO vs LLMO: The New Search Optimization Trinity” — https://www.indegene.com/what-we-think/reports/geo-vs-aeo-vs-llmo
- Tilipman Digital: “LLMO Guide 2026” — https://www.tilipmandigital.com/resource-center/articles/llmo-large-language-model-optimization-guide
- DigitalApplied: “LLMO Guide 2026” — https://www.digitalapplied.com/blog/llmo-guide-large-language-model-optimization-2026
- Search Engine Journal: “The State of AEO & GEO in 2026” — https://www.searchenginejournal.com/aeo-and-geo-in-2026/563856/
- Evergreen Media: “SEO Trends 2026” — https://www.evergreen.media/en/guide/seo-this-year/
Statistical Data
- Ahrefs (Feb. 2026): Top ranking results CTR -34.5%
- Seer Interactive (Sept. 2025): Organic CTR -61% for AI Overviews
- Gartner (2026): 25% of organic traffic shifting to AI [NOT VALIDATED — Gartner forecast, not measured data]
- First Page Sage: AI chatbot market shares 2026 — https://firstpagesage.com/reports/top-generative-ai-chatbots/
- Perplexity 370% annual growth — https://seoprofy.com/blog/perplexity-ai-statistics/
Created: 2026-03-09 | cc-research pipeline Status: Research document — some statistics are marked [NOT VALIDATED] because they are based solely on industry estimates or forecasts Project: LLMOFUTURE
Strategic Synthesis
- Map the key risk assumptions before scaling further.
- Monitor one outcome metric and one quality metric in parallel.
- Run a short feedback cycle: measure, refine, and re-prioritize based on evidence.
Next step
If you want your brand to be represented with context quality and citation strength in AI systems, start with a practical baseline and a priority sequence.