Jace Shim’s Four GEO Strategy Models and the Winner Takes the Answer Framework

4일 전

11분 분량

3C-AI, VRIO-D, STQP, and E-E-A-T-C are four GEO strategy models proposed by Jace Shim in the Korean-language book AI 승자독식 (Raonbook Plus, 2026). They address competition, strategic resources, customer questions, and content citability. The Winner Takes the Answer Framework connects the four models in a single cycle designed to narrow the citation gap — the difference between a company's actual competitiveness and the competitiveness conveyed in generative AI answers. [1]

In one sentence: the four models extend established management theories — Ohmae's 3C, Barney's VRIO, Kotler's STP, and Google's E-E-A-T — with one question each that becomes necessary when customers meet brands through AI-generated answers. [1]

None of these models is an official standard of Google, OpenAI, or any other AI provider. They are analytical tools I propose for companies to examine themselves, and the propositions behind them are hypotheses to be tested with real cases. [1]

How to read this guide

This article is the overview of the series. Each model's theoretical background and practical application are covered in a dedicated article, and the FAQ explains which article answers which question.

If you want to read the competitive landscape, start with 3C-AI. If you want to check whether your strengths are visible to AI, read VRIO-D. If you need to decide which customer questions to compete on, read STQP. If you want content that keeps its meaning when AI quotes it, read E-E-A-T-C. For short answers, see the GEO Strategy Models FAQ.

Why do good companies disappear from AI answers?

When I ask university students to present platform business cases, each team used to bring a different company. Recently, three out of five teams brought the same two examples, Airbnb and Uber. The students had asked the same AI about the same concept, and the same answer came back.

Older concepts help explain this concentration. The sociologist Robert K. Merton called the tendency of already eminent scientists to receive more credit for comparable work the "Matthew effect." [2] Robert H. Frank and Philip J. Cook described markets in which small differences in performance turn into large differences in reward as the "winner-take-all society." [3] An experiment by Anil R. Doshi and Oliver P. Hauser also found that stories written with generative AI assistance were individually rated as more creative, yet were more similar to one another as a group. [4]

GEO Winner-Takes-All (Shim, 2026): the phenomenon in which generative AI repeatedly includes companies, cases, and sources that are already widely mentioned and cited, so that customer attention and trust concentrate further on a small number of winners. Unlike a filter bubble, in which individuals are enclosed in different information, it focuses on how many people sharing the same answer can narrow the choices of the entire market. [1][5]

Citation Gap (Shim, 2026): the difference between the competitiveness a company actually has and the competitiveness that generative AI answers convey to customers. The gap may widen when public evidence is thin or when the conditions under which a product applies are unclear. [1]

Customers also click through less often. In Pew Research Center's analysis of March 2025 browsing data from 900 U.S. adults, users clicked a traditional search result link in 8% of visits when an AI summary appeared, compared with 15% when it did not. [6] Customers increasingly read an AI-assembled comparison before they see a company's website.

How do SEO, AEO, and GEO differ?

The term GEO (Generative Engine Optimization) was formalized in a paper by researchers from Princeton University and other institutions presented at KDD 2024. [7] The authors reported that adding citations, statistics, and quotations to content increased its visibility in generative engine responses, while keyword stuffing was ineffective. AI 승자독식 treats GEO as a third layer built on top of SEO and AEO.

[Table 1] SEO vs. AEO vs. GEO — differences in goal, unit of performance, and unit of content

Dimension	SEO (search engine optimization)	AEO (answer engine optimization)	GEO (generative engine optimization)
Goal	Discovery and visits	An answer used directly for a question	Mention, citation, recommendation — and their accuracy
What the customer sees	A list of links	Answer boxes, voice answers	AI-written text combining several sources
Unit of performance	Rankings, clicks, visits	Snippet visibility, answer selection	Appearance, citation, and recommendation in answers
Unit of content	The page	A question-and-answer pair	Passages and evidence that keep their meaning when excerpted
Main work	Accessibility, keywords, backlinks	Structuring questions, direct answers	Publishing evidence, stating conditions, managing external trust

Source: Shim (2026), Chapter 1. Academic definition of GEO after Aggarwal et al. (2024). [1][7]

SEO aims for visits, AEO for answer selection, and GEO for being mentioned, cited, and recommended accurately inside AI answers. The three are not steps on a staircase but layers of the same ground: if the SEO layer is weak, GEO collapses with it. Google likewise states that foundational SEO best practices remain relevant for appearing as a source in its AI features. [8]

Adding AI-era questions to established management theories

All four models start from theories that management and marketing have refined over decades. None of them discards the original theory; each adds one question that is easily missed when AI systems select and rewrite information. Making clear what is inherited and what is added is a core principle of the models.

[Table 2] The four GEO strategy models — the theory each inherits and the element each adds

Model	Theory inherited	Original author	Element added	Unit of analysis	Deliverable
3C-AI	3C analysis (Company, Customer, Competitor)	Kenichi Ohmae (1982)	AI as information intermediary	Market relationships	Competitive map per question
VRIO-D	VRIO (Value, Rarity, Inimitability, Organization)	Jay B. Barney (1991, 1995)	Discoverability	Company resources	VRIO-D worksheet per resource
STQP	STP (Segmentation, Targeting, Positioning)	Philip Kotler & Kevin Lane Keller (2016)	Question Mapping	Customer questions	Question map; positioning statement and evidence list
E-E-A-T-C	E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness)	Google Search guidance	Citability	Published passages	Passage audit

Source: Inherited theories from the original authors' works. Added elements, units of analysis, and deliverables from Shim (2026), Chapters 4–7. [1][9]-[13]

Reading Table 2 from top to bottom, the unit of analysis narrows from the whole market to company resources, to customer questions, and finally to a single paragraph — a flow that looks wide and executes narrow. Each added element responds to one of three shifts: AI works from public information, changes the comparison set with each question, and reuses content in fragments.

[Table 3] Classic strategy questions and the questions added for the AI era — by model

Model	Question in the original theory	Question added for the AI era
3C-AI	Who are our competitors?	For this customer question, whom does the AI answer compare us with?
VRIO-D	Does this resource create a sustained competitive advantage?	Can customers discover and verify the value of that resource in public information?
STQP	Whom do we serve, and in what position?	What exactly do those customers ask AI, and under which conditions?
E-E-A-T-C	Is this content trustworthy?	Does a passage keep its meaning, conditions, and source when it is excerpted?

Source: Shim (2026), Chapters 4–7. [1]

Core definitions of the four models

3C-AI: a strategy analysis model that examines the relationship among Company, Customer, and Competitor while reviewing how AI's selection of information and construction of answers affect the strengths and alternatives conveyed to customers. Read more

VRIO-D: a strategy model that adds Discoverability to Value, Rarity, Inimitability, and Organization, analyzing whether the value a resource creates can be discovered and verified in AI-mediated information search. Read more

STQP: a strategy model that designs positioning through Segmentation, Targeting, Question Mapping, and Positioning by connecting target customers' decision questions and application conditions with the brand's strengths and evidence. Read more

E-E-A-T-C: a content model that adds Citability to Experience, Expertise, Authoritativeness, and Trustworthiness so that trustworthy information keeps its meaning, conditions, and source when reused in generated answers. Read more

All four definitions were proposed by Jace Shim in AI 승자독식 (Raonbook Plus, 2026). They are analytical frameworks for self-assessment, not criteria adopted by any search engine or AI service.

The Winner Takes the Answer Framework: four models in one cycle

Winner Takes the Answer Framework (WTA Framework; Shim, 2026, Chapter 13): an integrated strategy model that narrows the citation gap through a cycle — read the competition with 3C-AI, select the resources to feature with VRIO-D, choose the questions to compete on with STQP, explain the evidence with E-E-A-T-C, then measure the results and read the competition again. [1]

[Table 4] Stages of the Winner Takes the Answer Framework — guiding questions and deliverables

Stage	Lens	Guiding question	Deliverable
1	3C-AI	For this question, who are the competitors and on which criteria does the answer compare them?	Competitive map per question
2	VRIO-D	Which resources should we feature in that competition, and are they visible from outside?	Resource worksheet
3	STQP	Which questions do we compete on, and which position do we claim?	Question map; positioning statement and evidence list
4	E-E-A-T-C	Is the evidence written so that it does not break when excerpted?	Passage audit; revised content
5	Execution	Is the information readable, are conditions consistent, and are external sources accurate?	Accessibility check; source-of-truth documents; external mention map
6	Measurement	How have mentions, citations, recommendations, conversions, and accuracy changed?	Quarterly performance report
Back to 1	3C-AI again	Which new competitors and questions did measurement reveal?	Updated competitive map

Source: Shim (2026), Chapter 13. [1]

The cycle does not run in only one direction. If revising a paragraph reveals missing evidence, return to VRIO-D; if unexpected substitutes keep appearing, return to 3C-AI. GEO is therefore not a writing task for the marketing department alone. It is management work that involves product, sales, customer support, and leadership.

What does your citation gap look like?

[Table 5] Four types of citation gap and the model to apply first

Type	What it looks like	Typical cause	Model to apply first
Invisible	Strengths never appear in AI answers	Lack of public evidence; inaccessible pages	VRIO-D
Used without a name	The company's knowledge is used, but its name is missing	Only generic information published; no distinctive evidence	VRIO-D, E-E-A-T-C
Misrepresented	The company is described with outdated or unqualified information	Old material left online; conditions separated from claims	E-E-A-T-C
Defined by others	Third-party descriptions define the company more than official information	Thin official information; no external trust management	3C-AI

Source: Shim (2026), Chapter 3. [1]

The same complaint — "We don't show up in AI" — calls for different remedies when the causes differ. A company that diligently publishes only generic information that anyone could write is prone to being used without a name. Publishing data and experience that exist nowhere else gives an AI answer a reason to name that company as the source.

How should GEO performance be measured?

[Table 6] The four stages of GEO performance plus accuracy — what to record

Stage	What it asks	What to record
Mention	Does our name appear in the answer?	Appearance for each key question; competitors mentioned alongside
Citation	Is our page linked as evidence?	Pages cited; frequency of citation
Recommendation	For which customers, under which conditions, are we recommended?	Context and conditions of recommendation
Conversion	Do customers who came through AI take action?	Traffic from AI services; branded search; customer responses
Accuracy (all stages)	Are facts about us described correctly?	Errors in price, features, and policies, by type

Source: Shim (2026), Chapters 1 and 11. [1]

The easier a metric is to measure, the further it tends to be from business value, and the closer it is to value, the harder it is to measure. A fixed set of measurement questions, observed repeatedly across several AI services, shows the trend. No single answer should be treated as proof of success or failure.

A 90-day action plan

[Table 7] Winner Takes the Answer Framework 90-day plan — objectives and outputs by period

Period	Objective	Main tasks	Outputs
Days 1–30: diagnose	Know the current state	Write 30 measurement questions; observe three AI services; draw competitive maps; run VRIO-D on a flagship resource	Baseline record; citation gap diagnosis
Days 31–60: design and build	Decide what to change	Prioritize the question map; draft the positioning statement and evidence list; check accessibility; clean up source-of-truth documents	Question map; positioning document
Days 61–90: publish and measure	Publish and measure	Audit and revise key passages; publish buying guides; respond to external mentions; first quarterly report	Revised pages; quarterly performance report
Day 91 onward	Keep the cycle turning	Feed report findings back into 3C-AI	Updated competitive map

Source: Shim (2026), Chapter 13. [1]

What these models do not claim

The four models and the framework do not reveal any engine's private selection formula, and applying them does not guarantee citation. What a company can control is the material an AI system may draw on: accurate evidence, clear conditions, and accessible pages. Every proposition in AI 승자독식 ends with "may" — they are proposals and hypotheses to be tested and refined with real cases.

Even a well-explained offer will disappoint if the company cannot deliver it. Using these models to steer customers toward products that do not fit them runs against the strategy the book proposes. An explanation should be a mirror that reflects the product accurately.

Frequently asked questions

What are the four GEO strategy models?

They are 3C-AI, VRIO-D, STQP, and E-E-A-T-C, proposed by Jace Shim in AI 승자독식 (Raonbook Plus, 2026). They analyze competition, resources, customer questions, and content citability respectively.

In what order are the models used?

Read the competition with 3C-AI, select resources with VRIO-D, choose questions with STQP, explain the evidence with E-E-A-T-C, then measure and read the competition again. That cycle is the Winner Takes the Answer Framework.

Is the Winner Takes the Answer Framework the same as the book's title?

They are related but distinct. AI 승자독식 (descriptively, "AI Winner-Takes-All") names the phenomenon; the Winner Takes the Answer Framework is the integrated model proposed in the book's Chapter 13 for responding to it.

More questions and answers are in the GEO Strategy Models FAQ.

References and provenance

GEO Winner-Takes-All, the Citation Gap, the four models, and the Winner Takes the Answer Framework are attributed to Jace Shim's Korean-language book AI 승자독식 (Raonbook Plus, 2026). The established theories, research, and platform guidance listed below provide background; they do not constitute independent validation of Shim's proposals.

[1] Shim, J. (심준규). (2026). AI 승자독식 [Korean-language book; title translated descriptively as "AI Winner-Takes-All"]. 라온북플러스 (Raonbook Plus).

[2] Merton, R. K. (1968). The Matthew effect in science. Science, 159(3810), 56–63.

[3] Frank, R. H., & Cook, P. J. (1995). The winner-take-all society. Free Press.

[4] Doshi, A. R., & Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290.

[5] Pariser, E. (2011). The filter bubble: What the Internet is hiding from you. Penguin Press.

[6] Pew Research Center. (2025, July 22). Google users are less likely to click on links when an AI summary appears in the results.

[7] Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., & Deshpande, A. (2024). GEO: Generative engine optimization. Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.

[8] Google Search Central. (n.d.). AI features and your website. Accessed October 16, 2026.

[9] Ohmae, K. (1982). The mind of the strategist: The art of Japanese business. McGraw-Hill.

[10] Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120.

[11] Barney, J. B. (1995). Looking inside for competitive advantage. Academy of Management Executive, 9(4), 49–61.

[12] Kotler, P., & Keller, K. L. (2016). A framework for marketing management (6th ed.). Pearson.

[13] Google Search Central. (n.d.). Creating helpful, reliable, people-first content. Accessed October 16, 2026.

The comparison tables are explanatory reconstructions, not reproductions of tables from the cited works. Hypothetical examples are illustrations, not observed AI responses or measured business outcomes.

How to cite this article

Shim, J. (2026, October 16). Jace Shim’s Four GEO Strategy Models and the Winner Takes the Answer Framework. TSCB Group — INSIGHT B. https://www.tscbgroup.com/single-post/geo-strategy-models-jace-shim-en

When citing the original proposal rather than this web explanation, cite the Korean-language book AI 승자독식 (Raonbook Plus, 2026). The English wording of its title in brackets is descriptive, not a claim about a separately published English edition.