S2: Make information usable together

English edition of S2 — the capstone of the first block. S3–S9 are published in Japanese and are being translated. With S1 and S2 in English, you have a complete, self-contained starting path.

Choose where you raise your voice, yourself. Do not burn out in a place with no demand.

On this page

• What you get
• Why it matters
• The core: breaking your cognitive frame
• The method (5 steps)
• Step 1: put yourself and your company into words
• Step 2: turn media into structured knowledge
• Step 3: market research through your Ontology
• Step 4: integrate across sources, find the gap
• Step 5: decide the market, write the action plan
• Capstone: build a mid-term business plan

What you get from this skill

Two things:

1. A base of outside-world information, taken in and structured (research/) — information from the web, books, videos, competitor sites, kept not as bookmarks and screenshots but as Markdown files, organized into a form AI can read.
2. A market and target chosen with data — your skills redefined outside the frame of your job title, and the market where money actually flows and the people you will aim at, identified. “Where, to whom, with what you fight” is decided.

In S1 you made a map of yourself. In S2 you make the outside-world map that connects to it. And in the capstone exercise you integrate everything into a single mid-term business plan — S2 does not stop at choosing a market; it draws the full three-year design.

Why it matters

People who fail at market selection usually trip on one of two things.

① Information is scattered in an unusable form. The article you liked is in a browser tab, the competitor screenshot is in your phone’s camera roll, the book flag is on your shelf. Collected this way, information — however much of it — cannot be used for judgment. Ask AI “what do you think of this market?” and, with no structured data in hand, AI can only return generalities.

② You choose the market subjectively. Choosing by “seems like demand” or “because I like it” lands you in “I tried it and nobody paid.” A market is chosen from the data of where money flows.

S2 solves both at once: take in information from every medium and turn it into structured MD (make information usable), and choose the market from that data (do not choose subjectively). Making information “usable together” is the foundation for choosing a market correctly.

The core: breaking your cognitive frame

Before choosing a market, grasp the essence: break your own cognitive frame and reach outside your perception.

When you have a specialty, you unconsciously think along the “socially given shape” of that field:

• A restaurant owner → “how do I make my restaurant popular?”
• A lawyer → “which firm do I join / how do I run my office?”
• A designer → “which client do I get work from?”

As long as you think this way, you can only move inside the role that social norms defined. But the same restaurant owner who reframes as “I have the power to design food experiences” can re-choose a main battlefield where that power is most alive — restaurant consulting, regional food branding — which never appears from inside “how do I make my restaurant popular?” It is what you see when you look at the market from outside your perception.

The Ontology you built in S1 becomes a weapon here. Lay your expertise, experience, and values back out as concepts, stripped of the job-title frame, and “what you can actually do” comes into a completely different resolution.

[Before: seen through the job frame]
Restaurant owner → market = the restaurant industry

[After: conceptualized through the Ontology]
"can design food experiences"
"can create places where people gather"
"can put the prep process into words"
"deeply connected to the local community"
  ↓
the field of markets you can fight in widens
(food-business consulting / online food courses / regional branding …)
  ↓
choose ONE you can win as your main battlefield  ← the goal of S2

This is what it means to expand perception itself. In S2 you choose one main battlefield from the widened field. Attacking several markets at once (multi-front expansion) is the theme of S4. Whether the chosen market is real, you verify by taking in outside information with the method below.

The method: take in, structure, decide

S2 runs in five steps. Step 1 takes in yourself and your company; Steps 2–4 take in and structure the outside world; Step 5 decides.

The new folder for this skill is research/, added to your S1 myproject.

myproject/
  ├── inbox/
  ├── ontology/  (concepts, people, organizations, relations)
  ├── actionlog/
  └── research/          ← added in S2: where outside information lands
        ├── sources/         primary material taken in (summarized MD of articles, books, videos)
        ├── market/          market-size and trend research MD
        ├── competitors/     competitor players (1 company = 1 file)
        └── analysis/        gap analysis, market choice, entry strategy, action plan MD

---

Step 1: put yourself and your company into words

You put two things into words: ① yourself (the individual) and ② the company you belong to. The most important part is to draw a clear line between them in the Ontology.

Why separate them — the skills you hold and the assets/brand the company holds are different things. Mixed together, AI cannot tell where your own power ends and the company’s begins. You can carry yourself to another company or into independence; the company you cannot. So they are separate files from the start.

Gather your own information into one MD (people/your-name (me).md), then the company’s into one MD (organizations/company-name (my company).md). That they live in different places — you in people/, the company in organizations/ — is itself the line.

Tip: name files with real names, not “me” / “my company.” As the Ontology grows and dozens of people and organizations line up, an agent running a systematic analysis easily loses track of “who was the owner of this map again?” Fix it with a real name plus the (me) / (my company) mark, and the subject never drifts no matter how deep the analysis goes.

① Gather yourself. First, gather the “you seen from outside” — if you do business under your real name, your information is already scattered online (company site, interviews, talks, SNS, reviews). Your inner self-image and how you look from outside are usually misaligned — and that gap is a market hint. (“What I thought was obvious is described outside as a strength.”) This is also a rehearsal for Step 2’s “turn information into structured knowledge,” done on the nearest subject: you.

Search the web for "(my name / company name)" and compile the me-seen-from-outside
into structured knowledge.
- how I'm introduced as a person / business
- what's written as my strength / expertise
- the achievements, career, involvements that appear
- always attach the source (URL)
→ save to research/sources/

Then the inside. Before skills, start from conviction — the vision of your mid-term plan sources here. Do not tell AI “make me a vision” (you get generalities). The real conviction is already spilled in your S1 inbox/ as daily voice memos. Pick it up from there.

Read all the voice memos / notes in inbox/ and extract my "conviction as a business owner / life design."
- what I want to realize with this business
- what I want the company to be in 3 years (grow / hold / succeed / sell)
- how long I want to stay involved
- what I want to protect in life (time, family, money, connection to society)
→ reflect into people/your-name (me).md under "conviction / life design"

Then add skills and experience. Your own skills are hard to see from your own viewpoint — what you think is “obvious” is often something others simply cannot do. Talk to AI concretely, cross-check with the externally gathered information, and have it organized objectively. Include: your career (years, what you did), what you can do (with numbers/examples), what you dislike, qualifications and unique methods, achievements you can state in numbers, and the resources you can put into the business (time available, self-funds, risk tolerance).

Tip: have AI throw back questions that doubt the “obvious.” “Of what I just said, which can almost no one do?” “Which would surprise someone outside the industry?” — the biggest differentiator is often what you did not count as a strength.

Then ask AI: “list, broadly, the industries and roles that need this skill,” “widen it beyond my own industry,” “here are acquaintances I can think of — help me consider who this might land with.” (Building the full people database is the theme of S6; here, naming a few aloud is enough.) Finally, have AI consolidate conviction, outside-view, self-reported skills, and who-it-lands-with into one file: people/your-name (me).md — your portable self, the origin of all research that follows.

② Gather your company. Consolidate what the company holds as a unit, separate from you. Collect not only what S2 needs now but what later skills will use — the Ontology is grown, not filled in at once. Company items include: philosophy/vision, business & products, customers & their problems, revenue composition and dependence, financials, size/organization/key people, work process and bottlenecks, assets/strengths (equipment, brand, clients, location, licenses), relationships (clients, suppliers, partners), and management issues/risks. External information (site, registry, interviews, reviews) works here too.

Tip: do not mix “me” and “the company” — this is the crux of the line. “10 years on the floor” is you (portable if you leave); “prime location by the station” is the company (gone if you leave). Separated, Step 3’s 3C analysis can evaluate Company from the company file and your market value from the individual file, separately. Mixed, AI misjudges your true individual strength.

These two files — your-name (me).md and company-name (my company).md — are the foundation referenced in every following step.

---

Step 2: turn any medium into structured knowledge

This is S2’s core skill. Convert the outside world — web, books, videos, competitor sites, images — into “structured knowledge” MD and take it into your Ontology. Bookmarks and screenshots cannot be used later; only when converted into a structure AI can read for judgment does information become “usable.”

What “structured knowledge” is: not a mere summary. An MD with four things: ① source (who, when, where) ② key points ③ implication for me ④ links to related concepts. A human note is “to remember later”; structured knowledge is “for AI to use across sources” — so source and links are its lifeblood.

Do structuring and Ontology-connection in one instruction. Splitting the steps breeds hesitation in practice. Hand over the information (image / file / URL) and, in one instruction, do “save the structured-knowledge MD” and “save the Ontology link” together.

(hand over an image / file / URL, then:)
Turn this into structured knowledge.
1. read it and save it to research/sources/ as a structured MD
   (source, capture date, summary, implication-for-me, quotes — as a set)
2. reflect related concepts into ontology/concepts/ (new file if it's a new point)
3. append the connection to ontology/relations/links.json

The best input in S2 is a book — more organized than web articles, a mass of primary information the author verified systematically. Scan the pages with Google Drive’s mobile “scan” (less distortion than a photo, better text extraction), then run the one instruction above on the synced file.

Tip: always verify AI’s summary against the primary source. AI can summarize a URL but may mix in stale knowledge or guesses, written plausibly. Check numbers and facts against the original; mark the unverifiable as “unconfirmed.” A source-less MD is garbage inside an Ontology.

Tip: do not separate “collecting” from “structuring.” The moment right after reading a source is when you grasp its implication best. “Collect 50 interesting articles, structure later” piles up unread MDs and doubles the cost. Narrow to a few good sources per theme; take one in, and wire its concept links all the way — the same principle as S1’s “don’t separate collecting from reading.”

---

Step 3: market research through your Ontology

S2’s market research is not only about finding new markets to enter. For most people it is optimizing the main business you already run along the axis of “where, to whom, with what you fight.” See it with data, not subjectively.

Here the Ontology pays off. Having AI read your your-name (me).md and company-name (my company).md alongside every piece of research — that is what “doing market research through the Ontology” means. The same “research the ○○ market,” fed your own data, returns not generalities but answers tied to your and your company’s actual situation — down to “which of your strengths works where in this money flow.”

Run research in two stages: gather base data, then read the structure with frameworks.

Base data (results accumulate in research/market/ and competitors/): market size, competitors, regional players, trends, technology, money flow, subsidies/programs, cost trends, hiring/wages, financing, benchmarks, regulation/tax. Each request adds a file and raises the resolution.

Tip: limiting the area raises accuracy. Ask only “the ○○ market” and you get national generalities. SMEs’ trade areas are local, so add geography — “in the Yokohama area,” “in the Colombo area” — and competitors, money flow, wage rates, and (locally specific) subsidies snap toward your reality.

Tip: keep competitors one-file-per-company. In competitors/, one file each with the same fields (services, price, target, strengths, weaknesses) makes them easy to turn into a comparison table in Step 4.

Frameworks. Once base data has accumulated, run the classic frameworks. Every framework needs a “Company” view — which you can fill instantly from your company/individual files. Fed research/ and ontology/, they return analysis tied to your business, not textbook generalities.

Framework	What it reveals
3C	from market, competitor, company — the position you can win
PEST	how macro change (political, economic, social, technological) hits your business
Five Forces	whether the industry is “structurally profitable at all”
TAM / SAM / SOM	the size of the market and the share you can realistically take

After reading research/ and ontology/organizations/company-name (my company).md and
ontology/people/your-name (me).md, do a 3C analysis of "the ○○ market."
- Customer: who they are and what troubles them
- Competitor: strengths/weaknesses of each in competitors/
- Company: the company's strengths (company file) and my individual strengths (me file), separately
Cross the three and propose the position where I can win.
→ save to research/analysis/3C.md

Tip: a framework is not there “to be filled.” Filling the 3C or PEST boxes and feeling satisfied is the classic failure. Always converge to one conclusion at the end: “so, where do I fight?” Ask AI: “from this analysis, name the one move I should make.”

Tip: if you have a running business, you can get primary information. Secondary web data is not the only research. Ask a few existing customers, suppliers, or floor staff “what have you struggled with lately? what did you pay for recently?”, turn their words into structured knowledge in research/sources/, and your frameworks jump a level above secondary-only.

---

Step 4: integrate across sources and find the gap

In Step 2, each source was structured and connected to concepts at intake. Step 4 is not “structure now” — it is laying the already-structured pieces across into one map: comparison, mapping, gap-finding, which only becomes possible once several sources are in.

Read the whole research/ folder and compile into one file (save to research/analysis/cross-integration.md):
1. a competitor comparison table (each company in competitors/ by price, target, value)
2. a market map (which layer, who sells, at what price)
3. candidate gaps where money moves but no one has filled it yet

Established categories always have competitors; entering means price wars and low, stable returns. Aim at the “gap” in an established category. Analysis lenses: what has no one done yet? which customer layer are incumbents failing? what is still done manually that AI lets one person do? where is the trend heading into a skills shortage?

Example: food × AI
[Established (hard to enter)]  lunch service, delivery → crowded, thin margins
[Gap (differentiable with AI)]
  AI-adoption consulting for restaurants → very few "knows the floor × can use AI" people
  productizing prep operations → repeatable, decoupled from time
  food-experience branding → local restaurants lack SNS know-how

Tip: add the cross-source connections newly visible here to relations/. The new thing in integration is relations that span sources — “competitor-A’s weakness is exactly the flip of my strength,” “several sources point at the same gap from different angles.” Have AI add these to relations/links.json; the map grows more precise. (Connections to acquaintances deepen further once S6 fills the people database.)

---

Step 5: decide the market and write the action plan

With yourself and your company in Step 1, and the outside world in Steps 2–4, the last move is to decide on one point, concentrate resources, and draw the plan to take it.

Three patterns for choosing: create an original market (define your own category — no competitors, slow recognition, monopoly if it works), attack a niche (a small but certain demand, easy to enter, differentiate by depth), or catch the gap in a trend (pick up spillover from a growing market, demand already proven). Which is right depends on the situation. What matters is that two conditions hold: your skill is maximally alive and money is flowing sufficiently. The data from Steps 1 and 3 is the basis for that judgment.

Once the market is set, define the target concretely (e.g., “owner-run restaurants in the Yokohama/Kawasaki area, ¥5–20M revenue, interested in SNS but unsure how, run by 3 or fewer staff”). Write the decided market, target, and entry strategy into one MD: research/analysis/entry-strategy.md — the flag planted where you will stand. Keep the reason (which data was decisive) so you can retrace and update the judgment when the environment changes.

Then the action plan — market and target alone do not move you. Have AI read the whole Ontology and research/ and drop the strategy into concrete action and milestones:

After reading research/analysis/entry-strategy.md and all of ontology/ and research/,
make an action plan to take "the ○○ market."
- the first move (something you can start right now)
- what to do in 90 days / 6 months / 1 year
- business milestones (customer count, revenue, deals)
- resources needed at each stage (time, money, people) and bottlenecks
- optimistic / pessimistic scenarios and the likely sticking points
→ save to research/analysis/action-plan.md

S2’s Ontology has three centers: cross-integration.md (the outside map) / entry-strategy.md (where you fight) / action-plan.md (how you move). When the market moves, these three are what you update first. The mid-term plan in the next exercise is the “final work” that binds these three — but what lives and grows daily is these three analysis files.

---

Capstone: build a mid-term business plan with your Ontology

S2’s finish. Using everything accumulated in the Ontology, build one mid-term business plan — the design of where and how you take the company over the next three years, the layer between “the 10-year vision” and “this year’s budget.”

Normally this is heavy labor (environment analysis, self-analysis, stacking numbers — millions of yen to a consultant, or months on your own). But having done S2, all the material is already in your Ontology. Each part of the plan maps to an S2 output:

Plan part	Ontology / research used
Current state (internal)	your-name (me).md · company-name (my company).md
Current state (external)	research/analysis/cross-integration.md (+ market/, competitors/, PEST, Five Forces)
Strategy	research/analysis/3C.md · entry-strategy.md
Priority measures / milestones	research/analysis/action-plan.md
Targets / resources	company revenue scale + market size (TAM/SAM/SOM)

Building the plan is integrating the separately-made structured knowledge into one story — not writing from zero, but having AI assemble existing material.

The four steps: (1) have AI assemble a draft from all of ontology/ and research/, citing the source file for each part; (2) read it in full and rewrite it with your own will — an AI-made plan, however tidy, is otherwise someone else’s essay; put in “this is how I fight these three years”; (3) tighten the numeric targets against market size and current state (back-calculate from current revenue and the market’s SOM, in optimistic/standard/pessimistic, decomposed into “count × unit price”); (4) tie it to the action log so it stays a living plan — reconcile monthly with actionlog/ and have AI analyze the gap and update.

Tip: the mid-term plan is both “the culmination of S2” and “the entrance to the next.” It is the core deliverable of Milestone 1, and each part is deepened in later skills — financials in S5/S8, organization and hiring in S6/S7. Not finished at S2; it is the trunk grown across the whole framework.

---

Are you done?

• Consolidated yourself (outside view + skills + resources + conviction) into people/your-name (me).md
• Consolidated the company into organizations/company-name (my company).md, drawing the line from the individual
• Took in web/book/video/competitor/public sources as structured MD in research/sources/, with source and date
• Ran 3C / PEST / Five Forces / market size tied to your business
• Cross-integrated research/ into a comparison table, market map, and gap in research/analysis/cross-integration.md
• Narrowed to one market and a concrete target; wrote entry-strategy.md and action-plan.md
• Assembled a mid-term-business-plan.md and rewrote it in full with your own will

Related skills

• Prerequisite: S1 — Build your own information system
• Next (currently Japanese): S3 — Control your own distribution

---

This is the English edition of S1’s sequel, S2. With S1 and S2 you have the complete first block of SOVREN Framework. The remaining skills (S3–S9) and three milestones are published in Japanese at sovren-framework.pages.dev. We run this method in weekly live classes in Yokohama and publish a report after each session.