solanabr/startup-builder-skill/market-sizing
Claude Code and agent skill by solanabr/startup-builder-skill
- community
Checking that the SKILL.md file is still in the repository.
Size the market bottom-up with a source on every number, and rank entry wedges by problem quality × how much each builds the vision. Writes .claude/context/market.md. Use for "TAM", "market size", "market slide", "beachhead", "wedge", "where do we start".
Install
npx vskill@latest install solanabr/startup-builder-skill/market-sizingSKILL.md
Market sizing
Rank where to enter → count the units it covers → price them → check against a named top-down figure → write .claude/context/market.md. Output is one bottom-up number you can defend factor by factor and a ranked list of entry wedges. Asked for both flows, run wedges first, then size the top two or three, plus idea.md's wedge if it isn't among them: the slide is sized on that one.
What investors and filed decks ask of a market slide: method-sources.md. Traps: traps.md. Read it before adding anything up.
Context handoff
- At start, read whichever of these exist, and only ask what they don't answer:
.claude/context/market.md: a re-size. Re-open every source before reusing its number.idea.md(format): the founder's wedge, a candidate in the ranking.positioning.md(format): the buyer and the alternatives.pricing.md(format): the price and the floor cost.build.md(format): What works today sets
each wedge's distance D; Traction is customer evidence (Kamps: a valid alternative to a formula).
- The price ceiling is what the alternative costs the buyer. No context file stores it: work
it out with pricing's cost of the bad alternative, or ask.
- On completion, write
.claude/context/market.md(format) for
pitch-deck's market slide and fundraising's diligence answers. The slide is sized on idea.md's wedge, the one the deck pitches; if the top-ranked wedge differs, say so under Open decisions and leave idea.md to idea-sprint.
Rules for both flows
- Units before dollars. Count accounts and units from a dated source, then price them.
- Only what was opened counts. A snippet-only number is
secondary: never summed, never on a slide. - Every number is one line in the number-with-source
format: [inferred] for a reading, (est.) for a calculation written out on the same line.
- Pages for other people hold principles and the result. Process stays in the notes.
Flow A: sizing
- Name what the market is: the spend the product replaces, not an industry category.
Ceiling = what the alternative costs the buyer; floor = the marginal cost from pricing.md.
- Bottom-up first. Accounts × units per account per year × price; a monthly count is
converted on its own line, never with a loose × 12. With a bps price, the unit is the volume per account per year. The beachhead is the annual revenue at 100% share of the first market (Aulet).
- Top-down only checks, inside
market.mdand never on the slide: a named firm, its latest
edition, the same cut, within an order of magnitude of the bottom-up.
- Three kinds of money, never added together (money-types.md):
C, current spend on the check or work the product would do; S, spend on the activity it touches; R, money at risk, which goes on the problem slide. TVL, volume and market cap are Context, outside every sum.
- Bands. Beachhead US$20M to 100M a year (Aulet's 2012 draft, marked preliminary); vision
above US$1B with the most persuasive evidence (YC). Outside the band, run the reverse calculation and label it calibration, never a price.
- SOM = reachable accounts × units per account per year × price, never a share of TAM.
Expansion from a material improvement (Gurley) is a mechanism with no number.
- Declare what is weak: source concentration, data years, the token or FX price and its
date, cuts with no number, and the billing unit.
- Propose the slide the way pitch-deck's Market row
asks: one bottom-up number, the multiplication with each factor sourced, the source on each card. Why now is its own slide. The price and the number are the founder's call.
Flow B: wedges
- Vision, cycle and pillars, one line each (qxa-rubric.md).
- Filters before the search, from how the customer operates: their decision depends on
what the product observes or does; the gap costs more as they grow; solving it builds the vision. One real worked case each; the founder approves them before any search.
- Five blind fronts in parallel, one by analogy, searches capped, reports kept whole:
- Merge and filter before listing, then one card per candidate.
- Rank by Q×A, D only breaking ties (rubric).
- One blind review of the map, by a reader with none of the context: input, not orders.
- [Result page](references/result-page.md). A buyer conversation decides among the top few.
Onchain buyers
The account count is often public; read it instead of estimating it. Protocols: DefiLlama's Solana fee overview, protocols only, grouped by parent (worked count). Users: transacting wallets, deduplicated (traction checks), in humans or teams (crypto ICP). DAOs buy through proposals and grants, so their C is in past ones. Tokens convert at a dated price. Time or work saved instead of money: accelerates-x.md.
Output
.claude/context/market.md, committed.- A one-paragraph summary for the founder: the top wedge and why, the beachhead with its
calculation, the proposed slide number, and the open decisions.
- On request, the wedge result page as one self-contained HTML file.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/market-sizing/SKILL.md
Popularity
- Repository stars
- 2
- Repository forks
- 1
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/solanabr/startup-builder-skill/market-sizing)