Size the market from the outside in.
Name a product and its constraints. Your assistant estimates TAM top-down and bottom-up, narrows it to SAM and SOM with stated reasons, projects 2–3 years ahead and numbers every assumption with a confidence level.
- TAM~$110Mper year, range $75–180M, two methods reconciledA1–A3
- SAM~$19M~600 qualifying offices, ~17% of TAMA4
- SOM$1.26Mrun-rate today; $3.2M at the 100-office goalA5, A7
From one saved run on “Tidewell Desk”, an invented Portland lunch-kit startup, with no web search. A1–A8 are the run’s numbered assumptions; six of eight are rated low confidence, so read these as order-of-magnitude estimates.
What it is
A single instruction file that makes a capable AI assistant act as a strategic market analyst. It defines the market boundary first, then estimates the Total, Serviceable and Obtainable markets, and cross-checks the top-down number with a bottom-up build of customers × price × frequency.
Every estimate comes with its reasoning, a summary table of today versus 2–3 years out, growth drivers, and a numbered list of assumptions with confidence, a way to validate each one and what would change the answer.
Who it is for
- Founders preparing an investor pitch or board discussion
- Product managers weighing a new segment or market entry
- Strategy and BD teams sizing an opportunity before committing budget
- Students and analysts learning TAM / SAM / SOM with worked reasoning
Works well for
- Investor decks that need a defensible market slide
- Go / no-go on a new geography or segment
- Turning your own traction into a market view
- Finding which assumption moves the number most
Not designed for
- Audited figures or a substitute for paid market research
- Revenue forecasts or valuations you can bank on
- Sourced numbers when no search tool or data is available
- Financial, legal or investment advice
What you provide, what you receive
You provide
- The product and the problem it solves
- Constraints: geography, vertical, customer type
- Pricing and current traction
- Any market research, reports or competitor data
You receive
- Market definition and scoping decisions
- TAM top-down, bottom-up and reconciled
- SAM as % of TAM, SOM as % of SAM, with reasons
- Summary table: today and 2–3 years
- Growth drivers; numbered assumptions with confidence
Where it runs
Any AI assistant that can read a Skill file, including Claude Code and Codex. Download the free plaintext package from Finch and load it with your assistant’s Skill method, or ask the assistant to read SKILL.md.
No extra API key, script or service. You need access to an assistant model, which may have its own cost. The instructions tell the assistant to use web search and cite sources; that only happens if your assistant has a search tool. Without one, it should label numbers as assumptions, as in the saved run.
Limits and licence
Market sizes are estimates. Their quality depends on the data supplied or found; check every source and assumption before a number reaches a deck.
MIT licence, Copyright (c) 2026 Pawel Huryn. Source: phuryn/pm-skills at commit 2a12295a7717, path pm-market-research/skills/market-sizing. The package adds a README, NOTICE, provenance note and a cover; the instructions are unchanged.
How it was checked
One controlled run on a fictional company (Claude Code only, no web search) produced all nine required parts, gave both top-down and bottom-up estimates, and labelled every market-level count as an assumption. Because no sources were available, it recommended data sources instead of citing any. The published package was downloaded back from Finch and matched the audited files. Web-search behaviour, native installation and real business outcomes were not tested. See the Benchmark tab.