market_research_sizing v1.0.0

name: "market_research_sizing" description: "Size a market for a product or business: TAM/SAM/SOM via top-down and bottom-up methods, buyer segmentation, competitive white-space mapping, and source grading on every load-bearing claim. Use when someone asks how big a market is, whether a business idea is worth pursuing, or needs numbers that hold up in a pitch."

Market Research Sizing

Purpose

Produce a defensible market sizing: TAM, SAM, SOM, segmented demand, and the competitive gaps worth entering. Every load-bearing number carries a named source and an evidence grade. The method triangulates independent estimates instead of trusting one.

Workflow

1. Define the boundary. Write one sentence: whose problem, what geography, what buyer type, and which product form count. Too broad (whole industry) or too narrow (current product shape only) both poison the number. Get this sentence agreed before any math.
2. Top-down. Start from a broad credible figure (analyst report, public spend data, population x penetration) and narrow with a chain of explicit assumptions: TAM -> SAM -> SOM. Each filter is a named assumption with a source.
3. Bottom-up (independent). Count buyers, not dollars. Countable proxies: company databases, community sizes, job postings, competitor customer counts, publication subscriber counts. Buyers x ARPU x purchase frequency. Do this without looking at the top-down number.
4. Triangulate. Compare. Within roughly 20-30%: treat the range as reasonably confident. Diverging by multiples: find the specific assumption driving the gap (usually penetration rate or ARPU), state which side is more defensible, and do not average them to hide the disagreement.
5. Segment the demand. Split by buyer need (jobs-to-be-done beats firmographics for explaining behavior; layer firmographics on after). For each segment: size, evidence grade, and the filter chain that produced it.
6. Map white space. For each demand segment, grade supply quality (price, product fit, availability). Plot segments on demand-strength vs supply-quality 2x2. The high-demand, low-supply quadrant is where a new entrant can win without fighting an entrenched incumbent head-on.
7. Stress test. State what must be true for the SOM to be bigger (funding, partnership, distribution you don't have yet). Never bake those into the base case.

Output Contract

• Boundary sentence, then TAM / SAM / SOM each with: number, method (top-down / bottom-up), filter chain or buyer math, and one source per claim.
• Evidence-grade table: every load-bearing claim graded strong / moderate / weak, with the weak ones flagged and their impact on the range stated.
• Triangulation result: the two estimates, the gap, and which assumption drives it.
• Buyer segments: need, size, grade, and what action each segment implies (pricing, message, or product decision).
• White-space map: the demand/supply 2x2 with quadrants labeled and the recommended where-to-play options ranked.
• No numbers without a source. No unsourced percentages, no "everyone knows" figures.

Operating Rules

• Bottom-up from countable buyers; never start from "x% of a $Y billion market" as the primary estimate.
• Evidence and assumptions are different things: label each. An assumption chained on three weak sources stays weak.
• Divergence between methods is signal, not noise. Reconcile explicitly or state which side wins and why.
• A segment is only useful if it drives an action (message, route, price, priority). Otherwise it is labeling, not strategy.
• Refresh before any fundraise. Note each source's date; market sizes move.
• Use bin/size_market.py to compute estimates and sensitivity ranges; hand math is error-prone and unverifiable.