Market Sizing

Two independent methods, one triangulated answer, and every assumption on the page.

What this skill is

A workflow for estimating Total Addressable Market (TAM), Serviceable Addressable Market (SAM), and Serviceable Obtainable Market (SOM) with a dual methodology. A top-down estimate narrows an industry total to the relevant segment; a bottom-up estimate multiplies out from customer counts and unit economics. Comparing the two exposes weak assumptions, and an assumption log makes the result defensible under questioning.

What it solves

• A single headline market number pulled from one analyst report with no path to the company's segment
• TAM presented as if it were the revenue the company could capture
• Bottom-up and top-down numbers that disagree by an order of magnitude with no explanation
• Optimistic estimates that lose credibility with investors and boards
• Assumptions buried in a spreadsheet that nobody can audit or update
• Market boundaries that change between slides, so the numbers cannot be compared

When to invoke

• Building the business case for a new product or product line
• Evaluating expansion into a new customer segment, use case, or geography
• Preparing an investor pitch, board memo, or annual plan that needs a market size
• Reviewing a market size produced by a team, consultant, or startup
• Deciding which of several segments to enter first

Phase 1: Define market boundaries

A market size is only as good as its boundary. State each explicitly:

Then define the three layers for this boundary:

• TAM - total annual spend on this job if every possible customer bought
• SAM - the part of TAM the product can serve with its current offering, channels, and geography
• SOM - the part of SAM the company can realistically win in a stated time frame

Phase 2: Top-down analysis

Start from a published total and narrow it with explicit filters.

TAM = industry spend on the category
SAM = TAM x segment share x geography share x use-case share
SOM = SAM x achievable market share in the time frame

Sources: industry reports, public company filings, government statistics, trade association data. Record the report, year, and definition used, because category definitions differ between sources.

Phase 3: Bottom-up analysis

Build from countable units.

TAM = number of potential customers x annual contract value
SAM = reachable customers x annual contract value
SOM = reachable customers x expected win rate x annual contract value

Phase 4: Triangulate

Compare the two methods at each layer.

• Within about 30% - high confidence; report a range spanning both
• Divergent - investigate before reporting; usually a boundary mismatch, an inflated report, or an unrealistic price or adoption assumption
• Report the more conservative figure as the base case unless the gap is explained

Phase 5: Document assumptions

Every number gets a source or an explicit assumption marker.

Flag the two or three assumptions that move SOM the most and show SOM at low and high values for each.

Phase 6: Present the result

• Lead with SOM and SAM, not TAM; investors and boards discount TAM-led stories
• Show the funnel visually: TAM → SAM → SOM with the filter applied at each step
• State the time frame and boundary on the same slide as the numbers
• Be conservative: optimistic estimates lose credibility faster than conservative ones gain it

Output

• Market boundary statement
• Top-down TAM, SAM, and SOM with sources
• Bottom-up TAM, SAM, and SOM with unit inputs
• Triangulation table with gap explanation
• Assumption log with sources, confidence, and sensitivity
• Summary range and recommended base case

Operating rules

Always

• Define the boundary before calculating anything
• Run both top-down and bottom-up methods
• Give every number a source or an explicit assumption marker
• Explain any gap between methods before reporting
• Show the sensitivity of SOM to its biggest assumptions

Never

• Present TAM as achievable revenue
• Rely on a single analyst report without checking its category definition
• Mix boundaries or years across methods
• Hide assumptions inside formulas
• Round up to make the story more compelling