Opportunity Sizing

Figure out if something is worth building before you build it.

How to use

• /opportunity-sizing Apply opportunity sizing constraints to this conversation.
• /opportunity-sizing <idea> Size the opportunity for the described product or feature idea.

Constraints

Opportunity Statement

MUST state the opportunity in one sentence before sizing: "If we build [X], [target users] will be able to [outcome], which drives [business metric]." If you can't fill this in, the opportunity isn't defined enough to size.

Market Sizing

Work through three layers:

• TAM (Total Addressable Market): everyone who could theoretically use this. Your ceiling.
• SAM (Serviceable Addressable Market): the portion you can realistically reach. Usually 10-30% of TAM.
• SOM (Serviceable Obtainable Market): what you can actually capture in a reasonable timeframe. Usually 1-5% of SAM.
• MUST calculate all three. NEVER cite TAM as if it's achievable.

Revenue Estimation

• Bottom-up (preferred): reachable users × conversion rate × price × retention rate
• Top-down (sanity check): total market revenue × realistic market share
• If bottom-up and top-down diverge wildly, assumptions need work
• MUST run three scenarios: conservative (50% reach, 70% conversion, 150% cost), base case, optimistic (150% reach, 120% conversion)
• Opportunity is worth pursuing if conservative case is acceptable and base case is compelling

Cost Assessment

MUST include all four cost types:

• Build cost: engineering, design, product time (include QA, docs, infrastructure)
• Go-to-market cost: marketing, sales, partnerships, education
• Ongoing cost: infrastructure, support, maintenance, iteration
• Opportunity cost: what you're NOT building while you build this. The hidden cost most teams ignore.

Key Assumptions

• MUST list top 5 assumptions the sizing depends on
• For each: confidence level (1-10), impact if wrong, how to test before committing
• Low confidence + high impact = validate this first, before going all-in

Quick Sizing (Smaller Decisions)

For features or experiments that don't warrant full analysis:

• Who benefits × how much per user = total value
• If maximum possible impact < cost to build, stop here
• If maximum possible impact > 10× cost, stop analyzing and start building

Anti-Patterns

• Optimism bias: founders consistently overestimate. Deliberately discount first estimates by 30-50%.
• TAM delusion: "$50B market" means nothing about your ability to capture any of it
• Ignoring alternatives: users have options including doing nothing
• Static analysis: size where the market will be, not just where it is
• Ignoring time dimension: $10M over 5 years is very different from $10M over 1 year