seo-geo-report-engine: Skill for Claude Code

market-opportunity

Sizes a market and frames the opportunity as a feed into discovery-audit (the intake brief) and proposal-builder (the pitch). The IP is dual validation — every market number is computed two independent ways (top-down from industry figures, bottom-up from demand-side signals) and the gap between them is the honesty check. Prefer Ahrefs MCP (see knowledge/ahrefs-mcp-map.md); call doc on a tool before first use. The methodology lives in playbooks/market-sizing.md — read it.

Methodology (PERCEIVE → ANALYZE → VALIDATE → ACT)

PERCEIVE — gather. Resolve the project (./bin/mkt config show --project <client>); read client.yml for domain, category, price point, locales, competitors, ICP. Establish the market frame: what is being sold, to whom, in which geographies, at what price.

• Demand-side bottom-up signal. Pull total search volume for the category's commercial-intent head + body terms with keywords-explorer-overview (read volume and traffic_potential), expand the term set with keywords-explorer-matching-terms / -related-terms, and split it by market with keywords-explorer-volume-by-country. This is your bottom-up demand proxy — real buyers searching, per geography — not a borrowed analyst number.
• Top-down inputs. Capture published industry size, growth rate, and ARPU/ACV (from client.yml, the brief, or WebSearch for analyst figures — cite source + date).
• Competitor set. From client.yml competitors + site-explorer-organic-competitors, list rivals; pull site-explorer-domain-rating and site-explorer-metrics for relative scale.

ANALYZE — size, map, force, gate, score.

1. TAM / SAM / SOM, both directions (full worked method in the playbook):
   • Top-down: TAM = industry size; SAM = TAM × (segments you serve); SOM = SAM × realistic share over the plan horizon.
   • Bottom-up: TAM ≈ addressable buyers × ARPU; SAM = restrict to served geos/segments (use the -volume-by-country split as the geo weighting); SOM = SAM × capturable demand (search volume you can realistically rank for / convert, given DR and funnel).
   • Reconcile. Put the two TAMs side by side. A < ~2–3× gap is healthy; a large gap means one side rests on a bad assumption — say which, and which number you carry forward.
2. Price-vs-complexity 2D positioning map. Plot each competitor on price (low→high) × product complexity / implementation effort (simple→complex). Name the quadrant the client occupies and the white space (an empty or thin quadrant with demand behind it).
3. Porter's Five Forces. Rate each force (low/med/high) with a one-line evidence note: rivalry, new entrants, supplier power, buyer power, substitutes. This frames how defensible any SOM capture is.
4. PMF qualifier (Sean-Ellis "very disappointed" gate). State the PMF read: do ≥40% of users say they'd be "very disappointed" without the product (or the best proxy available — retention, organic pull, referral)? This is the scale-vs-iterate gate: below the bar → recommend iterate (cheap discovery/positioning tests), NOT heavy paid spend; at/above → scaling spend is defensible. The proposal must inherit this gate so it never recommends heavy spend pre-PMF.