market-sizing

Produce a rigorous, sourced TAM/SAM/SOM market sizing for any product or business. Use this skill whenever a user asks about market size, total addressable market, SAM, SOM, or market opportunity — across any industry including SaaS, AI tools, consumer brands, F&B, fashion, beauty, packaging, automotive, and more.

# TAM / SAM / SOM Market Sizing

Produce both a bottom-up and a top-down analysis for every engagement. Both methods are required — they serve different purposes and must be reconciled. The goal is a defensible, sourced output — not a fast estimate.

---

## When to use this skill
Trigger on any prompt that contains:
- "TAM", "SAM", "SOM", or "market size"
- "How big is the market for X"
- "Total addressable market", "serviceable market", "obtainable market"
- "Market sizing for [product/company]"
- Investor deck context requiring market opportunity quantification

---

## Loading examples
Before calculating, check the `examples/` directory for a file matching the product's category. Load the closest match to calibrate data sources, filter logic, and ACV anchors before you begin.

| File | Use when product is... |
|------|----------------------|
| `examples/b2b-saas.md` | Software / AI tool sold to businesses |
| `examples/b2c-consumer-brand.md` | Consumer packaged goods, DTC brands, subscriptions |
| `examples/b2b-physical.md` | Physical goods or materials sold to businesses |

If no exact match exists, load the closest file and adapt.

---

## Pre-flight: Classify the product before doing anything else

Before searching or calculating, classify the product along these two axes. The classification directly determines which data sources to use and which pricing method applies.

### Axis 1 — Business model
| Code | Type | Revenue unit | Pricing anchor |
|------|------|-------------|----------------|
| **B2B-SaaS** | Software sold to businesses | Annual contract (ACV) | 10–20% of economic value created |
| **B2B-Physical** | Physical goods sold to businesses | Per-unit price × annual volume | Gross margin benchmarks by industry |
| **B2C-Brand** | Consumer product (any category) | Revenue per customer per year | Avg. purchase price × purchase frequency |
| **B2C-Subscription** | Consumer subscription | Monthly/annual subscription fee | Stated price or comp pricing |
| **Marketplace/Platform** | Takes % of GMV | Take rate × GMV | Industry take rate benchmarks |

### Axis 2 — Market geography
- **US-only**: Use US Census, NAICS codes, US industry associations
- **Global**: Use global market reports, then apply regional share (US ~25–30% of global GDP; NA ~32%; APAC ~38%)
- **Single city/region**: Use metro area population data + category penetration rates

## Step-by-step workflow (universal)

### Step 0 - Data Gathering
Since accuracy and reliability are key of this task, make sure you collect enough reliable data from reliable sources before the calculation. Limited estimation is fine with enough evidence support, but any kind of data making up or hallucination without evidence during the calculation and output phase is a fraud. 

### STEP 1 — Define the revenue unit
Before any market data search, lock down exactly what the product sells, to whom, and at what price.

**Required inputs:**
- Who is the buyer? (job title / consumer demographic)
- What is the unit of sale? (per seat, per SKU, per kg, per subscription, per transaction)
- What is the price? (stated, or estimated by comp analysis)
- What is the purchase frequency? (one-time, monthly, annual, recurring)

**If price is unknown:** Search `"[product category] average price" OR "[closest competitor] pricing"` and anchor to the median. For B2B-SaaS, also apply the 10–20% value rule: price = 10–20% of annual economic value the product creates for the customer.

**Output of Step 1:** A single sentence — *"The revenue unit is [X] sold to [Y] at [Z]/year."*

### STEP 2 — TAM (Total Addressable Market)

TAM = the maximum theoretical revenue if every potential buyer purchased the product.

#### 2A — Bottom-up method (PREFERRED for all product types)

**For B2B products:**
1. Search for the number of firms in the target industry
   - US: Use NAICS code lookup — search `"NAICS [code] number of firms US"` or `site:siccode.com NAICS [code]`
   - Global: Use IBISWorld, Statista, or national business registries
2. Estimate the % that are genuinely relevant (apply product-fit filter at this stage only if very obvious — e.g. residential-only firms for a commercial B2B product)
3. TAM = Total relevant firms × ACV

**For B2C products:**
1. Find the total consumer population in the target geography (US Census, APPA, Statista, trade associations)
2. Find the % in the target demographic/behavioral segment
3. TAM = Segment population × annual spend per person

**Search queries to use:**
- `"[industry] number of [businesses/firms/brands] US 2024"`
- `"NAICS [code] firm count employees revenue"`
- `"[consumer category] number of [households/users/owners] US"`
- `"[category] market size 2024"` (for cross-reference only — not the primary method)

#### 2B — Top-down method (SECONDARY cross-check)
1. Find total category market size from analyst reports (Grand View, IMARC, Mordor, Precedence, IBISWorld)
2. Apply funnel: Total market → relevant segment → product-type share → geography
3. Search: `"[category] software/product market size 2024"` or `"[category] industry revenue US 2024"`

**Rule:** Bottom-up is the primary method. Top-down is the cross-check. Both results should be reported. If they diverge significantly (>2×), explain why (e.g. top-down only counts current spenders; bottom-up counts total potential).

#### TAM output format:
- Single dollar figure (round to nearest $100M for large markets, $10M for mid-size)
- Conservative range (low–high)
- The math shown explicitly: `N firms × $X ACV = $Y` or `N consumers × $Z annual spend = $Y`

### STEP 3 — SAM (Serviceable Addressable Market)

SAM = the subset of TAM that the specific product can actually serve today, given its business model, geography, language, and product fit.

**Universal filter checklist — apply all that are relevant:**

| Filter | Typical reduction | How to estimate |
|--------|------------------|----------------|
| Geography | Varies | If US-only product, filter out non-US. If city-level, apply metro population fraction. |
| Product-fit segment | 30–60% reduction | Remove segments the product doesn't serve (e.g. residential-only for commercial SaaS; solo operators who can't justify the cost) |
| Tech/channel readiness | 10–30% reduction | % of target customers with internet access, cloud tools, or relevant distribution channel access |
| Language/regulatory | Varies | Only relevant for global products |
| Willingness to pay tier | 20–40% reduction | Remove segments priced out of the product's tier |

**Formula:** SAM = TAM × Filter₁ × Filter₂ × ... × Filterₙ

**Blended ACV for SAM:** If there are multiple customer segments (SME vs. mid-market, or mass vs. premium), calculate a weighted average ACV:
`Blended ACV = (Segment_A_ACV × weight_A) + (Segment_B_ACV × weight_B)`

**Search queries for filter data:**
- `"[industry] percentage [commercial/residential/enterprise/SMB]"`
- `"[industry] small business cloud software adoption rate"`
- `"[consumer category] premium vs mass market share"`

#### SAM output format:
- Single dollar figure with conservative range
- Every filter listed with its % reduction and data source
- Blended ACV (if applicable)

### STEP 4 — SOM (Serviceable Obtainable Market)

SOM = the portion of SAM the product can realistically capture in a defined timeframe (typically 3 years / Year 1–3 ramp).

**Penetration rate benchmarks by product type:**

| Product type | Year 1 | Year 3 | Source basis |
|---|---|---|---|
| New B2B SaaS (no category) | 0.1–0.3% of SAM | 1.5–3% of SAM | Andreessen Horowitz, OpenView benchmarks |
| New B2B SaaS (established category) | 0.3–0.8% | 3–7% | Category incumbents' early growth data |
| New B2C consumer brand (retail) | 0.05–0.2% of SAM | 0.5–2% | Nielsen new brand launch data |
| New B2C DTC subscription | 0.1–0.5% | 1–4% | DTC cohort benchmarks |
| New B2B physical product | 0.2–0.5% | 2–5% | Trade distribution ramp benchmarks |

**Cross-validation method:** Find 2–3 comparable companies (same