# Market Sizing Workflow > Pipeline | Sizing Flow | Multilingual Queries | Step 5b: Counterfactual | Pitfalls | Output (Market Size, Share, Concentration, Growth, Gaps) **Question pattern:** "What's the market for X?" / "How big is the CMP slurry market?" / "What's the TAM for EUV pellicles?" ## Investigation Pipeline ``` User Question: "What's the CMP slurry market breakdown?" │ ▼ ┌─────────────────────────────────┐ │ 1. DEFINE THE SEGMENT │ What exactly are we sizing? │ │ ├── Material? Equipment? Service? │ Scope the question │ ├── By application? (oxide CMP vs metal CMP) │ │ ├── By node? (advanced <7nm vs mature) │ │ └── Geographic? (global vs China-domestic) └──────────┬──────────────────────┘ │ ▼ ┌─────────────────────────────────┐ │ 2. TOP-DOWN: ANALYST REPORTS │ Published market size estimates │ │ ├── TrendForce (memory, foundry utilization) │ Search for existing sizing │ ├── Yole Group (packaging, compound semi) │ │ ├── SEMI (equipment, World Fab Forecast) │ │ ├── TechCET (materials — CMP, gases, wet chem) │ │ ├── Omdia (broad market data, expensive) │ │ └── Broker reports: │ │ ├── Naver Finance → 리서치 tab (Korea) │ │ ├── EastMoney → 研报中心 (China) │ │ ├── Kabutan (Japan) │ │ └── MoneyDJ / MOPS (Taiwan) │ │ │ ⚠ ALWAYS flag source + │ Most analyst data is paywalled. │ date + whether you accessed │ Press summaries often quote numbers. │ it or citing from memory │ Search: "[material] market size 2024" └──────────┬──────────────────────┘ │ ▼ ┌─────────────────────────────────┐ │ 3. BOTTOM-UP: COMPANY DATA │ Sum individual company revenues │ │ │ Build from public filings │ Revenue sources by filing system: │ │ ├── DART → 매출 현황, 주요 제품 현황 │ │ ├── EDINET → セグメント情報, 事業の内容 │ │ ├── MOPS → monthly 月營收 (unique — fastest) │ │ ├── cninfo → 主营业务分析 │ │ └── 10-K → segment disclosures (US) │ │ │ Cross-reference: │ Company A segment revenue │ Sum known players │ + Company B segment revenue │ vs analyst total = gap │ + Company C segment revenue │ Gap = unlisted/private cos │ ───────────────────────────── │ │ = Bottom-up estimate │ │ Compare vs top-down → consistency check └──────────┬──────────────────────┘ │ ▼ ┌─────────────────────────────────┐ │ 4. TRADE DATA CROSS-CHECK │ Validate with material flows │ │ │ Directional validation only │ ├── Comtrade: HS code bilateral flows │ Trade data ≠ market size │ │ (3818=wafers, 3707=resist, 8486=equip) │ but exposes inconsistencies │ ├── e-Stat: Japan exports at 9-digit HS │ │ └── Taiwan customs (MOPS trade data) │ │ │ Example sanity check: │ If analyst says "Japan photoresist market │ Trade flow vs market claim │ = $4B" but Japan HS 3707 total exports │ │ = $3.3B → either market includes domestic │ │ consumption or definition differs └──────────┬──────────────────────┘ │ ▼ ┌─────────────────────────────────┐ │ 5. CONCENTRATION ANALYSIS │ Who holds how much? │ │ │ Market structure matters │ ├── Top-1 share → monopoly risk? │ more than total TAM │ ├── Top-3 share → oligopoly? │ │ ├── HHI index (sum of squared shares) │ │ ├── Geographic concentration │ │ └── Compare vs chokepoints list │ │ (chemistry/precursor-chains.md) └──────────┬──────────────────────┘ │ ▼ ┌─────────────────────────────────┐ │ 6. OUTPUT │ │ Grade every number │ Market size: $X.XB ± range │ Source everything │ Growth: X% CAGR (source, period) │ Flag what's estimated │ Share breakdown by company │ vs confirmed │ Concentration metrics │ │ Source + date for EACH figure │ │ "What I Could Not Size" section └─────────────────────────────────┘ ``` ## Market Sizing Flow (Material Example) > **Note:** The figures below are illustrative methodology, not verified data. > They show how to structure a market sizing exercise, not actual confirmed numbers. ``` Global CMP Slurry Market (~$2.5B est.) │ ├── By application │ ├── Oxide CMP (colloidal silica) ── ~40% │ ├── Metal CMP (Cu, W, barrier) ──── ~35% │ ├── STI CMP (ceria) ──────────────── ~15% │ └── Other (poly, ILD) ────────────── ~10% │ ├── By supplier (approx. share, UNVERIFIED) │ ├── Entegris/CMC Materials (US) ──── #1 │ ├── Fujimi (5384.T, Japan) ────────── #2-3 │ ├── AGC (5201.T, Japan) ───────────── #2-3 │ ├── DuPont (US) ──────────────────── Top 5 │ ├── Anji Micro (688019.SS, China) ── Growing │ └── KC Tech (281820.KQ, Korea) ───── Regional │ ├── By geography (demand) │ ├── Taiwan (TSMC) ───── ~30% │ ├── Korea (Samsung/SK) ── ~25% │ ├── China ──────────────── ~20% (growing fast) │ ├── Japan ──────────────── ~10% │ └── US/EU ─────────────── ~15% │ └── Upstream chokepoint └── Colloidal silica → Fuso Chemical (4368.T) [VERIFY] └── Ceria → China rare earth processing [VERIFY] ``` ## Multilingual Search Queries ### Korean (시장 규모): ``` [소재명] 시장 규모 → "[material] market size" [소재명] 시장 점유율 → "[material] market share" 반도체 [소재명] 시장 전망 → "semiconductor [material] market outlook" [소재명] 세계 시장 → "[material] global market" ``` ### Japanese (市場規模): ``` {材料名} 市場規模 → "[material] market size" {材料名} シェア → "[material] share" 半導体材料 市場 予測 → "semiconductor materials market forecast" {材料名} 世界市場 → "[material] global market" ``` ### Chinese — Mainland (市场规模): ``` [材料名] 市场规模 → "[material] market size" [材料名] 市场份额 → "[material] market share" [材料名] 国产化率 → "[material] localization rate" [材料名] 行业研究报告 → "[material] industry research report" ``` ### Chinese — Taiwan (市場規模): ``` [材料名] 市場規模 → "[material] market size" [材料名] 市佔率 → "[material] market share" (Taiwan phrasing) 半導體 [材料名] 市場 展望 → "semiconductor [material] market outlook" [材料名] 全球市場 → "[material] global market" ``` ## Step 5b: Counterfactual check on sizing When top-down (analyst reports) and bottom-up (summing company revenues) estimates diverge by >20%, run the [Counterfactual Consistency Check](counterfactual-check.md). Don't average the numbers. Surface the contradiction: which source uses what market definition? Does the analyst include adjacent segments? Are private companies creating a blind spot in the bottom-up sum? A documented disagreement is more useful than a false consensus. ## Common Pitfalls 1. **"Market" definition varies wildly.** CMP slurry market — does that include pads? Conditioners? Diamond discs? Always clarify scope. 2. **Currency matters.** A "$2B market" in a 2020 report at ¥110/$ is different from 2024 at ¥150/$. Japanese company revenues in yen can look like growth when it's just depreciation. 3. **Analyst reports age fast.** A 2022 market sizing doesn't account for the 2023-2024 AI/HBM demand surge. Always note the report date. 4. **China double-counts.** Chinese broker reports sometimes count domestic + import as "China market" and sometimes only domestic production. Clarify which. 5. **Private companies are black holes.** Morita Chemical (HF), SEMES (equipment), Tanaka Precious Metals — no public revenue data. Their share is estimated by subtraction. 6. **MOPS monthly revenue is gold.** Taiwan is the only market where you get company revenue within 10 days of month-end. Use it for real-time market tracking. ## Output Format ```markdown # [Material/Equipment] Market Assessment ## Market Size - Global: $X.XB (YYYY) [SOURCE, accessed/not accessed] - CAGR: X% (YYYY-YYYY) [SOURCE] - ⚠ Confidence: [CONFIRMED from accessed report / APPROXIMATE from press summary / ESTIMATED bottom-up] ## Market Share Breakdown | Company | Approx. Share | Revenue (segment) | Source | Confidence | |---|---|---|---|---| ## Concentration - Top-1: X% | Top-3: X% | HHI: XXXX - Geographic: [X% from Country A] - Chokepoint: [tier-2 dependency if any] ## Growth Drivers - [driver 1] - [driver 2] ## What I Could Not Size - [gaps, unavailable data, private company blind spots] ```