unit-economics

Anthropic ｜ 金融與投資 ｜ 原文 ｜ Apache-2.0

分析 PE 標的的 ARR cohort、LTV/CAC、淨留存和營收品質

---
name: unit-economics
description: 分析 PE 標的的單位經濟：ARR 世代（cohort）、LTV/CAC、淨留存、回收期、營收品質和利潤率 waterfall。對軟體／SaaS、經常性營收和訂閱制事業特別重要。適用於評估營收品質、建立 cohort 分析或評估客戶經濟。觸發詞："unit economics"、"cohort analysis"、"ARR analysis"、"LTV CAC"、"net retention"、"revenue quality"、"customer economics"。
---

# 單位經濟分析

## 工作流程

### 步驟 1：辨識商業模式

判斷營收模式，以便量身調整分析：
- **SaaS／訂閱**：ARR、淨留存、cohort
- **經常性服務**：合約價值、續約率、加購（upsell）
- **交易／用量計價**：每筆交易營收、交易量趨勢、抽成率（take rate）
- **混合型**：依營收來源拆開

### 步驟 2：核心指標

#### ARR／營收品質
- **ARR bridge**：期初 ARR → 新增 → 擴張 → 縮減 → 流失 → 期末 ARR
- **依 cohort 看 ARR**：vintage 分析，每個年度 cohort 的留存和成長狀況如何？
- **營收集中度**：前 10/20/50 大客戶佔總額的 %
- **依類型看營收**：經常性 vs. 非經常性 vs. 專業服務
- **合約結構**：ACV 分布、多年期 %、自動續約 %

#### 客戶經濟
- **CAC（客戶獲取成本）**：總 S&M 支出 / 新獲取客戶數
- **LTV（生命週期價值）**：(ARPU × 毛利率) / 流失率
- **LTV:CAC 比率**：健康的事業目標 >3x
- **CAC 回收期**：回收獲取成本需要的月數
- **整體 vs. 分群**：依客戶區隔拆開（企業 vs. 中小企業 vs. 中型市場）

#### 留存與擴張
- **毛留存（Gross retention）**：期初 ARR 中保留下來的 %（不含擴張）
- **淨留存（NDR）**：期初 ARR 中保留下來的 %（含擴張）
- **Logo 流失**：流失的客戶 %
- **金額流失**：流失的營收 %（常常跟 logo 流失不同）
- **擴張率**：加購 + 交叉銷售佔期初 ARR 的 %

#### Cohort 分析
建立 cohort 矩陣，呈現：

| Cohort | 第 0 年 | 第 1 年 | 第 2 年 | 第 3 年 | 第 4 年 |
|--------|--------|--------|--------|--------|--------|
| 2020 | $1.0M | $1.1M | $1.2M | $1.1M | |
| 2021 | $1.5M | $1.7M | $1.8M | | |
| 2022 | $2.0M | $2.3M | | | |
| 2023 | $3.0M | | | | |

同時呈現絕對金額 $ 和指數化（第 0 年 = 100%）兩種檢視。

#### 利潤率 Waterfall
- 營收 → 毛利 → 貢獻毛利 → EBITDA
- 完全分攤後的單位經濟：獲取、服務、留住一個客戶要花多少成本？
- 依營收來源看毛利率（訂閱 vs. 服務 vs. 其他）

### 步驟 3：基準比較

把單位經濟跟相關基準比較：
- **SaaS Rule of 40**：成長率 + EBITDA 利潤率 > 40%
- **SaaS Magic Number**：淨新增 ARR / 前一期 S&M 支出 > 0.75x
- **NDR 基準**：最佳 >120%，良好 >110%，令人擔憂 <100%
- **LTV:CAC**：最佳 >5x，良好 >3x，令人擔憂 <2x
- **毛留存**：最佳 >95%，良好 >90%，令人擔憂 <85%
- **CAC 回收期**：最佳 <12 個月，良好 <18 個月，令人擔憂 >24 個月

### 步驟 4：營收品質評分

整合成營收品質評估：

| 因素 | 分數 (1-5) | 備註 |
|--------|-------------|-------|
| 經常性 % | | |
| 淨留存 | | |
| 客戶集中度 | | |
| Cohort 穩定性 | | |
| 成長持續性 | | |
| 利潤率輪廓 | | |
| **整體** | | |

### 步驟 5：輸出

- Excel 活頁簿，包含 ARR bridge、cohort 矩陣、單位經濟儀表板
- 附關鍵指標和基準的摘要投影片
- 紅旗警訊和需要進一步盡職調查的領域

## 重要注意事項

- 如果有的話，一律索取客戶層級的原始資料，彙總指標可能掩蓋問題
- 擴張夠強時，NDR 高於 100% 可能掩蓋很高的毛流失，所以兩個都要呈現
- Cohort 分析是評估營收品質最重要的單一視角，要積極爭取這份資料
- 區分合約 ARR 和實際認列的營收
- 用量計價的模式，重點放在消耗趨勢和擴張模式，而不是傳統的 ARR 指標
- 專業服務營收要另外評估，它不是經常性的，利潤率通常也比較低

---
name: unit-economics
description: Analyze unit economics for PE targets — ARR cohorts, LTV/CAC, net retention, payback periods, revenue quality, and margin waterfall. Essential for software/SaaS, recurring revenue, and subscription businesses. Use when evaluating revenue quality, building a cohort analysis, or assessing customer economics. Triggers on "unit economics", "cohort analysis", "ARR analysis", "LTV CAC", "net retention", "revenue quality", or "customer economics".
---

# Unit Economics Analysis

## Workflow

### Step 1: Identify Business Model

Determine the revenue model to tailor the analysis:
- **SaaS / Subscription**: ARR, net retention, cohorts
- **Recurring services**: Contract value, renewal rates, upsell
- **Transaction / usage-based**: Revenue per transaction, volume trends, take rate
- **Hybrid**: Break down by revenue stream

### Step 2: Core Metrics

#### ARR / Revenue Quality
- **ARR bridge**: Beginning ARR → New → Expansion → Contraction → Churn → Ending ARR
- **ARR by cohort**: Vintage analysis — how does each annual cohort retain and grow?
- **Revenue concentration**: Top 10/20/50 customers as % of total
- **Revenue by type**: Recurring vs. non-recurring vs. professional services
- **Contract structure**: ACV distribution, multi-year %, auto-renewal %

#### Customer Economics
- **CAC (Customer Acquisition Cost)**: Total S&M spend / new customers acquired
- **LTV (Lifetime Value)**: (ARPU × Gross Margin) / Churn Rate
- **LTV:CAC ratio**: Target >3x for healthy businesses
- **CAC payback period**: Months to recover acquisition cost
- **Blended vs. segmented**: Break down by customer segment (enterprise vs. SMB vs. mid-market)

#### Retention & Expansion
- **Gross retention**: % of beginning ARR retained (excludes expansion)
- **Net retention (NDR)**: % of beginning ARR retained including expansion
- **Logo churn**: % of customers lost
- **Dollar churn**: % of revenue lost (often different from logo churn)
- **Expansion rate**: Upsell + cross-sell as % of beginning ARR

#### Cohort Analysis
Build a cohort matrix showing:

| Cohort | Year 0 | Year 1 | Year 2 | Year 3 | Year 4 |
|--------|--------|--------|--------|--------|--------|
| 2020 | $1.0M | $1.1M | $1.2M | $1.1M | |
| 2021 | $1.5M | $1.7M | $1.8M | | |
| 2022 | $2.0M | $2.3M | | | |
| 2023 | $3.0M | | | | |

Show both absolute $ and indexed (Year 0 = 100%) views.

#### Margin Waterfall
- Revenue → Gross Profit → Contribution Margin → EBITDA
- Fully loaded unit economics: what does it cost to acquire, serve, and retain a customer?
- Gross margin by revenue stream (subscription vs. services vs. other)

### Step 3: Benchmarking

Compare unit economics to relevant benchmarks:
- **SaaS Rule of 40**: Growth rate + EBITDA margin > 40%
- **SaaS Magic Number**: Net new ARR / prior period S&M spend > 0.75x
- **NDR benchmarks**: Best-in-class >120%, good >110%, concerning <100%
- **LTV:CAC**: Best-in-class >5x, good >3x, concerning <2x
- **Gross retention**: Best-in-class >95%, good >90%, concerning <85%
- **CAC payback**: Best-in-class <12mo, good <18mo, concerning >24mo

### Step 4: Revenue Quality Score

Synthesize into a revenue quality assessment:

| Factor | Score (1-5) | Notes |
|--------|-------------|-------|
| Recurring % | | |
| Net retention | | |
| Customer concentration | | |
| Cohort stability | | |
| Growth durability | | |
| Margin profile | | |
| **Overall** | | |

### Step 5: Output

- Excel workbook with ARR bridge, cohort matrix, unit economics dashboard
- Summary slide with key metrics and benchmarks
- Red flags and areas for further diligence

## Important Notes

- Always ask for raw customer-level data if available — aggregate metrics can hide problems
- NDR above 100% can mask high gross churn if expansion is strong enough — always show both
- Cohort analysis is the single most important view for revenue quality — push for this data
- Differentiate between contracted ARR and actual recognized revenue
- For usage-based models, focus on consumption trends and expansion patterns rather than traditional ARR metrics
- Professional services revenue should be evaluated separately — it's not recurring and margins are typically lower