← All posts
Unit Economics for Product Leaders — Part 2

Unit Economics for Product Leaders, Part 2: The Bridge to the P&L

In Part 1, we covered the core unit economics formulas and the mistakes that show up everywhere. This post answers the question I see product managers struggle with most: how does "contribution margin per customer" connect to the income statement that your CFO is looking at?

The answer is simpler than it seems, but the implications are not.


Two Views of the Same Business

Unit economics and the P&L are two views of the same underlying reality, sliced differently.

Unit economics is the per-customer, per-period view. It answers: "Is each incremental customer accretive?" It sees only variable costs and revenue at the individual customer level. Fixed costs don't exist in this view — and that's by design, not an oversight.

The P&L is the portfolio, per-period view. It answers: "Is the business profitable?" It loads in the full fixed cost base on top of the total contribution across all customers.

The bridge between them is multiplication:

Portfolio Revenue = ARPU × N customers

Portfolio Variable Costs = Variable Cost Per Customer × N customers

Portfolio Contribution = CM Per Customer × N customers

Portfolio Profit = Portfolio Contribution − Total Fixed Costs

A company can have positive unit economics and still lose money if fixed costs exceed total contribution. Conversely, negative unit economics means scaling makes things worse — every new customer digs the hole deeper. That's the whole reason for computing them separately.


The P&L Waterfall

The income statement is read top-down. Each line subtracts a cost layer from revenue, producing a progressively narrower margin.

Line ItemFormulaWhat It Tells You
RevenueTotal invoiced sales
− COGSDirect costs of delivering the product
= Gross ProfitRevenue − COGSMargin on the product itself before operating costs
− Variable Costs (below COGS)Additional costs that scale with volume
= Contribution MarginRevenue − All Variable CostsWhat each revenue dollar contributes toward fixed costs
− Opex (Fixed Costs)R&D, G&A, office leases, exec salaries
= EBITContribution − Fixed CostsProfitability of core operations
+ D&ATypically 0–3% of revenue for cloud-native SaaS
= EBITDA
− Interest & Taxes
= Net IncomeThe "bottom line"

There's a subtlety here that trips up most people. Read on.


Why Gross Margin ≠ Contribution Margin

COGS — cost of goods sold — often contains both fixed and variable components. In SaaS, COGS typically includes things like:

  • Cloud infrastructure (variable — scales with customers/usage)
  • Data subscriptions or third-party APIs (often fixed — annual subscription regardless of customer count)
  • Platform engineering salaries (fixed — same team whether you have 10 or 10,000 customers)
  • Payment processing fees (variable — scales with transaction volume)

Because COGS blends fixed and variable costs, Gross Profit does not tell you about unit economics. A company with 70% gross margin might have excellent or terrible unit economics depending on how much of COGS is fixed vs. variable.

Contribution Margin is the metric that tells you about unit economics, because it isolates variable costs regardless of where they sit on the P&L. The formula is:

Contribution Margin = Revenue − All Variable Costs

"All" means variable costs inside COGS and variable costs below COGS (sales commissions, per-customer support, usage-based infrastructure). The fixed/variable distinction cuts across P&L line items — it doesn't map neatly to specific rows.

This also means you can't compute Contribution Margin by reading down the P&L sequentially (Revenue → minus COGS → minus more stuff). You have to decompose COGS into its fixed and variable components first. That decomposition is analytical work that the P&L format doesn't do for you.


The Cost Taxonomy That Matters

For unit economics, there's only one question that matters for each cost line item:

"If I add one more customer, does this cost increase?"

If yes, it's variable. If no, it's fixed. That's the entire taxonomy.

Variable costs (included in unit economics):

  • Per-customer cloud compute and storage
  • Per-customer data delivery or API calls
  • Usage-based third-party costs
  • Sales commissions tied to individual deals
  • Per-customer onboarding labor
  • Payment processing fees

Fixed costs (excluded from unit economics, included in portfolio P&L):

  • Annual platform subscriptions or data licenses
  • Engineering salaries
  • R&D
  • Office leases, insurance, exec compensation
  • Marketing spend (typically — though some PMs treat it as variable via CAC)

A few costs are genuinely ambiguous. Customer success headcount, for instance: one CSM might handle 20 accounts, and you hire another CSM when you hit 21. Is that fixed or variable? In the short run, it's fixed (adding customer #15 doesn't change the cost). In the long run, it's step-variable. For unit economics purposes, if the cost scales roughly proportionally with customer count, treat it as variable.


Costs with Multiple Dimensions of Scale

Some businesses have costs that scale along one dimension but not another. A clear example: per-customer onboarding costs. Whether a customer's contract is 50Kor50K or 500K, the onboarding effort might be roughly the same — the integration work, configuration, and training don't scale with contract value. They scale with customer count.

Meanwhile, delivery costs (compute, storage, API calls) scale with usage, which usually correlates with contract value because larger customers use the product more.

This creates a two-level structure:

Level 1 — Per delivery unit: Revenue per unit minus variable delivery cost per unit = contribution per unit. This is the atomic level where both revenue and cost are expressed in directly comparable terms.

Level 2 — Per customer: (Contribution per unit × customer's usage volume) minus per-account costs = customer-level contribution. This is where minimum deal size and pricing floor logic comes from.

The question connecting the two levels:

Minimum viable account size = Per-account costs / Contribution per delivery unit

If per-account costs are 48K/yearandcontributionperunitis48K/year and contribution per unit is 100/unit-year, you need at least 480 units of delivery before a customer is contribution-positive. Customers below this threshold are unprofitable regardless of how good the per-unit economics look.

This isn't a "fixed cost allocated to customers." Per-account costs are variable at the customer level — add a customer, they appear. They just happen to be fixed at the delivery-unit level. The language matters: calling them "fixed" obscures the fact that they scale with customer count and must be included in customer-level contribution calculations.


The Bridge, Made Explicit

Here's a complete bridge from unit economics to portfolio profitability, using concrete numbers:

Unit Level:

  • ARPU: $200K/year
  • Variable cost per customer: $50K/year
  • CM per customer: $150K/year
  • CM%: 75%

Portfolio Level (40 customers):

  • Portfolio Revenue: 200K×40=200K × 40 = 8M
  • Portfolio Variable Costs: 50K×40=50K × 40 = 2M
  • Portfolio Contribution: 150K×40=150K × 40 = 6M
  • Total Fixed Costs: $4M
  • Portfolio Profit: 6M6M − 4M = $2M

Breakeven:

  • 4M/4M / 150K = 26.7 → 27 customers to cover fixed costs

At 40 customers with positive unit economics, this business is profitable. At 20 customers with the same unit economics, it loses $1M. The unit economics haven't changed — the portfolio math has. That's the whole point of separating the two views.


What Belongs Where: A Cheat Sheet

MetricLevelFixed costs?What it tells you
ARPUUnitNoRevenue per customer per period
Variable Cost / CustomerUnitNoCost that scales with each additional customer
Contribution Margin / CustomerUnitNoIs each incremental customer accretive?
CM-Based LTVUnitNoTotal profit contribution over customer lifetime
CAC PaybackUnitNoHow long to recover acquisition cost
Gross ProfitPortfolioYes (partial)Margin on the product after COGS (fixed + variable)
Portfolio ContributionPortfolioNoTotal CM across all customers
Portfolio Profit / EBITPortfolioYesIs the business profitable after all costs?
Breakeven CustomersPortfolioYesHow many customers to cover fixed costs?

The bright line: unit economics never includes fixed costs. The moment you allocate a fixed cost to a per-customer metric, you've crossed from unit economics into P&L territory. That's fine if it's intentional, but conflating the two leads to bad decisions (more on this in Part 3).


A Note on NRR and the Portfolio View

Net Revenue Retention modifies the portfolio view across time periods, not within them. Current-period portfolio economics use actual current ARPU and current customer count. NRR tells you what next period's ARR will be from the existing customer base before new logo acquisition.

If NRR is 110%, a 8MARRportfoliobecomes8M ARR portfolio becomes 8.8M next year from the existing customer base alone, before adding any new customers. If NRR is 85%, that same portfolio erodes to 6.8Mnewsaleshavetofilla6.8M — new sales have to fill a 1.2M hole before the business can grow.

NRR doesn't modify any current-period unit economics calculations. It's a projection mechanism, not a cost or revenue input.


In Part 3, we'll use unit economics and the P&L bridge to make actual product decisions: when to kill a product, how to choose between investments, why bundling can destroy value, and how to prioritize when you're capacity-constrained.