PK
Back to archive

Projects & Notes

Unit economics

Business metrics for engineers who want their technical decisions to connect to the P&L.

Business Terms Every Engineer Must Know 🧵

1/10

"You build the product. You ship the features. You fix the pager alerts.

But when the VP asks 'how does this affect our GMV?' — what do you say?

Business terms aren't optional reading for Staff+ engineers. They're your native language at that level."

🧵

2/10

Revenue ≠ GMV ≠ Profit

GMV (Gross Merchandise Value): Total value of all goods/services sold. If customers buy ₹100Cr worth of stuff on your platform — that's ₹100Cr GMV. You don't keep this money.

Revenue: What the platform actually earns. For a marketplace: commissions, fees, ads. If your take rate is 5%, revenue = ₹5Cr on ₹100Cr GMV.

Profit: Revenue minus all costs (hosting, salaries, payment gateway fees, office chai).

3/10

The waterfall every engineer should memorize:

GMV (total transaction value) → Net Revenue (GMV × take rate, minus refunds/chargebacks) → Gross Profit (Net Revenue - COGS) → EBIT (Gross Profit - S&M, R&D, G&A) → Net Income (EBIT - interest - taxes)

At each step, your engineering decisions leak or preserve value.

4/10

Where engineering hits P&L:

COGS (Cost of Goods Sold) — In SaaS, this is infra. Your DB provisioning, CDN egress, GPU compute. Every query you optimize, every instance you right-size, directly improves Gross Margin.

Latency as a P&L item: A 500ms p99 vs 50ms p99 isn't just UX. At 10M requests/month with ads at checkout, that 450ms costs ₹X in drop-off. Engineering latency = revenue impact. Be able to quote the number.

5/10

P&L = Profit & Loss statement

The company's report card over a period (monthly, quarterly, annual).

Structure:

Revenue:               ₹100 Cr
- COGS:                ₹30 Cr
= Gross Profit:        ₹70 Cr   (70% margin)
- Opex:                ₹50 Cr   (Sales + Eng + Admin)
= EBIT:                ₹20 Cr
- Interest & Tax:      ₹5 Cr
= Net Income:          ₹15 Cr

Key metric for engineers: Gross Margin (Gross Profit / Revenue). If it's dropping, your costs are growing faster than revenue.

6/10

Unit Economics — the engineer's superpower

Two numbers that separate "building features" from "building a business":

CAC (Customer Acquisition Cost): Total sales & marketing spend / new customers acquired. If you spend ₹50L on ads and get 1000 customers, CAC = ₹5,000.

LTV (Lifetime Value): Average revenue per customer × avg months retained. If each customer pays ₹1,000/mo and stays 24 months, LTV = ₹24,000.

The rule: LTV > 3× CAC. If your infra costs push LTV below that, the business model breaks. You have a P&L problem, not a code problem.

7/10

ARPU, ARR, MRR — the growth dials

  • ARPU (Avg Revenue Per User): Total revenue / total users. Engineering directly influences this through pricing tiers, feature gating, upgrade flows.

  • MRR (Monthly Recurring Revenue): The SaaS lifeline. Predictable monthly subscription revenue. Churn is the enemy — every failed API call, every slow page load is a churn risk.

  • ARR (Annual Run Rate): MRR × 12. The number investors care about most.

8/10

The metrics that reveal product health

Net Revenue Retention (NRR) — Revenue from existing customers including upgrades/downgrades/churn. >120% NRR means your product sells itself (existing customers spend more each year). <100% NRR means you're running on a treadmill — you need new sales just to stay flat.

Gross Merchandise Value (GMV) take rate — For marketplaces. If your GMV grows but take rate drops, you're growing unprofitable volume. Engineering can fix this via better fraud detection, payment optimization, automated dispute handling.

9/10

How to speak business in a design review

Before proposing an architectural change, ask yourself:

  1. "How does this affect GMV/revenue?" (Slows checkout? Enables a new pricing tier?)
  2. "How does this affect COGS?" (Reduces infra cost? Adds a new dependency?)
  3. "How does this affect NRR?" (Reduces churn? Enables upsells?)
  4. "How does this affect CAC?" (Self-serve onboarding vs sales-led?)

If you can't answer all four, you don't know what you're optimizing for.

10/10

Bottom line:

The best Staff+ engineers don't just understand the code. They understand the P&L statement, can trace their infra decisions to Gross Margin, and argue in terms of LTV:CAC ratio — not just latency percentiles.

The golden rule: Every engineering decision is either improving Gross Margin, growing Revenue, or managing Risk. If you can't say which — you're building in the dark.

#EngineeringLeadership #BusinessMetrics #StaffEngineer #PAndL #UnitEconomics