The burn multiple.
How much cash each new dollar of ARR costs: the formula, what the ranges mean, and which half comes from Stripe.
The burn multiple is net burn divided by net new ARR. Burning $300,000 in a quarter to add $150,000 of ARR gives a multiple of 2.0. Lower is better. Net new ARR comes from your Stripe billing, and net burn from your bank and accounting.
- Divide net burn by net new ARR: lower is better
- Take net new ARR from Stripe billing
- Take net burn from your bank and accounting
What is the burn multiple?
The burn multiple tells you how much cash you burn for each new dollar of ARR. A burn multiple of 2 means you spent $2 of net cash to add $1 of ARR. Lower is better.
David Sacks popularized it in an April 2020 post, "The Burn Multiple" (Sacks, Substack, checked Nov 2, 2026). His point: growth on its own says little if it costs too much cash. The burn multiple puts growth and burn in one number.
The burn multiple formula, with an example
Burn multiple = net burn ÷ net new ARR
- Net burn: cash out minus cash in over the period. If you're profitable, net burn is zero or negative and the multiple stops being useful.
- Net new ARR: ARR at the end of the period minus ARR at the start. It already nets out churn and contraction.
Example: over the quarter your bank balance fell by $300,000 after all income. MRR went from $80,000 to $92,500, so net new MRR is $12,500 and net new ARR is $12,500 × 12 = $150,000.
Burn multiple = $300,000 ÷ $150,000 = 2.0.
Sacks called a 2x multiple reasonable for an early-stage startup and a 5x multiple terrible. He read 3x or more as a sign of a possible product-market fit problem. Measure it quarterly: one month is too noisy.
See the revenue side of these numbers, MRR growth and churn, in your own Stripe.
What Stripe data covers, and what it doesn't
| Input | In Stripe? | Where to get it |
|---|---|---|
| Net new ARR | Yes | End-of-period ARR minus start-of-period ARR, from your subscriptions |
| Net burn | No | Your bank and accounting: total cash out minus cash in |
North Metric can't calculate the burn multiple, because it reads your billing and not your bank. It can give you the denominator. ARR is MRR × 12, and MRR movement shows what made up the change: new, expansion, reactivation, contraction and churn.
What to do with your number
- Look at the denominator first. Net new ARR drops when churn rises, even if sales stay the same. A quarter where the burn multiple jumps is often a churn quarter.
- Split net new ARR into its parts with MRR movement. If churn and contraction eat most of new MRR, retention is the cheaper fix.
- Watch net revenue retention. Expansion from existing customers adds ARR with little extra spend, which pulls the multiple down.
- Compare quarters, not single months.
What North Metric does here, and what it costs
North Metric reads your Stripe account and ranks paying customers by money at risk. The free churn report shows the risk; the plan drafts the moves and waits for your approval.
| Free churn report | Plan | |
|---|---|---|
| Price | $0 | $199 a month or $1,990 a year; $109 a month for 12 months under $30K MRR |
| Customers at risk | The first 5 to act on, each with a plain reason | Every paying customer ranked by money at risk, with a 0–100 risk score |
| Failed payments | Money at risk from failed payments | Card-update notes drafted once Stripe’s retries end |
| Cancellations | Top cancel reasons, in Stripe’s own reason names | Save emails and win-back notes drafted in your name |
| Usage drops | — | From PostHog or Stripe Meters, matched to how each customer pays |
| Changes in Stripe | None | Only the ones you approve, or rules you set with limits |
| Proof | Your last 12 months replayed: who would have been flagged | Kept MRR tied to invoices, measured against a comparison group |
| How to start | Connect Stripe; refreshed monthly by email | 14-day trial, no card |
Prices as listed on the pricing page.
Fair questions.Straight answers.
David Sacks called 2x reasonable for an early-stage startup and 5x terrible, and read 3x or more as a warning sign. Lower is better, and a falling multiple over several quarters matters more than any single reading.