Capital efficiency

Burn multiple: the growth metric you cannot improve by spending more

By Assaf Schwartz · 10 September 2026 · 10 min read

Every other efficiency metric has a denominator you can argue with. CAC depends on which costs you decide to load into it; LTV depends on a churn assumption; payback depends on both. Burn multiple has a numerator that comes straight off the bank statement, and that is the entire reason investors reach for it first.

What the ratio is

burn multiple = net burn ÷ net new ARR, measured over the same period — usually a quarter, sometimes a trailing year. Net burn is the cash that actually left the business. Net new ARR is the recurring revenue you ended the period with minus what you started it with, after churn and downgrades.

A multiple of 1.0x means you burned a dollar for every dollar of durable revenue you added. At 2.0x you burned two. The number is a price: it tells you what a dollar of ARR currently costs you, in cash, all-in — sales, engineering, support, the office, everything. There is no allocation step, no attribution model and nothing to categorise, which is precisely why it is so hard to flatter.

Two properties make it more useful than it looks. First, it is symmetrical: both halves respond to management. You can improve it by burning less or by adding more ARR, and the ratio does not care which. Second, it is self-punishing. Anything you do that increases spend has to produce more than a proportional increase in net new ARR, or the number gets worse. Growth bought with cash shows up in the numerator immediately and in the denominator only partly, and only later.

How this model computes it

The simulator reports one burn multiple for the whole five-year run, and it is worth being precise about what that figure is: peak monthly net burn × 12 ÷ ARR added over the horizon. Peak net burn is the worst single month in the projection; ARR added is ending ARR minus the ARR the business held in month one.

On the default input set that resolves to a peak burn of $18,036, reached in month 1, against ARR growth from $1.5M to $3.4M — $1,984,547 added, for a headline multiple of 0.11x.

A proper quarterly burn multiple differs on three things. It uses the burn that actually occurred in the quarter rather than a peak annualised; it matches numerator and denominator to the same window, so an early quarter is not being rewarded for ARR added three years later; and it produces a series rather than a single number, which is what you want, because the trend matters more than the level. On the default company that series runs from 0.99x in the first quarter down through break-even in month 24 — the model burns in 23 of its 60 months, $223,654 in total, which against the same ARR added is 0.11x cumulative.

Read the site's figure as a summary of the whole path at its worst point rather than a current quarter: when the burn profile is flat or rising the two converge, and when it is front-loaded the horizon figure understates what an investor would calculate today.

Why spending more cannot help

This is the property that makes the metric worth reporting. Take the default company and pay more for each customer — raise CAC and change nothing else. The cost base is untouched, so the numerator does not move at all. The denominator does.

Quadrupling CAC from $3,600 to $14,400 stretches payback from 11 months to 44 and removes $398,604 of ARR from the five-year path. The burn multiple moves from 0.11x to 0.14x. It gets worse, because the same dollars bought less revenue.

Now spend on the cost base instead. Raise Engineering headcount and watch both halves of the ratio move in the wrong direction at once.

Going from 8 engineers to 16 multiplies peak burn by a factor of 6.5 and reduces ARR added by $496,095, because operating profit is what funds acquisition in this model and there is none left. The headline multiple goes from 0.11x to 0.95x; the cumulative version, which is the honest one here, goes to 4.08x. The business never breaks even and is out of cash by month 7.

Five companies, five multiples

The five saved scenarios on this site span the range, and they are a good demonstration of what the ratio does and does not capture.

The bootstrapped scenario reports 0.00x for the honest reason: it never burns a month, so the numerator is zero while it adds $1,813,536 of ARR. The churn crisis reports 0.00x for a dishonest one — it burns $214,212 in its worst month while ARR falls by $662,560, and the engine's guard returns zero rather than a negative ratio.

Between those poles, the seed-stage scenario runs 0.25x and the Series A scale-up 0.10x: both efficient on the horizon measure, but only one of them still has cash at the end. The seed company runs out in month 18. Efficiency and solvency are different questions, and the burn multiple only answers the first.

The bands, and what they mean

These are the bands in general use, and they are quarterly figures for a company past roughly $1M of ARR. They mean very little for a business in its first few quarters, when the denominator is small and volatile.

The number is most informative as a trend. A company moving from 2.5x to 1.6x over four quarters is telling a far better story than one sitting flat at 1.4x, because the first has demonstrated that its spending responds to management and the second has not been tested.

Payback, churn and the denominator

Burn multiple is not an independent measurement. It is a compressed statement about acquisition efficiency and retention, which is why it works as a cross-check on both.

Its relationship to CAC payback

CAC payback governs how much of the acquisition spend comes back inside the measurement window. An acquisition cost that repays in 11 months has returned most of a quarter's spend within the year; one that takes 44 months returns almost none of it, so the cash sits in the numerator quarter after quarter while the ARR it bought arrives once. If your reported CAC looks excellent and your burn multiple is above 2x, the burn multiple is the number to believe.

Its relationship to churn

Churn attacks the denominator directly, because net new ARR is net. Hold everything else constant on the default company and move Monthly revenue churn: at 1.8% a month the model adds $7,260,646 of ARR over the horizon against a peak burn of $17,166, a 0.03x multiple. At 4% the same acquisition engine adds $0 — gross additions are being consumed as fast as they arrive — and peak burn rises to $44,381. Nothing about sales changed. The retention rate alone decided whether the spend produced anything.

What to do above 2x

Above 2x, the instinct is to cut costs, and that is usually the second-best answer. Work the list in this order, because the first two items are the ones that change the ratio without shrinking the company.

• Fix retention before acquisition. Every point of revenue churn is a reduction in the denominator that you are paying full acquisition cost to replace. It is the cheapest available improvement and the one with the longest lead time, which is why it needs starting first.
• Raise price. It moves gross profit per account immediately, shortens payback, and adds to net new ARR without adding a dollar of burn. In a business with real retention it is the single fastest lever on this ratio.
• Cut the spend that is not producing ARR. Not a flat percentage across the organisation — that damages the denominator as much as the numerator. Find the channels, segments and product lines whose net new ARR does not cover their fully-loaded cost, and stop those specifically.
• Slow hiring before you cut it. Payroll is the largest and stickiest part of most cost bases, and headcount growth committed today shows up in the numerator for years.
• Measure it quarterly and publish the series. A single number invites argument about the period. Four consecutive quarters do not.

For the retention half of the denominator, see NRR vs. GRR, which separates the revenue you keep from the revenue you expand into. For the acquisition half, see CAC payback vs. LTV:CAC — payback is the metric that determines how much of your spend is still sitting in the numerator when the quarter closes.

Educational content, not financial advice. Figures here are illustrative and exclude taxes, financing and one-off items. Model your own numbers in the simulator and check them with a qualified accountant before acting.