What a healthy startup burn rate looks like
Founders ask this in dollars. Is eighty thousand a month too much? Is thirty thousand fine? The number on its own cannot answer it, because the same burn is reckless at one company and conservative at the one next door. What separates them is not the spending. It is what the spending buys.
Why the dollar figure cannot answer the question
Two companies each burn $200,000 a month. One is adding $250,000 of new annual recurring revenue a quarter. The other is adding $40,000. The first is compounding. The second is buying time and calling it growth. No burn figure, and no runway figure either, tells those two apart. Runway says how long you have. It does not say which direction you are pointed, which is the distinction in default alive or default dead.
This is also why the question gets answered badly so often. Search for a burn benchmark and you will find confident per-stage tables. Read three of them and the numbers disagree, none of them names a sample size, and the figures drift with whatever year the page was written. A founder who quotes one of those in a board meeting is quoting nothing.
The burn multiple
The metric that does answer it is the burn multiple, which David Sacks of Craft Ventures published on 23 April 2020. It is one division:
Burn multiple = net burn ÷ net new ARR
Net burn is cash out minus cash in. Net new ARR is new plus expansion minus churn. The result reads as a price: how many dollars you spent to buy one dollar of new recurring revenue. At 1.0x you spend a dollar to add a dollar. At 4.0x you spend a dollar to add twenty-five cents.
Sacks published it with tiers:
- Under 1x: amazing
- 1x to 1.5x: great
- 1.5x to 2x: good
- 2x to 3x: suspect
- Over 3x: bad
The word doing the work is suspect rather than bad in the 2x to 3x band. A company can sit there for a quarter for an honest reason: a sales team hired ahead of the revenue it will close, an enterprise deal that slipped a month. Sitting there for a year is a different statement, and it is usually a statement about the product rather than the spending.
What the sampled data actually shows
Be careful with per-stage burn multiple tables, including the ones that cite Sacks. His own stage figures, roughly 3x at seed and roughly 2x after a Series A, were offered as illustration rather than as a survey result, and they are frequently reproduced as though they were sampled.
There is no large public per-stage matrix. The most useful sampled figure available is from Lighter Capital, across 83 private B2B SaaS companies: a median burn multiple of 1.12x among the companies that were burning cash, with a bimodal spread rather than a tidy cluster. About half were above 1.0x and roughly a quarter were below 0.33x.
That shape matters more than the median. Companies are not distributed around a healthy middle. They tend to sit in one of two groups, and which group you are in is a more useful thing to know about yourself than your distance from an average.
Ignore burn per employee
Burn per head is the benchmark founders reach for most and the one worth trusting least. It is quoted widely and sourced almost nowhere, and even a correct figure would mislead, because it treats every headcount as the same purchase.
Compensation is the majority of operating expense at most early companies, so per-head burn largely measures your salary bands and your location. A company paying senior engineers in a high-cost market will always look worse per head than one paying juniors elsewhere, whatever either is achieving. The ratio you want is not spend per person. It is spend per dollar of new revenue, which is the burn multiple again.
What legitimately changes the answer
A services business and a product business should not be held to the same multiple. Services revenue arrives roughly in step with the cost of delivering it, so the ratio stays flatter and improves slowly. Product revenue costs a great deal up front and very little to serve afterwards, so a product company that is not bending the curve down over time is not behaving like a product company.
Timing distorts it too. Net new ARR lands in the quarter the contract signs, while the spending that produced it happened over the two or three quarters before. One bad quarter after a hiring push is arithmetic. A trend is a trend.
And it only works if you have recurring revenue to divide by. For a company that is pre-revenue, the burn multiple is undefined and the honest question reverts to the runway one: how many months, and what has to be true before they run out. That is the ground covered in how to calculate and extend your runway.
So what is healthy
Healthy is a burn multiple under 2x that is trending down, at a level of spend you can sustain for long enough to prove the next thing. Both halves are load-bearing. A brilliant multiple on a burn that leaves you four months of cash is not health, it is a good ratio in a bad position.
If the multiple is wrong, cutting is the second move rather than the first, and cutting badly makes the ratio worse by removing the capacity that produced the revenue. The order to cut in is the subject of cutting burn without killing momentum.
Make it a number you can see
The burn multiple is one division, and almost nobody computes it monthly. The reason is not the arithmetic. Net burn lives in a bank account, expenses in a spreadsheet, revenue in an invoicing tool, salaries in someone's head. Assembling those four takes an afternoon, which means it happens before a board meeting and not before a hiring decision.
That is the gap Klerky closes: it keeps burn, revenue and runway current from your real expenses, invoices and salaries, so the ratio is this month's rather than last quarter's.
Sources
Burn multiple definition and tiers: David Sacks, Craft Ventures, 23 April 2020. Sampled burn multiples: Lighter Capital, 83 private B2B SaaS companies. Stage figures attributed to Sacks are illustrative and are not survey results. Checked 31 August 2026.