Is my growth efficient? Burn multiple and magic number for Seed to Series B

1. Blog
2. Measurement & Budget

By Saroj Jha · November 2, 2026 · 13 min read

Your growth is efficient if you are adding net new ARR at a cost the stage can justify, and two ratios tell you. The burn multiple (net burn divided by net new ARR) measures the whole company: David Sacks, who named it, suggests a seed company might sit around 3 and a post-Series A company around 2^[1]. The magic number (new annualized revenue divided by the prior quarter’s sales and marketing spend) measures only the go-to-market engine. Read them together, check which definition a benchmark used before you compare, and never let either one stand in for retention or cash runway.

Here is how to calculate both, what the published numbers really measure, and where each ratio misleads.

Running case: Example Co. Fictional

Example Co. is a fictional Seed-stage B2B SaaS company selling scheduling software to multi-site clinics. Last quarter it grew ARR from $2.4M to $3.0M, adding $750K of new and expansion ARR and losing $150K to churn and downgrades. It burned $1.5M of cash in the quarter, recognized $690K of revenue (up from $560K the quarter before), and spent $700K on sales and marketing in the prior quarter. The board has asked a simple question: is this growth efficient enough to raise a Series A on?

What is the burn multiple, and how do you calculate it?

Sacks introduced the burn multiple in April 2020 as net burn divided by net new ARR^[1]. It answers one question: how many dollars of cash does the company burn to add one dollar of ARR? He describes it as flipping Bessemer’s Efficiency Score (net new ARR divided by net burn) so that a higher number reads as worse, and he prefers it to ratios built on total capital raised because it judges only the latest period rather than “permanently” penalizing early mistakes^[1].

Sacks does not fix the period: the metric can be read for “any given month, quarter, or year”, and his worked example uses a quarter in which a startup burned $2M and added $1M of ARR, a 2x burn multiple he calls “reasonable for an early-stage startup”^[1]. He calls 5x “terrible”^[1]. In my view the quarter is the right default at Seed: a month is too noisy, and a year hides the trend the board is asking about.

Two details decide whether your number is honest. Net burn is cash out minus cash in for the period, so it includes engineering, G&A and hosting, not just go-to-market. Net new ARR is new plus expansion minus churn and contraction, so a leaky base raises the multiple even when sales is doing well.

Example Co., step 1 Fictional

Net new ARR is $3.0M minus $2.4M, which is $600K ($750K new and expansion minus $150K churn and contraction). Burn multiple is $1.5M divided by $600K, which is 2.5x. Example Co. burns $2.50 to add each dollar of ARR.

Do this

Calculate your burn multiple for each of the last four quarters, using net new ARR after churn, and look at the direction before the level.

Which burn multiple benchmarks can you trust?

Most burn multiple “benchmarks” online are a rules-of-thumb table passed from post to post with no sample behind it. Sacks himself frames his guidance as rules of thumb for venture-stage startups, and his stage markers are illustrations, not survey results: a seed company “might” be at 3, and after the Series A the multiple “might drop to 2”, improving toward zero as the company matures^[1]. I use those as reference points, not as targets.

Two investor datasets disclose who is in them, and both use a different numerator from Sacks. ICONIQ’s State of Software 2025 summarizes quarterly data from 127 software companies, including its own portfolio, from 2013 to the second quarter of 2025^[2]. It defines the burn multiple as free cash flow over net new ARR, for unprofitable companies only, and reports a median of 3.0x in 2022, 1.9x in 2023 and 2024, and 1.7x in 2025 for companies under $100M ARR^[2]. The 2025 figure rests on 42 company observations against 99 for 2022, and the companies are mostly later-stage than a Seed startup, so I read it as a later-stage reference point, not a Seed norm.

Scale Venture Partners published burn multiples by ARR band in 2022 from its dataset of “several hundred SaaS companies, private and public”, using operating income as the burn figure because it is “reported more accurately and consistently than cash burn”^[3]. Companies with $0M to $1M of ARR had a burn multiple of 3.4, those with $25M to $50M had 1.4, and top-quartile growers at $2.5M to $5M ran about 1.8^[3]. Scale’s summary is that the typical SaaS startup burns $1.60 for every $1 of net new ARR from seed to IPO^[3].

Example Co., step 2 Fictional

At 2.5x, Example Co. is better than Sacks’s seed reference point of 3, but worse than the roughly 1.8 Scale found for top-quartile growers in its own $2.5M to $5M ARR band, and worse than the 1.7x ICONIQ median for 2025. For a Series A story, the question is not whether 2.5x is “good” but whether it is falling quarter by quarter.

Do this

Before quoting a benchmark to your board, write down its numerator (net burn, free cash flow or operating income) and recalculate your own multiple the same way.

What is the magic number, and how is it calculated?

The magic number asks a narrower question: for each dollar spent on sales and marketing last quarter, how much new annualized revenue arrived this quarter? Its origin is often credited to different people and I have not found a first publication I can confirm, so I will define it without attribution. An early public write-up, a guest post from Will Price’s blog reposted on Hummer Winblad Venture Partners’ site (dated 2008 in its address), which presents the metric from a talk by Omniture’s chief executive Josh James, gives the classic formula: this quarter’s revenue minus last quarter’s revenue, times four, divided by last quarter’s sales and marketing expense^[4]. The same post says that below 0.75 you should “step back and look at your business”, that above 0.75 the business is primed to turn spend into growth, and that above 1.5 the author wants a call^[4].

The lag is the point. Spend in one quarter produces revenue in the next, so the denominator is the prior quarter. ICONIQ uses a “net magic number” built on ARR instead of revenue: current-quarter net new ARR divided by prior-quarter sales and marketing expense, and it notes that which quarter of spend to use “should depend on a given company’s sales cycle”^[2]. If your typical deal takes six months to close, a one-quarter lag credits this quarter’s wins to spend that did not cause them.

Example Co., step 3 Fictional

Revenue grew from $560K to $690K, a $130K increase, or $520K annualized. Divided by $700K of prior-quarter sales and marketing, the revenue magic number is 0.74. On ARR, the net magic number is $600K divided by $700K, which is 0.86. One definition sits just under the 0.75 line in the 2008 post and the other sits above it.

Do this

Pick one definition, write it into your board metrics glossary, and do not switch definitions in the quarter the number looks better.

Which magic number benchmarks have a sample behind them?

Few do. ICONIQ reports top-quartile net magic numbers by ARR band: 2.3x under $10M ARR and 1.5x from $10M to $25M, from the same portfolio-weighted dataset of 127 companies^[2]. Those are top-quartile figures, not medians, so a company below them is not necessarily below average.

ICONIQ also cut the top quartile by deal size: 1.7x for annual contract values under $10K, falling to 1.1x at $50K to $150K and above^[2]. That matches what I would expect: larger deals take more sales effort per dollar and longer to close, so the ratio falls even when the business is healthy. In my view, comparing a mid-market sales-led company with a product-led one on this number is an easy mistake and a misleading one.

Example Co., step 4 Fictional

Example Co. sells annual contracts of about $20K to clinic groups. Against ICONIQ’s 1.5x top quartile for $10K to $50K deals, its 0.86 net magic number says there is room to improve go-to-market efficiency, but it does not say spend should stop.

Do this

Compare your magic number only with benchmarks cut by a similar contract size and sales motion, and treat top-quartile figures as a ceiling, not a pass mark.

Why can the two metrics disagree, and what does each one miss?

They disagree because they cover different costs. The burn multiple counts every dollar the company burns, so a heavy engineering roadmap or a new office raises it even if sales and marketing are working. The magic number counts only sales and marketing spend, so it can look healthy while the company burns cash elsewhere. When one is good and the other is bad, the gap tells you where to look.

Each also has blind spots of its own. The burn multiple uses cash, so annual prepayments collected up front lower net burn and flatter the ratio in the quarter they land; Scale’s benchmarks avoid this by measuring burn as operating income^[3]. It does not work at all before revenue: Sacks notes that if net new ARR is zero, “the ratio will not even compute”^[1]. And it treats every dollar of ARR as equal, whether it is a three-year contract or a pilot that may not renew.

The magic number has the opposite problems. It ignores gross margin, so a product with expensive onboarding or inference costs can show a good ratio and still lose money on each customer. It credits all new revenue to sales and marketing, including deals the founder closed whose cost sits in G&A. The 2008 post lists the main drags as wasted spend, high churn and market problems such as saturation or competition^[4], which means a falling number tells you something is wrong without telling you which.

Example Co., step 5 Fictional

Example Co. collected $300K of annual prepayments in the quarter beyond what it recognized as revenue. Without that cash, its burn would have been $1.8M and its burn multiple 3.0x instead of 2.5x. Applying a 75% gross margin (fictional) to its $600K net new ARR gives a margin-adjusted net magic number of 0.64 rather than 0.86.

Do this

Show the board the burn multiple on both cash burn and operating loss in any quarter with large prepayments, and footnote which one you use.

Which lever moves the burn multiple fastest?

Once you know the number, the useful question is which input to change. Because net new ARR is new business minus churn, retention work moves the burn multiple directly, which is not true of simple CAC payback. Cutting burn works too, but it usually cuts the spend that produces next quarter’s ARR.

My rule of thumb, not a benchmark: if the magic number is healthy and the burn multiple is not, look at non-go-to-market spend first; if both are weak, look at churn before cutting acquisition. AAJ’s LTV:CAC article walks through why single-lever fixes rarely close a big gap, so I will not repeat it here.

Example Co., step 6 Fictional

To reach 2.0x, Example Co. could cut burn by $300K to $1.2M, or add $150K of net new ARR to reach $750K. Halving churn and contraction from $150K to $75K alone lifts net new ARR to $675K and brings the multiple to 2.22x without any new spend.

Do this

Model one quarter forward with three versions (less burn, more new ARR, less churn) and present the one you can deliver without hurting the next quarter.

What changes by stage?

• Pre-seed — With little or no ARR, the burn multiple will not compute or will swing wildly^[1]. Track cash runway and learning milestones instead, and start logging sales and marketing spend by quarter so the magic number has history later.
• Seed — Sacks’s reference point is around 3^[1], and Scale’s earliest ARR band ran at 3.4^[3]. What matters more is the direction over four quarters and whether net new ARR is held back by churn.
• Series A–B — Investors will compare you with later-stage medians such as ICONIQ’s 1.7x for 2025^[2]. Expect to show both metrics, a consistent definition, and the magic number cut by segment or contract size.

Free tool

Check both ratios on your own quarter

The SaaS Growth Efficiency Calculator takes one quarter’s revenue, prior-quarter sales and marketing spend, net burn and net new ARR, and returns your magic number, burn multiple, Rule of 40 score and CAC payback against each metric’s cited thresholds.

If your efficiency problem turns out to be acquisition cost or payback, read My LTV:CAC is below 3:1. Which number should I fix first?. If the question is how much to spend in the first place, How much should a startup spend on marketing? sets ranges by stage, and the Unit Economics Calculator checks LTV, CAC and payback on your own numbers.

If you want me to look at your own funnel, positioning and channels, my Growth Audit ends with three prioritized fixes and a 30-day action plan.