How much new annual recurring revenue does each dollar of sales and marketing generate?
№ 015 · Magic number
Definition
MetricNet new ARR in quarter, annualized, per dollar of prior-quarter S&M spend
Unitx (dollars of new ARR per S&M dollar)
Net new ARR divided by previous quarter's sales and marketing spend. Above 0.75 means the motion is fundable. Pull S&M expense from the P&L and net new ARR from billing each quarter, then divide. Example: $1.2M S&M produced $1.05M new ARR = 0.88x (healthy).
- ARR
- Annual recurring revenue. MRR at period end multiplied by 12, or the annualised value of active contracts.
- S&M
- Sales and marketing. All go-to-market cost: salaries, commission, media spend and tooling.
- P&L
- Profit and loss statement. Revenue minus cost of goods sold minus operating expenses, down to operating profit.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| Under 0.68x (below the 0.75 minimum investment floor) | 1.37x on CY2025 actuals (a sharp improvement year; longer-run medians sat below 1.0x) | Above 1.0x sustained |
Private B2B SaaS and AI-native software, CY2025, one-quarter-lag convention (current-quarter net new ARR x4 over prior-quarter S&M); n=132 reporting · The 2025 median crossing 1.0 for the first time in the four-year trend partly reflects the AI-native mix in the sample; traditional VC-backed SaaS still hovers near the 0.75 floor. High-ACV enterprise books show volatile quarterly magic numbers because deal timing dominates; use trailing 2-3 quarters. Gross vs net magic number (whether churn is netted) changes the reading; the figures here are net.
Category split omitted: Magic-number splits by segment or motion are published by single sources only; no two independent Tier 1-2 sources cover the same split.
When it looks bad
The bar chart oscillates around 0.5-0.7 for three or more consecutive quarters with S&M spend rising, meaning each incremental GTM dollar reliably returns less than the floor and scaling spend is scaling the loss.
Q1 0.66, Q2 0.71, Q3 0.58 while quarterly S&M grew from $1.2M to $1.8M. The engine is being fed faster while running below the 0.75 line.
What to do about it
- Freeze incremental S&M below 0.75 for two consecutive trailing-2-quarter readings and audit the funnel stage by stage; the floor exists precisely because spend below it compounds losses (Aleph x Benchmarkit framework).
- Reallocate before cutting: move budget from the lowest-magic-number channel or segment to the highest; segment-level magic numbers (compute per line of business) routinely differ by 2x inside one company.
- Raise the numerator with expansion: net new ARR includes expansion, and expansion dollars cost ~$1.40 vs $2.00 for new logos (Benchmarkit 2025), so an expansion-weighted quarter mechanically lifts the magic number.
- For enterprise motions, judge on trailing 3 quarters and pipeline-adjusted basis; a single slipped seven-figure deal can swing the quarterly print by 0.4x and trigger wrong decisions.
Sources
- Aleph and Benchmarkit Median 1.37, above 1.0 for the first time in the four-year trend; bottom quartile 0.68; 0.75 is the investment floor getaleph.com ↗
- ICONIQ Growth (as summarized by SaaStr) Top-quartile companies at $25M ARR: 100% growth, 110% NRR, 0.9x net magic number, $230k revenue per employee saastr.com ↗
- ICONIQ Growth Net magic number among the efficiency metrics stabilizing as companies rewire operations iconiq.com ↗
- Fiscallion Stage expectations: 0.6-0.75 acceptable at Series A while refining ICP; trending 0.75+ at Series B; sustained 1.0+ table stakes at Series C fiscallion.io ↗
- Drivetrain PE-backed companies posting higher magic numbers on efficiency focus; VC-funded hovering near 0.75; a 0.8 magic number at 80% gross margin implies ~19-month CAC payback drivetrain.ai ↗