What Is the SaaS Magic Number? Formula, 2026 Benchmarks, and When to Spend
The SaaS Magic Number is annualized net new recurring revenue divided by prior-period S&M. Benchmarkit's 2026 private median is 1.37; Leckie's floor is still 0.75. Formula, bands, and a 30-day test.
The SaaS Magic Number answers one board question: for every dollar you spent on sales and marketing last period, how many dollars of annualized new recurring revenue did you create this period? It is a GTM efficiency ratio, not a growth trophy and not a valuation multiple. Score above 0.75 and the classic rule is to keep investing. Score above 1.0 and each S&M dollar is returning more than a dollar of new ARR. Score below 0.75 and more spend on the same motion usually makes the problem worse.
Growth Centr publishes evergreen, research-backed analysis on SaaS unit economics for founders, marketers, and operators who need a spend signal they can calculate from the P&L, not a vanity KPI that only works in a pitch deck.
Lars Leckie popularized the metric in March 2008 on Will Price's blog, after Omniture's Josh James described how the company decided when to "pour on the gas." The formula has barely moved since. What moved is the market print: Benchmarkit's 2026 SaaS and AI-Native Metrics (Aleph × Benchmarkit, CY-2025 actuals, 342 B2B SaaS and AI-native companies; Magic Number n = 132) puts the private-company median at 1.37, up from 0.94 the year before — the first time that four-year series has cleared 1.0. The bottom quartile still sits at 0.68. The bar did not move. The typical reporter finally cleared it.
Key Takeaways
- Magic Number = (current-period recurring revenue − prior-period recurring revenue) × 4 ÷ prior-period sales & marketing expense. Lag the S&M. Do not divide this quarter's growth by this quarter's spend.
- Leckie's original bands still hold: below 0.75 = audit before you spend; 0.75–1.0 = efficient enough to keep investing; above 1.0 = lean into growth; above 1.5 was his "call me" zone.
- Benchmarkit CY-2025: private B2B SaaS median 1.37 (from 0.94), bottom quartile 0.68, 75th percentile 2.14. Companies growing >50% print a 2.40 median; the 11–30% growth band often sits below 0.75.
- It is roughly the inverse of CAC payback: high Magic Number, short payback. Pair both before you add headcount. Benchmarkit median payback is still 16 months.
- Churn and contraction hit the numerator. A leaky gross revenue retention print will look like a sales problem on this score even when the reps are fine.
How to Calculate the SaaS Magic Number
Magic Number = (QRev[X] − QRev[X−1]) × 4 / ExpSM[X−1]
That is Leckie's 2008 formula. QRev is quarterly recurring revenue (or ARR snapshots converted to the same basis). ExpSM is total sales and marketing expense in the prior quarter — the spend that produced this quarter's net new recurring revenue, not the spend you are booking while the deals close.
Worked example. Q1 recurring revenue $1.0M, Q2 $1.2M, Q1 S&M $800K → ((1.2 − 1.0) × 4) / 0.8 = 1.0. Q2 $1.2M, Q3 $1.5M, Q2 S&M $900K → ((1.5 − 1.2) × 4) / 0.9 = 1.33. Same company Leckie walked through in 2008; the arithmetic has not changed.
Two conventions that break the number if you mix them:
| Input | Clean version | Distortion risk |
|---|---|---|
| Numerator | Net change in recurring revenue (new + expansion − churn − contraction), annualized | Gross bookings ignore churn; GAAP total revenue mixes services and one-time items |
| Denominator | Prior-period S&M (fully loaded: salaries, commissions, ads, tools, agencies) | Same-period S&M flatters a ramp and punishes a cut; excluding demand gen understates cost |
| Lag | One quarter for most SaaS; two–three for long enterprise cycles | A nine-month sales cycle will understate efficiency on a one-quarter lag |
| Basis | ARR or recurring revenue, held constant | Mixing ARR growth with GAAP S&M against a peer's GAAP revenue Magic Number is not a peer |
Private operators who can see net new ARR should use that as the numerator and still lag S&M. Public comps usually substitute GAAP subscription (or total) revenue because that is what sits on the income statement — Scale Venture Partners defended that GAAP version precisely because it is comparable and hard to fudge. Label which version you are printing.
What never enters cleanly. One-time professional services. A mid-year price increase annualized as if it had run for twelve months. New-logo bookings that are not yet revenue. Stock-based compensation debates belong in a margin conversation, not a Magic Number rewrite — pick a definition and hold it for trend analysis.
2026 Benchmarks: What "Good" Actually Means
Leckie's 2008 rule of thumb and Benchmarkit's 2026 distribution answer different questions. Keep both.
Classic bands (Leckie / board vernacular):
| Magic Number | What boards usually do |
|---|---|
| < 0.5 | Model is not working. Fix targeting, conversion, or product before adding quota. |
| 0.5–0.75 | Investable but inefficient. Incremental spend buys expensive growth. |
| 0.75–1.0 | Healthy. Keep investing; tighten the motion. |
| ≥ 1.0 | Strong. Each S&M dollar returns ≥ $1 of annualized new ARR. |
| ≥ 1.5 | Exceptional on the original scale — rare, and usually temporary once you reinvest. |
Private B2B (Aleph × Benchmarkit, June 2026, CY-2025). Median 1.37. Bottom quartile 0.68. 75th percentile 2.14. The median crossed 1.0 for the first time in the four-year series; the bottom quartile is still below Leckie's floor. The spread between the 75th and 25th percentile is now large enough that "hitting the median" is not a soft middle — it is already the safer side of a bifurcating market (Aleph's write-up of the same book).
By growth rate, same book:
| 2025 growth rate | Median Magic Number | Read |
|---|---|---|
| >50% | 2.40 | Efficiency and growth coexist. Even the bottom of this cohort clears 0.75. |
| 11–30% | Below 0.75 | Spending without proportional return. Reset the motion, do not add budget. |
That split matters for planning. A mid-growth company below 0.75 is not "almost efficient." It is funding a motion that does not pay for itself, which is the same trap that shows up as a low Rule of 40 score when you add S&M to chase the growth half.
Public comps. Scale VP's original public heuristic treated >1.0 as compelling, <0.5 as unproven, and the middle as capital-inefficient but possibly workable. Public Magic Numbers run on GAAP revenue and a different lag structure than a private ARR book. Do not grade your Series B dashboard against a public GAAP print without saying so.
Magic Number vs CAC Payback vs Rule of 40
These three metrics travel together. They are not substitutes.
- Magic Number asks: how much annualized new ARR did last period's S&M buy?
- CAC payback asks: how many months until that customer (or cohort) pays back acquisition cost? Benchmarkit's 2026 median is 16 months; blended CAC ratio is $1.30 of S&M per $1 of ARR (full CAC breakdown).
- Rule of 40 asks: is the growth-plus-margin mix board-acceptable? Private median 25%, top quartile 43%.
A high Magic Number with a long payback can still strain cash — you are efficient on an annualized ARR basis but slow to recover cash. A short payback with a weak Magic Number usually means you are buying expansion cheaply while new-logo S&M is broken (or the reverse). The Benchmarkit guidance is blunt: when Magic Number, CAC payback, NRR, and GRR are all top-quartile, accelerating GTM spend lifts growth and the Rule of 40. When they are not, more spend mostly erodes both.
Retention is the silent input. GRR fell from 88% to 84% in the same 2026 book. Expansion already supplies 40% of net new ARR at the median. If churn and contraction eat the numerator, your Magic Number falls even when win rates look fine. Fix the bucket before you hire another AE. Customer lifetime value and LTV:CAC (Benchmarkit median CLTV:CAC 4.1×) tell you whether the customers you do keep are worth the spend the Magic Number is measuring.
A product-led motion can print a strong Magic Number with thin S&M — or a weak one if self-serve growth is mostly tire-kickers who churn before renewal. Read PLG efficiency next to GRR, not next to signups.
When to Spend, When to Stop
Use the Magic Number as a spend gate, not a goal to optimize directly.
- Below 0.75 for two consecutive quarters. Freeze net-new S&M headcount. Audit ICP, conversion by stage, ramp time, and win rate. More budget on a broken motion compounds the loss.
- 0.75–1.0 with stable retention. Keep investing. Tighten CAC mix toward expansion (Benchmarkit expansion CAC ratio $0.80 vs new-name $1.63) before you scale brand spend.
- Above 1.0 with top-quartile payback and NRR ≥ 100%. This is the classic "pour on the gas" zone. Cap the ramp so you do not crash the lag: a sudden S&M spike will tank next quarter's Magic Number even if the hires are good.
- Enterprise cycles >6 months. Recalculate with a two- or three-quarter S&M lag for your own trend. Still report the one-quarter version so the board can compare to Benchmarkit, but decide spend on the lag that matches your sales cycle.
Do not chase 1.5 by cutting S&M to the bone. A Magic Number that rises because you stopped hiring while ARR still lands from last quarter's pipeline is a one-quarter illusion. Trend it on a rolling four-quarter view.
A 30-Day Operator Test
Week 1. Lock definitions: recurring revenue basis, fully loaded S&M, one-quarter lag. Pull the last eight quarters.
Week 2. Compute Magic Number, CAC payback, GRR, and NRR on the same periods. Flag any quarter where bookings were used instead of net ARR.
Week 3. Segment: new-logo vs expansion contribution to the numerator; paid vs organic contribution to the denominator if you can. Identify the one lever that would move the score 0.2 points without adding headcount.
Week 4. Write the spend rule for next quarter in one sentence ("if Magic Number ≥ 1.0 and payback ≤ 16 months, add X AEs; else freeze"). Put it in the board pack next to Rule of 40. Revisit after the next closed quarter — not after every pipeline review.
FAQ
What is a good SaaS Magic Number in 2026?
Above 1.0 is strong. 0.75–1.0 is the classic invest zone. Below 0.75 means audit GTM before adding spend. The Benchmarkit private median is now 1.37, so clearing 1.0 is no longer rare among reporters — but the bottom quartile at 0.68 shows plenty of companies still fail Leckie's floor.
How do you calculate the Magic Number?
Subtract prior-period recurring revenue from current-period recurring revenue, multiply by four to annualize, and divide by prior-period sales and marketing expense. Lagging the S&M is the whole point.
Is the Magic Number the same as CAC payback?
No. Magic Number is new ARR per S&M dollar (efficiency). CAC payback is months to recover acquisition cost (cash timing). They usually move together, but a high Magic Number with slow collections or annual upfront billing quirks can still stress cash.
Why did the median jump from 0.94 to 1.37?
Aleph × Benchmarkit attribute the 2025 efficiency print to GTM rationalization, not a sudden demand boom — the same cost discipline that lifted Rule of 40 from 15% to 25%. Blended CAC ratio improved to $1.30. Whether that is durable depends on whether retention holds; GRR is moving the wrong way.
Does churn affect the Magic Number?
Yes. The numerator is net change in recurring revenue. High logo or revenue churn shrinks (or reverses) the delta even if new bookings look healthy. That is why GRR and NRR belong on the same dashboard page.
Should PLG companies use a different formula?
Same formula. PLG companies often print high Magic Numbers because S&M is lighter — until paid acquisition or sales-assist ramps. Keep the lag and the recurring-revenue basis; do not invent a "PLG Magic Number" that excludes sales-assist cost.
Methodology
Primary sources for this article: Lars Leckie's March 2008 post on Will Price's blog (original formula and 0.75 / 1.5 bands); Scale Venture Partners' Magic Number Math note (GAAP public-company version); Aleph × Benchmarkit 2026 SaaS & AI Performance Benchmarks (published 1 June 2026; CY-2025 actuals; 342 B2B SaaS and AI-native companies; Magic Number from 132 reporters; median 1.37, bottom quartile 0.68, 75th percentile 2.14; >50% growth cohort median 2.40). Related Growth Centr benchmarks on CAC payback, Rule of 40, NRR, GRR, and LTV are cited where they share the same Benchmarkit book. Figures labeled "2026" refer to the report edition and planning year; underlying actuals are CY-2025 unless noted.
