How do I measure sales efficiency at different ARR scales?

Question

Pulse RevOps · The Machine · Accepted Answer

Sales efficiency is not one number — it's a different engine at every ARR scale, and applying the wrong yardstick is how boards burn good operators. The single equation behind everything is CAC payback = (S&M spend in period) ÷ (New ARR × gross margin) — but the *target* moves with stage, and what you should also be tracking moves with it.

Stage-by-stage targets (2026 benchmarks, public + private SaaS)

Stage	CAC Payback	Magic # (qoq)	Gross Margin	NRR	Win Rate
<$5M ARR	18–24m	n/a (too volatile)	65–75%	95–105%	15–20%
$5–20M	12–18m	0.5–0.75	72–78%	105–115%	20–25%
$20–50M	9–12m	0.75–1.0	75–80%	110–120%	25–30%
$50M+	6–9m (best-in-class <12m public median)	0.7–1.0	78–82%	115–125%	30%+

Public-SaaS medians from Bessemer's *State of the Cloud 2026* and Meritech's BVP Cloud Index sit near 18-month CAC payback and ~0.6 magic number at the median, with top-quartile 12 months / 1.0+. KeyBanc's 2025 SaaS Survey of ~400 private companies puts private median CAC payback at 28 months — i.e. *most* private SaaS is below the public benchmark, which is why your private-stage targets above are stricter than what's actually shipping.

**The mechanics change by stage — what you are *actually* measuring**

Below $5M ARR — founder-led, idiosyncratic. Blended efficiency is noise. ACVs swing 5x deal-to-deal, sample size is too small for magic number to mean anything. Track: per-rep quota attainment, win rate by source, and the founder's *capacity ceiling* (hours/week in deals × close rate). Once founder selling tops out, the next $1M ARR is the most expensive ARR you'll ever buy. This is the danger zone.

$5–20M ARR — first repeatable motion. Magic number becomes meaningful: net new ARR ÷ prior-quarter S&M spend (the SaaStr / Scale Studio definition). 0.75+ = healthy, scale capital. 0.5–0.75 = workable but watch payback. <0.5 = the engine is broken; do not pour fuel on it. CAC payback should compress as inbound builds — if it isn't compressing, your motion isn't repeatable.

$20–50M ARR — land + expand split. This is where blended numbers start to lie. Split land CAC (new logos) from expansion CAC (cost-to-expand existing). Expansion typically pays back in 3–6 months because there's no acquisition cost. If blended payback looks fine but land payback alone is >18 months, your new-logo motion is dying and expansion is masking it. ICONIQ's *State of SaaS* tracks this as "land economics" vs "blended economics" — they diverge most at $20–50M.

$50M+ ARR — net new is a rounding error, retention is the engine. At this scale, NRR > 120% outperforms a higher-NRR-but-bigger-CAC peer on every long-run efficiency curve. Track: gross retention (the floor — must be >90% in mid-market, >95% enterprise), expansion ARR mix, and Rule of 40 (growth + FCF margin ≥ 40). Bessemer's data shows the top public quartile in 2025–2026 hit Rule of 40 around 45–55, mostly via NRR + gross-margin discipline, not via outbound efficiency.

Bear Case — where this framework breaks

The honest counter: CAC payback and magic number are gameable. Three failure modes I've watched in the field:

Discount-funded magic number. Reps close at -40% to hit quota, ARR booked looks great, payback shortens because S&M didn't change — but gross margin (and NRR-at-renewal) collapses 4 quarters later. The metric was right in the moment, the company was wrong. Always pair magic number with net price trend.
Capitalized-S&M trick. Some companies amortize S&M over the cohort's expected life rather than expense it in-period; payback "improves" by 30–40% with zero operational change. Cross-check the cash burn statement vs the P&L S&M line.
Mix shift hides a broken segment. Blended CAC payback at 11 months sounds healthy until you split it: SMB at 7 months, enterprise at 24 months, and you're tilting top-of-funnel toward enterprise. Six quarters later you've quietly become a 24-month-payback company. Ben Murray (the *SaaS CFO*) has written this exact pattern up — segment your efficiency or it will lie to you.

The framework is also lagging. CAC payback measured today reflects pipeline you generated 6–12 months ago. If macro turns, the metric won't tell you for two quarters. Pair efficiency with leading indicators: SQO conversion rate, average deal cycle length, win rate by segment.

Action — what to actually do this week

Pick the row that matches your ARR. That's your only valid benchmark.
Decompose blended CAC payback into land and expansion if you're past $20M. If you've never done this, the answer will surprise you (it usually surprises CFOs first).
Sanity-check magic number against net price realization and gross-margin trend — if they're diverging, your efficiency improvement is fake.
Set the next-stage target and *stop optimizing for the current stage* — efficient operators get to the next stage's metric set 1–2 quarters before they need to.

Cross-links — go deeper

[/knowledge/q98](/knowledge/q98) — What's the right CAC payback target — 12, 18, 24 months?
[/knowledge/q99](/knowledge/q99) — How is the Rule of 40 actually computed and why does it matter?
[/knowledge/q100](/knowledge/q100) — What's a good magic number for a public SaaS company?
[/knowledge/q103](/knowledge/q103) — How do I track burn multiple alongside efficiency metrics?
[/knowledge/q105](/knowledge/q105) — How do I calculate LTV when expansion is meaningful?
[/knowledge/q96](/knowledge/q96) — What's a good NRR for Series B SaaS in 2026?
[/knowledge/q106](/knowledge/q106) — What's the right ARR-per-employee benchmark for efficient SaaS?
[/knowledge/q91](/knowledge/q91) — What's a realistic CAC payback for SMB vs mid-market vs enterprise?

graph LR A[ARR Scale] --> B{$1-5M} A --> C{$5-20M} A --> D{$20-50M} A --> E{$50M+} B --> B1[CAC Payback 18-24m Per-rep attainment] C --> C1[Magic # 0.5-0.75 Payback compresses] D --> D1[Split LAND vs EXPANSION CAC] E --> E1[NRR 115-125% Rule of 40 = 45-55] B1 --> X[Founder ceiling] C1 --> Y[First repeatable motion] D1 --> Z[Blended hides truth] E1 --> W[Retention is the engine] Z --> BC[BEAR CASE: discount-funded / capitalized S&M / mix shift] W --> BC

TAGS: cac-payback, magic-number, arr-benchmarks, efficiency-metrics, saas-growth, rule-of-40, nrr, burn-multiple

How do I measure sales efficiency at different ARR scales?

What does the score mean?