Magic Number

A subscription business grows quarterly recurring revenue from $2.0 million to $2.3 million after spending $1.2 million on sales and marketing in the preceding quarter, and annualising the $0.3 million increase gives $1.2 million, so the ratio is 1.0 under this convention. The SaaS CFO describes that quarterly lag and cautions that the metric includes growth from existing customers, not just new business.

Bessemer's SaaS discussion treats sales efficiency as important but also uses other defined ratios, while Stripe's public guide, by contrast, labels CLV divided by CAC as a 'Magic Number', so that naming conflict makes formula disclosure essential. Use recurring revenue, since one-off implementation fees or hardware sales can make the numerator look strong without adding durable subscription value, and choose consistent accounting because recognised recurring revenue and annual recurring revenue bookings are different measures.

Set the quarter boundaries on a consistent basis, as calendar and fiscal quarters may differ, and apply the lag: preceding-quarter sales and marketing spend is a convention reflecting a short sales cycle, so longer enterprise cycles may need a longer horizon. Annualise correctly by multiplying the difference between quarterly revenue figures by four before comparing it with one quarter of expense, and do not multiply an already annualised ARR change again.

Define expense consistently, including sales salaries, commissions, marketing and allocated costs, because excluding large channels makes comparison weak, and check zero spend, since a near-zero denominator can produce a huge or undefined ratio that should not be read as extraordinary efficiency. Account for expansion, as existing customers may buy more and lift revenue with little new acquisition spend, which can be good business but changes what the metric measures.

Account for churn too, because lost customers can offset new sales, and a low result may reflect retention rather than poor campaign execution. Consider price changes, because raising subscription prices can increase revenue without a proportionate increase in new customer acquisition, and consider seasonality, since marketing spend and contract starts may cluster in different quarters and a single quarter can be noisy.

Check acquisition timing, as sales teams may work leads for months so the previous quarter's expense may not correspond to the current quarter's revenue, and check acquisitions because buying another company's customers can increase revenue without reflecting organic sales-and-marketing efficiency. Distinguish cash and accrual as well, since expense recognition and customer cash receipts can occur at different times and this is not a liquidity measure.

Do not equate the ratio with profit, because the formula omits gross margin and ongoing service costs and a 1.0 value does not guarantee one-year cash payback. Pair it with CAC payback, since gross-margin-adjusted CAC payback better connects acquisition cost with contribution from new customers though it also needs assumptions, and use cohorts to break out new-customer ARR, expansion and churn, which explains the net growth that enters the ratio.

Avoid universal cutoffs, because claims that 0.75 or 1.0 always means 'scale spending' ignore stage, margin, sales cycle and data quality. Compare peers carefully, as a self-serve monthly product and an enterprise software vendor have different expense timing and retention, and track several periods, since a rolling four-quarter analysis can smooth some noise if any changed method is disclosed.

Set action tests before increasing spend by examining lead quality, conversion, retention, gross margin and capacity, because one number should not set the budget, and report both numerator and denominator, as a high ratio from a small revenue increase and tiny spend has a different strategic meaning from large growth at scale. For owners, the quarterly Magic Number is a starting view of growth per sales-and-marketing dollar, and its formula should be named every time because even reputable sources use the label differently.