The formula
Magic number
Magic number = (This quarter's revenue − Last quarter's revenue) × 4 ÷ Last quarter's sales and marketing spend
- Why × 4: it annualizes the quarter's revenue growth, so you compare a year of new revenue with one quarter of spend.
- Why last quarter's spend: sales and marketing take time to land. A one-quarter lag is the convention.
- Sales and marketing, fully loaded: salaries, commissions, ads, tools, events, agencies. Not just the ad budget.
For a pure subscription business there is a cleaner version: net new ARR over the quarter divided by last quarter's sales and marketing spend. Same idea, no × 4 needed. Pick one and keep it.
A worked example
Example, fictional figures
Q2 revenue: $900k. Q3 revenue: $1.05M
Growth: $150k × 4 = $600k of annualized new revenue
Q2 sales and marketing spend: $750k
Magic number = 600 ÷ 750 = 0.8
What a good magic number looks like
| Magic number | What to do |
|---|---|
| Over 1 | Efficient. Spend more, as fast as you can hire well. |
| 0.75 to 1 | Healthy. Keep investing. |
| 0.5 to 0.75 | Watch it. Fix conversion before adding budget. |
| Under 0.5 | Stop and look under the hood before the next dollar. |
Four traps
- Revenue that sales did not bring. A price increase or product-led upsell lifts revenue without any sales effort. The magic number looks great. Know what you are looking at.
- Seasonality. A big Q4 makes Q4 look magic and Q1 look cursed. Use a trailing four-quarter average.
- Churn in the revenue line. Revenue change is net of churn, so a churn spike drags the number even if sales did their job. Check the MRR bridge before blaming the sales team.
- Half-loaded costs. Counting ads but not the salaries of the people running them. Your number will be magic. It will also be wrong.
Magic number, burn multiple, CAC payback
Three efficiency numbers, three scopes. The magic number looks at sales and marketing only. The burn multiple looks at the whole company. CAC payback adds gross margin and speaks in months. When the magic number is fine and the burn multiple is not, the problem is outside sales.
George reads them side by side every month, because one good number on its own proves nothing. Two that disagree are where the conversation starts.