What this covers
- What it actually tells you
- How to calculate it honestly
- The three ways it misleads
- What to pair it with
What it actually tells you
Customer acquisition cost is why blended CAC hides the problem. Used well it is a diagnostic; used carelessly it becomes a target, and any metric that becomes a target stops measuring what it used to.
Split by channel and calculate on margin, not revenue. The channel that looks fine in aggregate is usually being subsidised by another.
How to calculate it honestly
The arithmetic is rarely the hard part. The judgement is in the denominator and the window.
- Define the population — which records count, and which are excluded as test, duplicate or out of scope
- Fix the window — cohort by entry date rather than exit date, or improving numbers will simply reflect a slow month
- Use margin, not revenue, wherever money is involved
- Segment before averaging — a single figure across segments usually describes nothing real
The three ways it misleads
Aggregation. One number across segments hides the variation that would have told you what to do.
Timing. Measuring on exit date flatters slow periods and punishes fast ones.
Isolation. Almost every sales metric can be improved by damaging another one. Customer acquisition cost in particular moves when something upstream changes, so reading it alone invites the wrong conclusion.
What to pair it with
Read customer acquisition cost alongside stage conversion and time in stage. Those two locate the problem; customer acquisition cost tells you it exists.
And where money is involved, pair it with collected revenue rather than bookings. In most B2B businesses the gap between the two is material, and metrics that stop at the booking answer a different question from the one finance is asking.