How to measure true CAC when attribution is broken
Blended CAC is honest and useless. Channel CAC is useful and usually wrong. Here is how to get a number you can act on without pretending your attribution is better than it is.
blended CAC = all spend ÷ all new customers · channel CAC = spend ÷ attributed customers
Start with blended CAC — it cannot be wrong, because it does not depend on attribution at all. Then use channel CAC to allocate within that total, and treat the gap between the two as your measure of how much attribution you are losing.
If your channel CACs multiply out to fewer customers than you actually acquired, your attribution is under-counting. If more, it is double counting.
Every CAC number you have ever been shown is one of two things: a number that cannot be wrong but cannot guide a decision, or a number that guides decisions and is probably wrong.
The trick is knowing which one you are looking at, and using both deliberately.
Blended CAC: trustworthy, undirected
blended CAC = total acquisition spend ÷ total new customers
Both numbers come from systems that cannot be wrong. Spend is in your bank. New customers are in your database.
No attribution is involved anywhere, which is why blended CAC is the only acquisition number that survives contact with a broken tracking setup — and why it should be the one on your board deck.
Its limitation is that it tells you nothing about where the next dollar should go.
Channel CAC: directional, distorted
channel CAC = channel spend ÷ conversions attributed to that channel
The numerator is solid. The denominator is a measurement, and measurements are incomplete.
The distortion is not random, which is what makes it dangerous: channels with the longest consideration cycles look worst, because their conversions are the ones that fall outside the identity’s lifetime. Prospecting, content and awareness get systematically punished; retargeting, email and branded search get systematically rewarded.
Act on channel CAC without correction and you will defund the top of your funnel while congratulating yourself on efficiency.
The reconciliation that tells you how bad it is
Sum your attributed conversions across all channels and compare to your actual new-customer count.
| Result | Diagnosis |
|---|---|
| Attributed ≈ actual | Attribution is roughly complete. Trust channel CAC. |
| Attributed < actual | Under-counting: identity expiry, ad blocking, blocked events |
| Attributed > actual | Double counting: cross-platform claims, or client+server without a dedup key |
That single comparison is more informative than any dashboard, and almost nobody runs it.
A useful correction while you fix the underlying problem: scale your attributed conversions up to the real total, proportionally, and recompute channel CAC on the adjusted numbers. It is not precise, and it is far better than pretending the missing conversions belong to nobody.
Fully loaded vs media-only
Two different numbers for two different audiences:
- Media-only CAC — ad spend only. Use it to compare channels to each other.
- Fully loaded CAC — plus salaries, agencies, tooling, creative production. Use it for the board and for unit-economics decisions.
Fully loaded is typically 1.4–2.0× media-only. Reporting one while your CFO assumes the other is a reliable way to have a bad quarter.
The number CAC only means something against
CAC alone is meaningless. Three companies with an identical $200 CAC:
| Business | LTV | Payback | Verdict |
|---|---|---|---|
| B2B SaaS, $200/mo, low churn | $4,800 | 3 months | Scale hard |
| DTC, $150 AOV, 30% repeat | $195 | 1 order, barely | Marginal, fix retention |
| DTC, $80 AOV, no repeat | $80 | Never | Terminal |
Same CAC, three completely different businesses. Which is why the useful sequence is always:
- Break-even ROAS — the revenue multiple you need at your contribution margin. (calculator)
- CAC — what you are actually paying.
- Payback — how long until it comes back, and how deep the cash valley gets. (calculator)
Only step 2 depends on attribution, which is worth remembering when someone says the tracking is broken so no decisions can be made.
Fixing the denominator
Everything above is how to work with an imperfect number. Making it a good number is a separate job, and it is the same short list every time:
- A durable, server-set visitor id so long consideration cycles stay attributable.
- Server-confirmed conversions so blocked sessions still count.
- A de-duplication key so client and server do not double count.
- An early email capture so identity outlives the cookie.
Do those four and the reconciliation gap closes to within a few percent. Then channel CAC becomes a number you can actually act on.
Frequently asked questions
What is blended CAC?
Total acquisition spend divided by total new customers, over the same period. No attribution required, which is exactly why it is trustworthy — and why it cannot tell you where to spend more.
Why is channel CAC usually wrong?
Because it divides real spend by attributed conversions, and attribution is systematically incomplete. The channels with the longest consideration cycles look worst, because their conversions are the ones falling outside the identity's lifetime.
Should I include salaries in CAC?
For a board-level number, yes — fully loaded CAC includes marketing salaries, agency fees and tooling. For channel allocation decisions, media spend only, or you cannot compare channels to each other.
How does CAC relate to break-even ROAS?
Break-even ROAS tells you the revenue multiple you need at your margin. CAC tells you what you are actually paying. Payback tells you how long until that spend comes back. You need all three, and only the second depends on attribution.
What is a good CAC?
Only meaningful against payback and LTV. A $200 CAC is excellent for a business with a $2,000 LTV and terminal for one with a $150 AOV and no repeat purchase.
Work out what you can afford to pay
The CAC payback estimator shows how many months until a customer repays their acquisition cost, and how deep the cash valley gets first.
Stop guessing which ad made the sale.
Flowsk Signals stitches the anonymous click to the email to the purchase — first-party, server-side, de-duplicated. One snippet, $29/mo, and every conversion comes with a receipt you can inspect.