Blended vs. per-channel CAC: which one to trust when
Blended CAC and per-channel CAC answer different questions. Trust the wrong one and you over-fund a channel that was just taking credit for organic sales.
AI can compute both from your spend and new-customer counts. You supply the real numbers, because it must never invent spend, conversions, or a channel's attributed sales.
Fastest path: one prompt, end to end
🤖 AI prompt — paste into ChatGPT / Claude
You are a growth-finance analyst. Use MY numbers only (invent nothing).
Total marketing/ad spend this period: [$]
Total NEW customers this period: [#]
Per channel, give me: channel name, spend [$], and attributed new customers [#]
(list every channel; note the attribution model each platform uses if you know it)
Do this and show the arithmetic:
1. Blended CAC = total spend / total new customers. Call this the north star.
2. Per-channel CAC = channel spend / that channel's attributed new customers, for each channel.
3. Sum the per-channel attributed customers and compare to total NEW customers. If the sum EXCEEDS actual new customers, flag over-attribution (channels double-counting the same sale).
4. Point out which channels look cheap only because they claim conversions others assisted.
5. Recommend: use blended CAC to judge overall health; use per-channel CAC ONLY for directional allocation, with the over-attribution caveat.
If a number is missing, ask; do not guess.
Output: blended CAC + a per-channel CAC table + an over-attribution flag + one allocation recommendation.
Or do it in 4 steps
- Compute blended CAC first, and treat it as the truth. Total marketing spend divided by total new customers, no attribution involved. It can't be gamed, and it answers the only question that pays the bills: is the whole business acquiring customers profitably?
- Compute per-channel CAC for allocation, not for scoring. Channel spend / that channel's new customers tells you where a marginal dollar might work hardest. It's necessary, but it's a directional signal, not gospel.
- Watch for attribution pollution. Add up every channel's claimed conversions. If the total beats your actual new-customer count, channels are double-counting. Retargeting and branded search especially claim sales other channels created, which makes their CAC look artificially cheap.
- Build a simple reconciliation habit. Each month, compare blended CAC to the spend-weighted average of your channel CACs. When they drift apart, attribution is lying somewhere: trust blended for the go/no-go and use channel CAC only to nudge budget.
Worked example (labeled): Spend $20,000, 500 real new customers, so blended CAC = $40. Channels claim: Paid Search 250 at $30, Meta 200 at $45, retargeting 150 at $15.
Claimed total = 600 > 500 actual, so 100 conversions are double-counted and retargeting's $15 is illusory. Blended $40 is the honest figure; the channel table only guides where to lean. Illustrative, run your own numbers.
How those attribution models actually decide who gets credit is a marketing topic: see attribution-true-performance under Marketing. Then bring blended CAC into cac-and-payback to check it against your margins.