Paid social at 1.06 LTV:CAC - cut, hold, or fix the measurement?
A direct-to-consumer skincare brand sells mainly through its own Shopify storefront. Average order value is $52 and first-order contribution - after COGS, outbound shipping, pick/pack and payment fees - is $18.50. Repeat behaviour is stable: 42% of new customers place another order within 12 months, and repeat buyers average 2.1 additional orders at $61 AOV with $28.25 contribution each. Paid social is the largest acquisition channel at a new-customer CAC of $41. Before the next board meeting the CFO wants paid social spend cut 30% to hit a profit target. The growth lead counters that paid social is 'profitable at 2.1 ROAS and the LTV/CAC is fine.' Paid branded search reports 8.9 ROAS, but 78% of its clicks come from customers who bought in the last 18 months. Email and SMS drive 22% of revenue.
1. Using these numbers, is paid social actually profitable on a new customer? Build the 12-month contribution case and give a recommendation on the CFO's 30% cut.
Read the worked answer
Fix the framing first. A 2.1 ROAS means 47.6% of revenue goes to media. First-order contribution before CAC is $18.50 on $52 of revenue - 35.6%. So the brand spends $41 to earn $18.50 on the first order, a $22.50 loss per new customer. 'Profitable at 2.1 ROAS' only holds if repeat contribution closes that gap, which is the question, not the premise.
12-month contribution per new customer
- Expected orders = 1 + (0.42 × 2.1) = 1.882
- Repeat orders = 0.882 × $28.25 = $24.92
- Total 12-month contribution = $18.50 + $24.92 = $43.42
- LTV:CAC = 43.42 / 41 = 1.06; net contribution = $2.42 per new customer
Payback. Cumulative contribution runs $18.50 after order 1 (45% of CAC), $46.75 after order 2, $75.00 after order 3. Cash payback lands roughly in month 7–9 once repeat timing is considered - long for a brand with no financing headroom.
Recommendation. The CFO is directionally right that the spend is thin, but a flat 30% cut removes the best ad sets along with the marginal ones. Cut in tranches (10%, hold 14 days, 10%, hold), targeting the marginal spend where new-customer CAC sits above roughly $38–40, and watch 60-day repeat rate by cohort to confirm the cut did not degrade quality. Do not cut below the volume needed to feed email/SMS, which drives 22% of revenue.
2. How should the team treat the 8.9 ROAS on branded search, and how would you establish its real value?
Read the worked answer
What 8.9 ROAS is measuring. With 78% of clicks from customers who purchased in the last 18 months, the channel is largely harvesting demand that would have arrived via direct visit or organic result anyway. The number reflects attribution, not incrementality.
Establish the real value cheaply
- Geo holdout: pause branded search in matched DMAs (e.g. 12 test vs 12 control markets) for four weeks, then compare total orders and total revenue per capita - not channel-attributed revenue.
- Expect test markets to lose much less than the channel report implies; the shortfall is the true incremental contribution.
- Check for affiliate or partner bidding on the same brand terms, which double-counts the same conversion.
Decision rule. Fund branded search only to the point where holdout-estimated incremental contribution covers media cost. If incremental ROAS lands near 1.5–2.0, narrow to the top converting terms and lower bids rather than protecting the budget because of the 8.9 headline.
3. If the profit target must be met this quarter, in what order should levers be pulled, and what should be instrumented so the next decision is not guesswork?
Read the worked answer
Sequence by contribution recovered per dollar of risk taken
- Branded search and non-incremental retargeting - cheapest to trim, lowest revenue risk, test-backed.
- Paid social ad sets above the CAC ceiling - tranche cuts with 14-day holds before the next tranche.
- Discount depth and free-shipping threshold - real margin levers, but watch conversion and AOV.
- Retention spend (email/SMS) last - cheapest contribution in the business; cutting it mortgages next quarter.
Instrumentation to fix
- New-customer CAC by channel and cohort, never blended
- First-order contribution after shipping and fees, not revenue
- 60/90/180-day repeat rate by acquisition cohort
- Quarterly holdout-based incrementality for the two largest channels
- LTV:CAC reported on a contribution basis with payback period attached
The blended ROAS view hid the fact that the average new customer returns $2.42 of profit over twelve months; that number belongs in the board pack, not the ROAS.