FiberFuel: Blended CAC, Returns and Channel Attribution
FiberFuel sells a psyllium-fibre supplement direct-to-consumer in Thailand. Monthly media spend is THB 2.1 million: THB 1.8 million on Meta and THB 300,000 on nine TikTok creators. Meta's Ads Manager reports ROAS 2.4, but 40% of attributed conversions are view-through within seven days and a post-purchase survey says 28% of new buyers first heard about the brand on TikTok. Average order value is THB 890; first-order gross margin is 55%; 22% of first-time buyers reorder within 90 days at the same margin. Finance calculates blended CAC at THB 520 across 4,040 new customers last month and wants a 90-day payback. TikTok code redemptions produced 1,240 orders at THB 760 average, but 38% were cancelled or returned. The founders must allocate next quarter's THB 6.3 million budget and are arguing about whether to cut Meta or double TikTok.
1. Calculate 90-day contribution per new customer and the maximum CAC that would still break even within 90 days. Is THB 520 acceptable? Show the math and the assumptions that matter most.
Read the worked answer
90-day revenue per new customer = THB 890 + (22% × 890) = THB 1,085.80. Gross profit = 55% × 1,085.80 = THB 597.20. At CAC THB 520, 90-day contribution = THB 77.20 per customer, or THB 311,888 across 4,040 customers. That is a 14.9% return on the THB 2.1 million spend.
Breakeven CAC = THB 597.20. So THB 520 is technically acceptable, but the margin of safety is only 13%.
Three assumptions matter most. First, the 22% repeat rate is measured on a cohort that may include promotional buyers; if it falls, the ceiling falls. Second, returns and cancellations are excluded from the blended CAC calculation; FiberFuel's own TikTok data show 38% cancellations, so a 10% overall cancellation rate would push effective CAC above the ceiling. Third, the 55% margin may exclude fulfilment, payment fees and free-shipping subsidies.
Recommendation: treat THB 520 as a ceiling, not a target, and report 90-day contribution after returns by cohort and channel.
2. Diagnose the Meta ROAS decline using three plausible mechanisms and propose one test that separates creative fatigue from attribution inflation.
Read the worked answer
Mechanism 1: creative and audience fatigue. Frequency rises, CPM rises and click-through falls, so each additional impression buys fewer incremental orders.
Mechanism 2: attribution inflation. Forty percent of Meta conversions are seven-day view-through, which captures demand that TikTok, organic or retail search created. Meta's reported ROAS is then overstated and its incremental ROAS is lower.
Mechanism 3: mix shift at scale. The easiest buyers were acquired first. As the brand scales, marginal CAC rises even if the platform performs as before.
Test. Run a geo holdout: switch off Meta in matched regions for four weeks and compare total new-customer orders, not platform-attributed orders, against control regions. In parallel, run a creative-refresh test with frequency caps and compare incremental conversions per 1,000 impressions. If sales hold in the holdout, Meta is largely harvesting demand. If sales fall by close to the platform-reported number, attribution is credible. Triangulate with a forced-choice post-purchase survey question asking where the buyer first heard about the brand, which can validate or challenge the 28% TikTok figure.
3. Recommend a THB 6.3 million allocation across Meta, TikTok and a measurement reserve, with expected outcomes and guardrails.
Read the worked answer
Allocate THB 3.4 million to Meta, THB 2.2 million to TikTok and THB 0.7 million to measurement and creative testing.
Expected outcomes. Meta should deliver a 90-day CAC no higher than THB 550 after holdout calibration. TikTok should deliver a delivered-order CAC no higher than THB 450 after cancellations and returns, tracked with codes and post-purchase surveys. Blended new customers should reach at least 11,000 for the quarter at a blended CAC no higher than THB 570.
Guardrails. Pause any creator whose cancellation or return rate exceeds 25%. Cap Meta frequency and refresh creative every three weeks. Require positive 90-day contribution by cohort before scaling any channel. Do not increase TikTok beyond THB 2.2 million until a four-week holdout confirms incrementality.
The measurement reserve is not overhead. It funds the holdout, the survey and the creative tests that prevent next quarter's argument from repeating. The allocation deliberately reduces Meta's share without eliminating it, because the post-purchase survey and view-through rate suggest Meta is partly harvesting demand rather than creating all of it.