Northwind Meal Kits: Reallocating a $400K Acquisition Budget Under a Payback Rule
Northwind Meal Kits sells frozen dinner boxes by subscription and runs a $400,000 monthly acquisition budget split 30% paid search, 40% paid social, 20% affiliate coupon sites, 10% influencer placements. Finance reports 12-month contribution per customer net of product and fulfillment: paid search $145, paid social $160, affiliate $52, influencer $210. Current acquisition costs are $62, $88, $41 and $130 respectively. Affiliate volume doubled after a 50%-off code was syndicated across deal forums. Paid social frequency has climbed to 6.4 while its conversion rate stayed flat. Cash is tight, and the CFO will not approve any spend with payback longer than four months. Next quarter the board wants 20% more new customers on the same budget. The growth team proposes cutting affiliate and moving that money into paid social.
1. Does the data support the growth team's reallocation? What does the board's 20% target actually require?
Read the worked answer
Convert the budget into units first, because percentages hide the trade-off. At the stated CACs: search 120,000/62 = 1,935 customers; social 160,000/88 = 1,818; affiliate 80,000/41 = 1,951; influencer 40,000/130 = 308. Total 6,012 customers, blended CAC $66.50.
Contribution tells a different story than volume. 12-month contribution: search $280,575; social $290,880; affiliate $101,452; influencer $64,680 - $737,587 on $400,000 of spend, a 1.84 ratio and roughly 6.5-month payback. Affiliate delivers 32% of customers but 14% of contribution, and its volume is an artifact of a syndicated 50% discount, which selects for churn. Cutting it is defensible.
But the CFO's four-month rule, applied literally, shuts down the entire program. Four-month contribution is roughly a third of the 12-month figure: search $48 vs $62 CAC, social $53 vs $88, affiliate $17 vs $41, influencer $70 vs $130. Every channel fails. The rule must become a portfolio constraint (12-month contribution/CAC at or above 1.5, cash payback under seven months), not a per-customer screen.
The 20% target is the real problem. 20% more customers on flat spend means blended CAC of $400,000/7,214 = $55.40, a 17% improvement. Reallocation alone cannot deliver that: paid social at frequency 6.4 is near saturation, so its next dollars cost more than $88, not less. The gap has to come from conversion-rate and retention work. Recommend cutting affiliate, freezing social at current spend, funding an influencer test, and renegotiating the target to +20% contribution.
2. Write the 90-day test and guardrail plan for moving the affiliate budget, and state what evidence would reverse your recommendation.
Read the worked answer
Stage the move, do not flip it. Phase 1 (weeks 1-2): cut affiliate spend by 50%, not 100%, keeping the non-coupon publisher segment so you can separate discount-seekers from genuine content partners. Phase 2 (weeks 3-8): release the freed $40,000 into (a) a search incrementality test, since search is the highest 12-month net contribution channel at $83, and (b) a creative refresh on social rather than a budget increase.
Guardrails that stop the test from becoming a permanent subsidy.
- Marginal CAC ceiling per channel, not blended: search $95, social $120, influencer $150, based on 12-month contribution discounted at the finance hurdle.
- Frequency cap of 4.0 on social; if incremental spend pushes frequency past it with flat CVR, stop.
- A 10% holdout in each channel for four weeks so you can measure lift rather than attributed conversions.
- Weekly blended CAC and contribution-per-customer tracking against the $66.50 baseline; escalate on two consecutive weeks of deterioration.
Reversal evidence. Reconsider the affiliate cut if, after 30 days, the retained publisher segment shows 12-month contribution per customer above $110 and coupon-code redemption cohorts show repeat rates within 10% of organic cohorts. Reconsider the social freeze if a creative refresh lifts CVR by 25% or more at stable frequency, which would mean the saturation was creative, not audience. Do not reverse on a single week of ROAS noise.