Where should a coffee roaster spend?
The companies and data in this case are made up.
A coffee roaster has $1,000 for a marketing pilot. Compare three channels and recommend where to test the next dollar.
Suggested time: 20 minutes for a written response, then 5 minutes for the challenge. This is a practice exercise, not a full Regionals paper or a statement of the final exam format.
A · Completed eight-week pilot
| Channel | Marketing spend | New customers | Contribution per first order |
|---|---|---|---|
| Online | $600 | 60 | $8 |
| Grocery promotion | $400 | 100 | $3 |
| Café referral | $300 | 30 | $12 |
B · Repeat orders observed during the pilot
| Channel | Repeat orders per acquired customer | Contribution per repeat order |
|---|---|---|
| Online | 0.5 | $8 |
| Grocery promotion | 0.2 | $3 |
| Café referral | 0.8 | $12 |
Evidence limit: Contribution is after product and channel variable costs but before marketing and fixed costs. Pilot groups differ and are not randomized. Repeat orders are observed only within eight weeks; channel capacity and incremental effects beyond the pilot are unknown.
Your task
- Calculate customer acquisition cost and observed eight-week contribution per acquired customer for each channel.
- Calculate contribution after marketing for each pilot. What would a simple proportional projection imply for a $1,000 spend?
- Recommend an allocation or experiment. Explain why the proportional projection may fail.
Your answer stays on this page. It is not submitted or saved; copy it before leaving.
Follow-up challenge
Grocery acquired the most customers. Why is that not enough to select it?
Read a worked response
Acquisition cost is $10 online, $4 grocery and $10 café. Eight-week contribution per customer is $12 online ($8 + 0.5 × $8), $3.60 grocery and $21.60 café. After acquisition cost, that leaves $2, −$0.40 and $11.60 per customer respectively.
The pilots produce $120 online (60 × $12 − $600), −$40 grocery and $348 café. Scaling those returns mechanically to $1,000 suggests $200, −$100 and $1,160 after marketing. These are conditional projections, not reliable forecasts: capacity, audience saturation and customer mix may change with scale.
I would expand the café test in stages while preserving a comparison group, because its observed contribution after marketing is highest. Start with a spend comparable to the original pilot, check incremental new customers and repeat orders, then release more budget only if returns persist. Referrals might reward customers who would have bought anyway, and the café may have limited capacity. I would reconsider if measured incremental contribution falls below acquisition cost or if another channel wins under a comparable test.
Other recommendations can earn full credit when they follow from the exhibits and acknowledge the limits.
Score your response · 100-point practice rubric
This self-assessment assigns 20 points to each category. These weights apply only to this practice exercise; official competition weights remain unpublished.
| Category | Evidence needed for 20/20 |
|---|---|
| Interpretation | Distinguish customer counts, contribution, marketing cost and an eight-week observation window. |
| Evidence | Use both first-order and repeat-order information for every channel. |
| Reasoning | Calculate acquisition costs, net pilot contribution and clearly conditional projections. |
| Judgment | Recommend a staged allocation with a scale and incrementality check. |
| Defense & adaptability | Explain the grocery loss and avoid treating observed repeats as lifetime value. |
| Score per category | Practice standard |
|---|---|
| 0 | Missing, irrelevant or fundamentally incorrect. |
| 5 | Some relevant observations, with major errors or unsupported assertions. |
| 10 | Partly correct; important evidence or a reasoning step is missing. |
| 15 | Sound analysis with a minor omission or an underdeveloped limitation. |
| 20 | Accurate, supported and complete for the task; assumptions and limits are explicit. |
The worked response illustrates full-credit evidence in the first four categories. For Defense & adaptability, score your own response to the follow-up challenge; reading the model answer does not demonstrate that skill.
What to practice
How customers respond to a business decision.
- Segmentation
- Demand analysis
- Pricing
- Positioning
- Customer behavior
- Experimentation
- Channels
- Brand and customer economics
How to use this case
- Read the case and restate the decision in one sentence. List the quantities, units and constraints before calculating.
- Work each question without the model response. Write the calculation and the inference separately.
- Compare your work with the model. Mark one arithmetic error, missing assumption or unsupported claim to correct.
- Ask a partner to challenge your recommendation. Give a specific condition that would change it.
Skills checklist
| Skill | Application in this case |
|---|---|
| Interpretation | Distinguish customer counts, contribution, marketing cost and an eight-week observation window. |
| Evidence | Use both first-order and repeat-order information for every channel. |
| Reasoning | Calculate acquisition costs, net pilot contribution and clearly conditional projections. |
| Judgment | Recommend a staged allocation with a scale and incrementality check. |
| Defense & adaptability | Explain the grocery loss and avoid treating observed repeats as lifetime value. |
Common mistakes
- Quoting a number without explaining what it changes.
- Treating a forecast as an observed fact, or a correlation as a causal effect.
- Optimizing revenue while ignoring cost, risk or opportunity cost.
- Listing uncertainties without making a provisional decision.
A second practice session
Change one numerical assumption in the exhibit. Recalculate the decision threshold, explain whether your recommendation changes, and say which original assumptions still matter. Have a partner check units and arithmetic.
The content areas above describe preparation topics. They do not assign exam percentages or finalize time limits, permitted resources or grading weights. See the handbook draft for adopted rules.