Decide whether to keep a negative-margin promotion

A retail team is running a Mulch promotion. The promotion currently has a negative unit margin (i.e., discounted price is below unit cost), so on the promoted item itself the company is “taking a loss.”

You are the analyst asked to recommend whether to continue, modify, or stop the promotion.

What you’re given

• Historical transaction data with:
  • basket-level purchases (items, quantities, prices)
  • customer identifiers for a subset of sales (some are anonymous)
  • store/region and date
• Promo metadata: start/end dates, discount level, and eligible SKUs
• Cost data at SKU level (may be noisy / delayed)

Questions

1. What analyses would you run to decide whether the promotion is worth continuing?
2. Propose a primary metric , diagnostic metrics , and guardrail metrics . Explain tradeoffs.
3. How would you estimate incrementality (vs. cannibalization/stockpiling/seasonality)? Provide at least one experimental and one observational approach.
4. If leadership decides to continue the negative-margin promotion anyway, list plausible strategic reasons that could justify it and what data would validate each reason.

State assumptions you need and call out major risks (confounding, missing customer IDs, supply constraints, etc.).