Design analysis to validate reported decline

A beverage company claims 'Q2 sales dropped' but gives you only one dataset (the transactions table described above) and 45 minutes. Design the analysis to validate or refute the claim rigorously. Answer the following: - Define the precise decision goal: what action hinges on your conclusion, and what metric(s) must move? State primary (e.g., total revenue) and counter-metrics (e.g., units, avg price, same-store revenue) and why. - Choose the analysis grain and scope: what is the correct time aggregation and business grain (store x product x day), and how will you ensure alignment so averages are not misleading? - Outline a step-by-step plan using only the provided dataset: period definitions, baseline selection (Q1 2024 and Q2 2023), same-store filter, mix effects (region/product), returns handling, stockout proxy, and missing-days bias mitigation. - List at least 5 sanity checks to detect false alarms (e.g., store openings/closures, promo timing, price increases masking unit decline, Simpson’s paradox across regions, returns spikes, data latency). - Specify the minimal outputs you will present (tables/plots), the interpretation rubric for 'decline' (e.g., both revenue and units down by >5% overall and same-store), and how you will communicate assumptions and limitations in under 3 minutes. - If you had 5 more minutes and could request exactly one additional dataset, which would it be, why, and how would it change your conclusion?