The channel mix shift

Harbor & Pine sells through three channels: its website, its mobile app and a third-party marketplace. Daniel Foster, who runs the app, has noticed it taking a bigger slice of orders every year and wants to use that in next year’s planning. The question sounds simple. The trap is that a share can grow for two very different reasons: because the app is winning new business, or because the same customers are moving over from the website.

This case is the whole job:

1. Frame the question so it can be answered: which orders, which measure, which years, and what the shares can and cannot tell you.
2. Check the data: confirm the channel values are clean and see where the pending orders sit.
3. Analyse, in SQL, pandas or Excel, whichever you prefer: each channel’s share of completed orders and revenue by year, then the counts behind the shares and what an order is worth in each.
4. Recommend: pick the paragraph you would send Daniel, with the numbers and the caveats.

The case opens as soon as you finish dates and CASE in SQL, grouping in pandas, or pivot tables in Excel. It uses conditional counts, a window function for shares (or transform in pandas, or SUMIFS with mixed references in Excel); the lessons walk through each.