The operations team has received a small order feed in practice.review_feed. Some January orders were delivered twice by the import process. The repeated rows describe the same transaction, so a row count and an order count can differ. Keep that distinction visible while you investigate.
You already know the SQL needed here: select named columns, filter dates and statuses, group rows, count distinct keys and handle missing values. These tasks combine those skills without introducing new syntax. Read each requested output grain before choosing an aggregate.
Use the feed supplied for each exercise rather than the full orders table. The date range and number of repeated records change in the hidden copy. A correct query should still work. Finish with a customer list whose count means orders, even when the same source row arrived twice.