Commerce

Every commerce system sells from a catalog to a customer and owes someone a fulfilled order: a restaurant's menu pushed to three apps, a grocery store's shelf priced for delivery, a courier batching two orders on one trip, a fee a city caps at 15%, a membership a regulator says was too hard to cancel. Underneath, the restaurant, the store, the marketplace and the point of sale share one set of records (merchants and their locations, items and modifiers, orders and their lines, fulfillments, payments and customers), and the hard problems are keeping those records agreeing across every channel that sells from them. This book surveys how the industry does each piece, benchmarked across DoorDash, Uber Eats, Instacart, Grubhub, Shipt, Gopuff, Deliveroo, Just Eat Takeaway and Wolt and the POS and ordering vendors behind them, such as Toast, Square and Olo, with no single company as the model, and how you'd build it, extending one core data model chapter by chapter.

It reads in two parts so far, plus a reference. The Business: the parties, the channels and the core data model every later part extends. On-Demand Delivery: the landscape of marketplaces, logistics-only networks and first-party ordering with take rates from the filings, merchants, menus and grocery catalogs synced from the POS, orders, prep time, dispatch and batching, fees, commissions and the city and state laws that cap and police them, memberships and cancellation law, and refunds, missing items and fraud. Reference: a glossary of every term, grouped by part. The rest of the planned scope comes later, as parts that extend the same model: catalog, orders and inventory across your own store, the POS and marketplaces like Amazon, subscriptions and memberships beyond delivery, and regulated goods.