Generic reservation widgets treat every restaurant the same way: a grid of time slots with a capacity number. That works passably for a restaurant with uniform two-top tables and no walk-in traffic, and breaks down for almost everyone else — mixed table sizes, a bar section that takes walk-ins, private dining rooms with different rules.
Zuma Eats models each partner restaurant's actual floor plan — table sizes, sections, and which tables can be combined for larger parties — rather than a single capacity number. A booking checks availability against that specific configuration, and a host-side view shows the real floor plan filling up in real time, whether the booking came through the app or was entered manually by staff for a walk-in or phone booking.
The detail that matters here is the last part: the system has to be the single source of truth regardless of which channel a booking comes through, or double-bookings are inevitable. This is a recurring pattern in any ERP-style build I do for a business with both digital and in-person booking channels — the system only works if every channel writes to the same record, not if the digital channel is a separate calendar someone has to reconcile manually against the paper one.