Taking over a booking platform and rebuilding its front end

The context

A platform connecting customers with hospitality venues: online booking on the customer side and, on the venue side, the management that goes with it — quotes for groups, invoicing, communications, dashboards.

The project came to us when it already existed. After an initial takeover phase, the interface was rebuilt from end to end.

The challenge

This platform serves four audiences with opposing needs, which is the product’s central difficulty:

• The customer wants to book in three clicks, ideally without an account.
• The venue manager wants their schedule, wants to answer group enquiries, and wants to turn a quote into an invoice without retyping anything.
• The internal team creates and edits the listed venues.
• The administrator needs a global view of the platform.

A badly designed permission system is paid for twice: first in security holes — a user seeing what they should not — and then in complexity, when every new feature forces you to relitigate who is allowed to do what.

The second challenge is the one every takeover faces: when should you rebuild, and when should you leave it alone? Rebuilding for comfort is professional malpractice; never rebuilding is too, when the existing base blocks everything you want to build on top of it.

What we did

• Takeover before judgement. Understand what exists, run it, ship it — before proposing anything.
• A complete front-end rebuild, decided once the existing system was under control and not on day one. The previous interface was replaced while the back-end and the data stayed in place. That split is what makes a rebuild fundable: you don’t put what works back on the table.
• An explicit four-level permission model, enforced server-side — a check that only exists in the interface is not a check.
• A continuous quote-to-invoice flow, with no double entry, because that is exactly where a manager’s admin time disappears.
• Venue search designed for customer-side discovery.

The outcome

The platform is in active development, with a fully rebuilt interface on a preserved functional base.

The lesson here is about sequence. A rebuild proposed on the first day of a takeover is an opinion. The same rebuild proposed after running the existing system in production is a diagnosis — and only one of those is something a client is right to pay for.