← Back to work

Scaling a Product with Empathy

Reserve's restaurant product had hit its ceiling. We rebuilt the vision from scratch — and closed a $10M Series B.

Reserve
VP, UX & Design
100 restaurants · 7 markets
$10M Series B
[ Restaurant product — hero image ]

The situation

Reserve's first B2B reservation system had been live for nearly a year and had nearly maxed out its potential market. Experienced GMs loved it. Less experienced operators — the majority of the prospective market — couldn't use it.

The challenge wasn't fixing the product. It was reimagining the product vision to serve a wider audience without losing the operators who already depended on it.

Starting in the field

Before touching a single wireframe, we spent months inside restaurants. We established relationships with over 100 restaurants and 27 operators across 7 markets, and set up a permanent restaurant feedback program to inform all future development.

We staged — the restaurant term for shadowing — through both front-of-house and back-of-house operations. We learned the workflows, the workarounds, and the small things that made or broke a service.

Helping a host fold menus at the beginning of service helped us earn street cred with the crew. That kind of access doesn't come any other way.
[ Field research — staging in restaurants ]

What we learned

The core tension: some operators want full automation. Others want total control. Most want something they can dial up or down based on the night, the staff, and how many covers are on the books.

The problem statement: build a reservation system that becomes an extension of the operator and scales to their level of sophistication.

Cracking the algorithm: Reservation Tetris

Before designing any interface, we had to solve the logic underneath it. Restaurant seating is not like airline seats — you're dealing with combinable tables, varying turn times by party size, and the deeply personal judgment of individual operators.

We printed a large grid representing tables and times, gave experienced operators a stack of reservation post-its, and watched how they reshuffled. We documented the rules they were following intuitively — then built the algorithm from their behavior, not our assumptions.

We called it Reservation Tetris. The engineering team had a proof of concept in two sprints.

We built the algorithm from their behavior, not our assumptions.
[ Reservation Tetris — ideation and algorithm ]

The product

A cross-platform system — responsive web for back-of-house operations, native iOS for front-of-house during service — with five levels of control operators could dial based on their preference and sophistication: from full automation to complete manual control, with everything in between.

Each control level was designed around operator behaviors we had directly observed. Not hypothetical users. People we had stood next to during a 7:30 dinner rush.

[ Product screens — responsive web + iOS tablet ]

The outcome

A $10M Series B. A product that worked for experienced operators and first-timers alike. And a design process rigorous enough that engineering was able to start building in parallel — because they had been part of the research from the start.