Case study
2020–2021

A Platform Staff Worked Around — Not With

A restaurant management platform had every feature a hospitality team could need: reservations, floor plans, table status, zones, waitlists, service flows. But during the 45 seconds between seating a party and picking up a phone, nobody had time to figure out which icon did what. Support calls were the workaround.

Over a year, I led the redesign of the core workflows — from the reservation list to the floor plan to the information architecture of every card. The goal wasn't to add features. It was to make the existing ones usable under pressure.

My Role
End-to-end Design · Research + UX + UI
Team
Developers · me
Duration
1 year (part-time)
Deliverables
Research · UX · Design System · Prototypes
Restaurant platform — redesigned integrated view
01 — The Challenge

The tool had everything. Nobody knew how to use it — and support was paying the price.

The platform covered every workflow a hospitality team needed. In practice, staff called support instead of using it. The support team documented the feedback. The patterns were consistent across all 4 user types.

Before
Before: original restaurant platform — dense and overwhelming
1
Overwhelming main view — too many features on one screen, no clear entry point, icons with no recognizable meaning.
2
Customer names and key reservation data were hard to read — font size, contrast, and density made scanning impossible during service.
3
Finding a reservation on the floor plan took too long — the visual language didn't communicate status quickly enough.
4
Onboarding was slow — new staff took weeks to become comfortable with the software.
Key signal
The support team wasn't fielding technical bugs — they were fielding questions about how to do basic things. Every support call was a UX failure.
02 — Research

9+ users. A support team's worth of documented pain. One clear direction.

The support team had been collecting user feedback systematically — the closest thing to a continuous usability study. Combined with direct observation sessions, it gave us a precise map of where the tool was failing.

👁️
Observation — 9+ users
Watched staff use the tool during actual service hours across all 4 roles. What they clicked, what they skipped, what they said out loud.
📋
Support log analysis
Reviewed the support team's documented feedback — categorized by interaction, frequency, and severity. Turned a support backlog into a prioritized design brief.
🗺️
Workflow mapping
Traced the most common flows step by step: new booking → assignment → arrival → seating → table turnover. Identified exact moments where users abandoned the tool.
User flow: login to reservation types
Key finding
Staff developed mental shortcuts around broken flows: memorizing which icon did what, using external notebooks to track what the floor plan didn't show, calling support for things they felt they should already know. Every workaround was a design failure.
03 — Users

Four roles. All working under pressure. Each failing in a different place.

Hosts

Manage walk-ins and reservations simultaneously. Need to assign tables in real time — often without looking away from the entrance.

Failure: can't find the right reservation fast enough → guest waits.

Managers

Oversee occupancy and service flow. Need floor visibility without interrupting service.

Failure: can't read the floor at a glance → manual counting → decisions delayed.

Floor Staff

Need table status and assignment updates during service. No time to navigate deep menus.

Failure: wrong table assignment → service errors.

Admins / Configurators

Configure zones, rules, turn times, floor layouts. Less time pressure — but system reliability is critical. Failure mode: misconfigured settings create cascading errors during service hours.

Shared need
Get in, do the task, get out. During service, every second lost to interface confusion is a second taken from a guest.
04 — Framing the Opportunity

If staff could read the interface as fast as they read the room, support calls would drop.

The opportunity wasn't to redesign the software architecture — the underlying logic was sound. The opportunity was to redesign the surface: the visual language, the information hierarchy, and the interaction patterns that staff encountered under pressure.

Color before text

Make reservation status scannable by color — not by reading a label or decoding an icon.

One view, not two

Reservation list and floor plan should work together — not as separate contexts that require switching.

Design system

A consistent visual language that scales to the platform's complexity without adding to cognitive load.

The constraint

This was a mature product with real users. A full redesign would require retraining. Every change had to feel like an improvement — not a replacement. Adoptable without a training session.

05 — Approach

Map the system first. Then design for the moment — not the feature.

🗺️
User flow mapping
Documented every path through the reservation lifecycle — from new booking to table turnover. Identified exact moments where users abandoned the interface.
📊
Support log analysis
The support team's feedback logs became a structured design brief — ranked by frequency and severity of interface failures.
🔁
Iterative prototyping
Rapid cycles shared with staff for quick feedback. No long review cycles — if it needed explaining, it needed redesigning.
Key principle
In high-pressure environments, a design that requires thinking is a design that fails.
06 — The Solution

Not a new product. A usable version of the one that already existed.

01
Information Hierarchy — The Reservation Card

The reservation card was the atomic unit of the entire platform. We mapped every field (state, time, party size, table, zone, name, date, prescriptor, comments, waiters, tags) and rebuilt the hierarchy from scratch — primary information at the top, status communicated by color at the left edge.

Design decision

Every field that required calling support to interpret was redesigned or renamed. The card was anchored around what staff needed in the first 2 seconds — name, time, table, status — not around database structure.

Reservation card anatomy — annotated
02
Reservation List Clarity

The original list buried customer names in small text among columns of equal visual weight. We introduced color-coded left borders for status, increased typographic hierarchy on names, standardized bottom navigation with count-per-status tabs, and added contextual tags as colored chips.

Design decision

Color-coded borders replaced icon guessing. A 1-second status read replaced a 10-second scan. The bottom tab bar — showing counts per status — gave managers an at-a-glance view of service state without opening a separate report.

Reservation list — color coded, readable names
03
Floor Plan Redesign

The original floor plan required reading text on individual table blocks to determine status. We rebuilt it using color as the primary communication layer — each table block's background indicates its status. Text (time, name, pax) is secondary. Multi-floor navigation via tabs. Table blocks support 1/2/3 stacked reservations with hover to expand.

Design decision

The floor plan should be readable from 3 meters away. Color is faster than text. Before this change, finding a reservation required searching every table — after, status was a color, and the reservation list handled search.

Floor plan — color-coded table status, multi-floor
04
UI Simplification

The main view tried to surface everything simultaneously — counters, icons, toggles, and controls competing for attention. We restructured the navigation, introduced a dark-themed data table with avatar-based customer identification, clear filter chips, and a high-contrast primary action button. Secondary actions were grouped and deprioritized.

Design decision

Removing features from view isn't the same as removing features. Staff needed 3 primary actions for 90% of their service interactions. The redesign put those 3 in reach — everything else accessible but not competing for attention.

Dark data table — filters, avatars, clear hierarchy
05
Design System

Every inconsistency in the original UI added to cognitive load. We defined an 8px base grid, standardized border radius (4px and 8px), selected Rubik as the primary typeface (Regular and Medium), and established a communication tone: warm, approachable, non-technical.

Design decision

Rubik and rounded corners weren't aesthetic choices — they communicate "non-technical" intentionally. The previous interface read like a developer tool. A hospitality platform serves people focused on guests, not systems. The design system made that distinction in every element.

Design system — 8px grid, Rubik, border radius, tone
06
Integrated List + Floor Plan View

Staff described a core frustration: finding a reservation required the list, placing it required the floor plan, confirming it required switching back. Three contexts for one task. We redesigned the main view to show both simultaneously — list on the left for search, floor plan on the right for spatial context. Actions in the list update the floor plan in real time.

Design decision

The two views weren't separate tools — they were two perspectives on the same data. Showing them together wasn't a layout decision, it was a workflow decision. It reduced the most common source of context-switching to zero additional clicks.

Integrated view — reservation list + floor plan side by side
07 — Outcomes

A clearer tool. A measurable goal. An incomplete project.

The goal was to reduce support call volume driven by interface confusion. The support team's documentation made it possible to measure this — every redesigned interaction had a corresponding support category. The design was delivered. Implementation was in progress when I left.

Reservation card rebuilt from scratch — hierarchy anchored on the 2-second scan: name, time, table, status
Reservation list redesigned — color-coded status borders, improved typography, status tab navigation with counts
Floor plan rebuilt — color-as-status visual language, multi-floor support, stacked reservation blocks
UI simplified — restructured navigation, primary action prioritized, visual noise reduced
Design system established — 8px grid, Rubik, rounded corners, warm hospitality tone
Integrated view designed and prototyped — list + floor plan simultaneously, zero context-switching
08 — Key Takeaways

Three things that changed how I think about operational tools.

01
In high-pressure environments, clarity is a survival feature
Restaurant staff don't have cognitive budget to spare during service. Every second spent deciphering the interface is a second taken from a guest. Design for the moment of use, not the moment of setup.
02
Support logs are the usability study you didn't have to run
The support team had been documenting user failures for months. That data shaped the redesign scope more than any interview session. In products with active support, the backlog is your research.
03
A design system is a speed investment, not an aesthetics investment
Every time a user encounters a pattern they've seen before, they process it faster. The design system wasn't about visual polish — it was about making the interface legible under pressure, at scale.
09 — What I'd Improve Next
1
Conduct usability testing with live restaurant staff during actual service hours — edge cases only appear under real pressure.
2
Explore a mobile-first version for floor staff who need status updates away from the main screen.
3
Add real-time analytics for managers: covers seated vs. reserved, average turn time, peak hour patterns.
4
Return to validate whether the redesign reduced support call volume — the metric was defined, measurement was pending when I left.
Closing Thought

Great restaurant experiences start behind the scenes — with tools that let staff focus on people, not processes.

Designing for operational environments taught me that clarity and speed aren't design goals. They're business requirements. And a support team's documentation is the most honest usability test you can have.