InternalWebDashboard

Microsoft - DISH

Client

Microsoft & Compass Group

Scale

4 Locations

Microsoft - DISH

Role

UI/UX Designer

Timeline

2 years

Tools

Figma, FigJam, Notion, Jira

The Problem

"Facility managers had full visibility into what was planned. Zero visibility into what was actually happening."

Issues were reported manually, hours after they occurred. By the time a manager knew something was wrong, guests had already experienced it. The data existed — equipment sensors, service logs, supply records — but it lived in disconnected systems with no unified view and no real-time signal.

What Made This Hard

01

Issues were invisible until they became problems

Dining service failures — stations down, supply shortages, equipment issues — were only reported when a guest or manager noticed. By then, the damage was done. There was no early warning system.

02

Data existed but lived in silos

Each location had its own reporting method. Some used spreadsheets, some used verbal handoffs, some used nothing at all. There was no unified view across locations, making it impossible to spot patterns or act proactively.

03

Managers were drowning in noise

Early prototypes showed everything equally — every metric, every location, every timestamp. Managers couldn't find what mattered. The challenge was designing for urgency, not completeness.

Process

01

Stakeholder Alignment

Ran a discovery workshop with operations, IT, and management to align on what 'healthy' actually meant for each location. Defined alert thresholds collaboratively — no assumptions about what numbers triggered action.

02

Information Architecture

Mapped the data hierarchy from network-level health down to individual station status. Designed for progressive disclosure — summary first, detail on demand. Managers needed a 10-second read, not a report.

03

Dashboard Design

Iterated through 5 layout directions. Landed on a status-first design where critical issues surface immediately without requiring any interaction. Color was the only visual signal we allowed ourselves — used sparingly and consistently.

04

Validation

Tested with 4 facility managers over 2 weeks in a live-prototype environment. Watched them work. Timed how long it took to identify a critical issue. Iterated until the answer was always under 30 seconds.

Key Screens

Health Overview

DISH Health Dashboard — multi-location status overview

The full picture, at a glance

Every location. Every service station. Every status — all visible from one screen. Critical issues surface at the top automatically. Managers don't need to look for problems. Problems come to them.

Alert Detail

DISH Alert Detail — incident timeline and response actions

From alert to action in under a minute

When something goes wrong, the alert detail view gives managers everything they need to respond — what happened, when, who's affected, and what actions are available. No digging through logs.

Station View

DISH Station View — individual station health metrics

Station-level detail when you need to go deeper

Each service station has its own health timeline, metric history, and maintenance log. For managers who need to investigate, the data is there. For managers who just need a status check, the overview is enough.

Outcomes

70%

Faster alert response time across monitored locations

100%

Real-time coverage across all 4 locations

30s

Time to identify a critical issue from dashboard open

Learnings

Hierarchy is more important than completeness

The instinct is to show everything and let users filter. The right answer was to make decisions for them — surface what matters, bury what doesn't. Every element on screen should earn its place.

Design for the worst moment

This dashboard would be used most when things went wrong. Designing for calm, normal operation meant it would fail exactly when it was needed most. We had to design for stress.

Color discipline is a feature

We committed early: color only means status. Nothing else is colored. That decision made every colored element immediately meaningful and reduced the time to understand what was wrong.

What I'd do differently

I'd push for a mobile-first version earlier. Most managers weren't at a desk when incidents happened — they were on the floor. The responsive design came late and could have shaped the information architecture from the start.

Next project

Component Factory

View work →