Portfolio / Project 03
Fulfillment Operations Control Center
A unified command interface for fulfillment center floor managers — combining real-time zone monitoring, associate task management, equipment status, and safety alerts into a single glanceable view designed for warehouse conditions.
Context
Concept Project
Role
Lead Product Designer
Timeline
5 months
Platform
Desktop + Handheld
Domain
Logistics Operations
Year
2024
The Problem
Fulfillment center floor managers operate in a uniquely demanding environment: they walk 8-12 miles per shift across 800,000+ sq ft facilities, managing 100+ associates across 6 operational zones while monitoring real-time throughput, equipment status, and safety compliance. Their current toolkit is fragmented — five separate systems, none designed for mobile or glanceable use.
The consequence: managers default to radio calls and walking to stations for basic status updates. Shift handoffs happen verbally, losing critical context. Equipment failures cascade because alerts sit in email queues. Safety near-misses go unreported because reporting requires a desktop login.
This project extends two core experiences: building the AIDE unified desktop that consolidated 23 tools for 80,000 Fidelity associates, and designing operational systems for physical environments at Shoft Shipyard.
Research and Discovery
Contextual Inquiry
I shadowed 8 floor managers across 3 fulfillment centers over 2 weeks, observing full 10-hour shifts. I mapped every tool interaction, radio call, and physical movement.
Floor Managers8
"I check five screens before I even start walking. By the time I get to a problem, it's already escalated."
Area Managers4
"Shift handoffs are our biggest information gap. We lose 20 minutes every transition."
Safety Leads3
"Near-miss reporting drops 60% on night shifts because nobody wants to walk to a terminal."
Process Assistants6
"The rate boards update every 15 minutes. By then it's too late to fix a bottleneck."
Pain Point Frequency
Mapped across 21 interviews by severity and frequency.
No single-pane operational view19/21
Shift handoff context loss18/21
Equipment alerts buried in email17/21
Safety reporting requires desktop16/21
Rate metrics lag 15+ minutes15/21
Cannot manage tasks while mobile14/21
No predictive bottleneck alerts12/21
Associate performance not real-time10/21
Floor Manager Journey Map — 10-Hour Shift
Physical movement, tool usage, and emotional state across a full shift
Shift Start
6:00 AM
Check 5 systems, read emails, walk floor
Mood: Neutral
20 min before walking
Desktop x5
Morning Ops
6:30-10:00
Monitor zones, reassign labor, fix jams
Mood: Frustrated
Radio-only updates
Radio + memory
Mid-Shift
10:00-12:00
Break mgmt, rate coaching, safety walk
Mood: Strained
No mobile dashboards
Clipboard + radio
Afternoon Push
12:00-3:00
Peak volume, bottleneck response, escalate
Mood: Overwhelmed
Reactive, not predictive
Radio + walking
Shift Handoff
3:30-4:00
Verbal debrief, open issues, priorities
Mood: Depleted
Context lost in translation
Verbal only
Tool Fragmentation Audit
5 disconnected systems, none mobile-optimized
System 1
Rate Tracker
Desktop only
15 min lag
No mobile view
System 2
Labor Mgmt
Desktop only
Manual refresh
No floor access
System 3
Equipment Mon.
Control room
Email alerts
Alerts buried
System 4
Safety System
Desktop only
End of shift
Low adoption
System 5
Shift Notes
Verbal + email
N/A
No structure
Information Architecture
Physical-to-Digital Mapping
Navigation mirrors the physical warehouse. Each screen answers a question managers ask dozens of times per shift.
Floor Plan
"Where are the problems right now?"
Constant
Process Flow
"Is the pipeline flowing or stuck?"
Every 10 min
Associates
"Who needs help? Who's on break?"
Every 20 min
Equipment
"Is anything down or degrading?"
On alert
Safety
"Is everyone safe right now?"
Constant
Handoff
"What does next shift need to know?"
End of shift
Design Principles
01
Glanceable at 10 Feet
Wall-mounted displays use oversized KPIs, color-coded zone status, and motion-based alerts visible from across the warehouse floor.
02
One-Handed Interaction
Every critical action reachable with a single thumb. Managers hold handhelds in one hand while the other gestures or carries a radio.
03
Warehouse-Optimized Visibility
High-contrast color system designed for bright ambient light. Pastel accents on white backgrounds maximize legibility under fluorescent lighting.
04
Zone-First Mental Model
Navigation mirrors physical warehouse zones. The UI map matches the real environment so managers never translate between layouts.
Two Worlds Converge
At Fidelity, I led the AIDE unified desktop that consolidated 23 legacy tools into a single interface for 80,000 associates — the same consolidation challenge fulfillment center managers face.
At Shoft Shipyard, I designed operational systems for physical environments — monitoring real equipment, tracking real people, managing real safety risks. The gap between a control room screen and the factory floor is physics, not pixels.
Fidelity AIDE
Unified 23 tools for 80K users
Same consolidation challenge
Shoft Shipyard
Physical-environment ops systems
Same physical-digital bridge
Interactive Prototype
Explore the System
Six screens covering the complete operations lifecycle — from zone-level monitoring to automated shift handoff.
FC Ops Center
Floor Plan
Process Flow
Associates
Equipment
Safety
Handoff
2 alerts
01
Floor Plan Overview
Where do I need to be right now?
The primary command screen. Six fulfillment zones displayed as a spatial grid matching the physical warehouse layout. Managers see throughput, headcount, and bottleneck status at a glance — designed for 10-foot legibility when displayed on wall-mounted monitors. Active alerts surface inline with one-tap triage actions.
Total Throughput
14,309/hr
Active Associates
88/96
Bottlenecks
1
Equipment Up
94%
Safety Alerts
2
Shift Progress
62%
RECVReceiving
On Track
Throughput
94%
Associates
12
Items/Hr
2,847/hr
STOWStow
On Track
Throughput
87%
Associates
18
Items/Hr
3,102/hr
PICKPick
Bottleneck
Throughput
62%
Associates
24
Items/Hr
1,890/hr
PACKPack
On Track
Throughput
78%
Associates
16
Items/Hr
2,340/hr
SHIPShip
On Track
Throughput
91%
Associates
10
Items/Hr
2,680/hr
SORTSort
On Track
Throughput
85%
Associates
8
Items/Hr
3,450/hr
Active AlertPICK Zone Bottleneck Detected
Pick throughput dropped 38% in last 20 minutes. Kiva Bot Cluster B at 68% (2 units offline). Backlog building at Mod-A3.
Reassign 4 Associates
View Kiva Status
FC OPS CENTER v2.1 | AUS-FC-04REBECKA RAJ
Key Design Decisions
01
Glanceable Wall Display Mode
Problem
Rate boards updated every 15 minutes on static TV screens. Managers walk to control rooms for real-time data.
Solution
Adaptive layout detects display type. Wall mode shows oversized KPIs with 10-foot legibility. Handheld mode provides drill-down detail.
Before
WAREHOUSE RATE BOARD — TV DISPLAY
RECV
2,847
STOW
3,102
PICK
1,890
PACK
2,340
SHIP
2,680
SORT
2,410
Last updated 14 min ago
No color-coding. No status. No alerts. Static refresh every 15 min.
After
FC OPS CENTER — WALL MODELIVE
RECV
95%
STOW
97%
PICK
68%
PACK
90%
SHIP
96%
SORT
91%
ALERT: PICK zone bottleneck — Kiva Bot Cluster B at 68%
Color-coded status. Live alerts. 10-ft legible. Adaptive layout.
87%
fewer control room trips
Research insight:Contextual inquiry revealed managers averaged 14 control room visits per shift just to check rate boards.
02
Automated Shift Handoff
Problem
Verbal handoffs averaging 20 minutes. Critical context lost between shifts. New managers start blind.
Solution
Auto-generated structured reports: open issues ranked by severity, zone trends, equipment alerts. One-tap acknowledgment.
Before
SHIFT HANDOFF — VERBAL + NOTES
"So the Kiva bots in B were acting up since around noon..."
"Oh, and watch out for lane 3, I think there was a near-miss but I didn't get a chance to file it..."
~20 min
Context lost
No audit trail
After
AUTO-GENERATED HANDOFFREADY
Units
48,240
Uptime
94.8%
Safety
2 incidents
Utilization
92%
CRITICAL: Kiva Cluster B — 2 units offline
Confirm
Add Notes
4 min
handoff time (from 20 min)
Research insight:Shadow sessions showed incoming managers spent their first hour re-discovering issues the previous shift already knew about.
03
Predictive Bottleneck Detection
Problem
Managers discover bottlenecks by walking to zones or radio calls. By then, throughput has already dropped.
Solution
ML-powered flow analysis monitors transfer rates and predicts bottlenecks 15-30 minutes before they materialize.
Before
REACTIVE BOTTLENECK DISCOVERY
2:12 PMPick rate starts declining
2:24 PMQueue builds at Pack stage
2:31 PMAssociate radios manager
2:38 PMManager walks to PICK zone
2:45 PMRoot cause identified
33 min delay — throughput already impacted
After
PREDICTIVE BOTTLENECK DETECTION
2:12 PMML detects velocity pattern change
2:13 PMAlert: Pick predicted 32% below target in 25 min
2:15 PMManager taps: Reassign 4 Associates
2:18 PMAssociates redistributed to PICK
2:30 PMThroughput stabilized at 89%
3 min to action — bottleneck prevented entirely
82%
bottlenecks caught early
Research insight:Transfer rate analysis showed detectable velocity patterns 18 minutes before visible queue buildup.
Projected Results
87%
fewer control room trips
4 min
shift handoff (from 20)
82%
bottlenecks predicted early
3.2x
safety report submissions
Designed by Rebecka Raj — Concept Project, 2024
This concept extends real experience designing unified operational tools at Fidelity Investments and physical-environment systems at Shoft Shipyard.
Back to Projects