PROFESSIONAL PROJECT · LOGISTICS OPERATIONS · NDA

ParCel

Realtime Operations Manager

Every lane, every handler, every shift, in real time.

ROLE

Sr. UX/UI Design Lead · Researcher

TIMELINE

2024 · 6 months

CONTEXT

Professional · Logistics · NDA

TOOLS

Figma · FigJam · Excel

[ HERO IMAGE — 4 device-wrapped screens: Home · Sort goals · Coverage · Volume forecast ]

PROCESS · 6 MONTHS · 4 PHASES

Six months, four phases.

Research found the problems. The problems set the features. The features hit the floor. The floor sent us back to redraw.

01

DISCOVERY

Sit with the floor.

4 workshops · 10+ interviews · 4 management levels

02

DEFINE

Find the patterns.

3 archetypes · 4 discovered problems · 3 journey maps

03

DESIGN

Structure, then skin.

IA · lo-fi for every flow · component library · hi-fi

04

TEST & REFINE

Put it in their hands.

5 usability sessions · 50+ observations · 3 redesigns

THE CHALLENGE

One manager. Ten areas.
A single shared terminal.

An operations manager runs a slice of a sorting hub through the overnight volume spikes: up to ten areas, each tied to a different shipping region. To see what's coming, they walk to a shared computer terminal and wait their turn. Staffing happens on the fly, on paper. Schedules go stale the second someone no-shows. Safety checklists get printed, filled by hand, and filed in a binder. The data to run the floor was all there. It just never reached the person running the floor.

10

areas

One ops manager's span of control, each a different region

Discovery workshops

1

terminal

Shared and stationary, the only real-time view, and you queue for it

Discovery workshops

100%

paper

Daily safety checklists, printed and filed by hand

Discovery workshops

“By the time I’ve walked to the terminal and logged in, the spike’s already on top of me.”

OPERATIONS MANAGER · DISCOVERY WORKSHOP

WHY THIS EXISTS

The dock was modernizing.
The manager’s tools weren’t.

The client was modernizing the sorting dock, and the operations manager was the person every shift leaned on. They make the call when a lane falls behind, when an area runs short on handlers, when tomorrow’s volume means calling people in tonight. Every one of those calls depended on data stuck at a terminal, on paper, or in a scheduler’s head. The brief was simple to say and hard to do: put real-time, decision-grade information in the manager’s hand, let them act on it without leaving the floor, and back it with volume prediction so staffing could get ahead of the spike instead of chasing it.

THE STATUS QUO

No tool met the manager where they worked.

Every tool the manager relied on assumed they were sitting still. The volume view lived on a desktop terminal. Staffing lived in hallway conversations and paper. Scheduling lived in a system that aged by the hour. Compliance lived in a binder. Nothing was real-time, nothing was mobile, and nothing was built for the role that had to react fastest. The gap wasn’t missing data. It was data that never reached the floor.

THE CORE GAP

No tool connected live volume → the manager on the floor → an action they could take in the moment: reroute a lane, pull a handler, call in coverage. That’s the gap ParCel fills.

WHO IT’S FOR · 3 ARCHETYPES

Three roles, one floor.

Workshops and interviews ran the full ladder, from regional and facility managers down to the package handlers. Three archetypes came out of it, each with a different job to do. The Operations Manager is the protagonist here; the app is built around their shift.

4 workshops · 10+ interviews · 4 management levels

PRIMARY

Operations Manager

SPAN

One facility section: up to 10 areas, each a different shipping region.

OWNS

Package flow, staffing, smooth daily ops; reallocates handlers to other managers’ areas during spikes.

JOBS-TO-BE-DONE

“When I log into my dashboard, I want to see the expected package volume for today and tomorrow, and know if I have adequate staffing. If I don’t, I need a way to quickly contact available employees from a pool of resources to fill the gaps.”

SECONDARY

Facility Manager

OWNS

The larger flow of packages in and out of the whole facility; builds systems and tackles structural problems.

JOBS-TO-BE-DONE

“When I log into my dashboard, I want to see any impacts on volume flow, including predicted spikes for the day, days, weeks, and months, so I can staff accordingly.”

RESOURCE

Package Handler

ROLE

The core floor workforce the manager schedules, reallocates, and supports.

TOUCHPOINTS

Shift performance inputs, punch-in correction, coverage requests, feedback.

JOBS-TO-BE-DONE

“When I start my shift, I want my punch-in to be right and to see which area I’m on, so I can get to work without tracking down a manager.”

RESEARCH SYNTHESIS · AFFINITY MAP

Everything we heard, clustered.

Workshop and interview notes went up on an affinity board, and the themes sorted themselves out. Four of them shaped the MVP and became ParCel’s discovered problems. The rest set the tone, the defaults, and the things we chose to leave out.

01

The terminal bottleneck

Managers queue for a shared, stationary terminal, with no quick real-time view of incoming volume.

→ Real-time home + sort

02

Staffing on the fly

No way to manage handlers live during a spike; clunky shift inputs; double punch-ins at shift start are a pain to fix.

→ Coverage from a resource pool

03

Schedules that age

Schedules go stale fast; no-shows are hard to backfill; no way to forecast staffing against expected volume.

→ Volume forecast + scheduling

04

Paper safety checks

Daily checklists on paper; no record when an incident happens; missed checks stay invisible.

→ Digital checklist + incident audit

IDEATION · PROBLEM → FEATURE

From many problems to main features.

Every problem got scored on two questions: how often did it come up across roles, and could a mobile feature actually fix it inside MVP scope? The four that scored highest became the build. The rest became backlog.

PROBLEM 01

Terminal bottleneck

Live Sort & Volume

The shift, in real time, in one hand.

PROBLEM 02

Staffing on the fly

Coverage Pool

See the gap, contact available staff, fill it.

PROBLEM 03

Schedules that age

Volume Forecast

Today and tomorrow’s volume against your staffing.

PROBLEM 04

Paper safety checks

Digital Checklist

Completion tracked, incidents auditable.

THE FRAMEWORK · 4 FEATURES

Four features, one shift.

ParCel turns a manager’s night into four jobs it does well. Each one traces straight back to a discovered problem.

01

LIVE

Sort & Volume

Real-time sort progress, hourly burndown, lane-by-lane status. Replaces the terminal.

📱 Replace with Screen Name

📱 Replace with Screen Name

📱 Replace with Screen Name

📱 Replace with Screen Name

📱 Replace with Screen Name

📱 Replace with Screen Name

02

COVERAGE

Resource Pool

See understaffed areas, contact available handlers from a pool, confirm coverage. The core job-to-be-done.

ANCHOR SCREENS

📱 Coverage

📱 Handler feedback

03

FORECAST

Volume ahead

Predicted volume today and tomorrow against current staffing, so you get ahead of the spike instead of chasing it.

ANCHOR SCREENS

📱 Volume forecast

04

SAFETY

Digital checklist

Daily safety checks done in-app, visible to managers, auditable after an incident.

ANCHOR SCREENS

📱 Safety checklist

📱 Checklist

Supporting flows: Attendance · Shift handoff · Doors · Van lines · Midnight load · Door stats

CONCEPT EVOLUTION · DIRECTIONS CONSIDERED

A few directions we tried first.

Before the four-feature mobile tool, a few other directions got explored. Two were tried and dropped. One was right, but out of scope for this release.

REJECTED

Desktop web dashboard

TRIED IT BECAUSE

The project started life as a desktop web app, the obvious home for a data-dense ops dashboard.

IT FAILED BECAUSE

Managers don’t sit at a desk during a spike; they’re out on the floor. A desktop tool just rebuilt the terminal bottleneck it was meant to kill.

WHAT SURVIVED

The move to Android mobile, glanceable and one-handed at the point of work. That pivot turned out to be the project’s biggest insight.

REJECTED

Spatial hub-map metaphor

TRIED IT BECAUSE

A literal map of the facility’s areas and lanes felt like the intuitive way to show flow.

IT FAILED BECAUSE

It didn’t shrink to a phone, and managers thought in lanes and status, not geography.

WHAT SURVIVED

The lane row and the status grammar (ON PACE / AHEAD / BEHIND) instead of a map. Spatial layout is parked for a possible iPad or kiosk version.

DEFERRED · OUT OF SCOPE

Automated staff reassignment

THE IDEA

If the app already knows volume and availability, it could reassign handlers on its own.

WHY WE PARKED IT

Managers wanted to keep the call and the relationship: an action they trigger, not one done to their crew. It also sat outside MVP scope.

WHAT SHIPPED INSTEAD

The Coverage request flow. The app finds the gap and the pool; the manager makes the move. Auto-assignment lives in the backlog.

WIREFRAMES · LO-FI

Four flows, framed early.

Each feature became a flow, wireframed at phone width with auto-layout from the first sketch so the structure carried into hi-fi unchanged. Below, the payoff screen from each flow.

FLOW 01

Live Sort & Volume

Payoff: the Home shift card

FLOW 02

Coverage Pool

Payoff: contact-the-pool confirmation

FLOW 03

Volume Forecast

Payoff: tomorrow vs. staffing

FLOW 04

Safety Checklist

Payoff: completion + incident audit

HI-FI

Built for 2am, under 1000 lux.

The system carries its weight without raising its voice. Lime marks what’s live and on pace. Warn-red is rationed for the one lane that actually needs you. Numbers do the talking: big, tabular, and never sitting next to nothing, always paired with a comparator. The cream hero card is the single warm surface on the screen, and it flips between light and dark so the thing you look at most always lights up.

Home

Sort goals

Current sort

Doors

Van lines

Midnight load

Door stats

Handler feedback

Attendance

Checklist

Coverage

Volume forecast

Safety checklist

Shift handoff

DESIGN SYSTEM

One system, two modes, eight atoms.

ParCel runs on a small, strict system, so every screen reads the same way at a glance. Geist sets the type for display, body, and numbers; Geist Mono handles codes, lane IDs, and timestamps; Source Serif 4 is there as an optional old-style numeral. Lime is the default accent and can be retuned to four other swatches, but the status colors, warn-red and info-blue, stay fixed. Light and dark are siblings, not a parent and a tint; each owns its full ink and surface scale. Eight atoms do almost all the work, and the whole thing sits on a three-tier token system, so swapping the typeface or flipping a mode is a single change.

ATOMS · BUILD REFERENCE

StatusPill

Sparkline

BurndownChart

SegBar

Delta

LaneTile

Avatar

Ticker

ShiftRing

NUMBER SCALES

Hero KPI 30px · screen-level sort percentage 56px

WHAT MADE THIS HARD

Four constraints that shaped the work.

01

PLATFORM

WHAT HAPPENED

The project began as a desktop web app, then pivoted to Android mobile mid-stream.

TRADE-OFF / DECISION

Rebuilt the IA for one-handed, glanceable use on the floor. The pivot was the insight.

02

CONTEXT

WHAT HAPPENED

The user is on a 1000-lux floor at 2am, moving, often gloved.

TRADE-OFF / DECISION

Numbers, a status color, and a status verb. No dense tables, no color-only signals.

03

STAKEHOLDERS

WHAT HAPPENED

Workshops spanned regional, facility, ops, and handler, each wanting their own view.

TRADE-OFF / DECISION

Role-aware JTBD, with the Operations Manager set as primary and the others secondary.

04

SCOPE & NDA

WHAT HAPPENED

A real client, a 6-month window, and an NDA.

TRADE-OFF / DECISION

Ship the MVP that clears all four discovered problems; document the rest as backlog.

TEST & REFINE · 5 SESSIONS · 3 REDESIGNS

The floor sent us back to redraw.

We put working screens in managers' hands under shift conditions: phone in one hand, on the move. Fifty-plus observations sorted down to the handful worth a rebuild. Three of them changed the design.

REDESIGN 01

Grading a borrowed handler

BEFORE

AFTER

WHAT TESTING SURFACED

Every handler we tested typed a sentence or two, then stopped and asked what the lead was actually supposed to do with it. The free-text box captured a mood, not a signal. Two leads said they skim these at 4 a.m. A paragraph is the last thing they want mid-shift.

“I can write ‘good worker’ but that tells the next lead nothing. Tell me what to rate — speed, safety, whatever — and I’ll tap it. I’m not writing an essay at the end of a twilight.”

Renata · Twilight sort lead · Usability test, 05/26

“If there’s no scale, everybody grades different. My ‘fine’ is somebody else’s ‘great.’” — Glen, dock handler

WHAT CHANGED

The free-text box became structured prompts: overall performance, strongest area, area to develop, lead-rotation flag, each a tap, not a sentence. Notes are now optional. A lead finishes a handler in under fifteen seconds and the next shift reads a signal, not a paragraph.

REDESIGN 02

Which doors are actually behind?

BEFORE

AFTER

WHAT TESTING SURFACED

The grid showed a number for every door, but testers couldn’t tell a healthy door from a failing one without reading each cell. Managers scanned left to right doing the math in their heads, and still missed the two doors that were red. The data was all there; the priority wasn’t.

“There’s twelve numbers on here and they all look the same weight. I need the floor to jump out at me — show me the three doors killing my hour, not all twelve at once.”

Mara · Operations manager · Usability test, 05/26

“By the time I’ve read every box the belt’s already backed up. I want to glance, not study.” — Cody, unload supervisor

WHAT CHANGED

Raw per-door counts became an hour-by-lane heatmap. Actuals fill in green, gaps stay dark, and exceptions surface as their own list below. A manager sees the shape of the shift in one look and reads the problem lanes by name: belt stop, staffing short, upstream delay.

REDESIGN 03

Where do I even start?

BEFORE

AFTER

WHAT TESTING SURFACED

The all-areas view listed every facility area with the same visual weight: pkg/hr, lanes full, percent loaded, all in one flat row. Testers said it read like a spreadsheet. Nobody could answer the only question that mattered at start of shift: which van do I work first?

“It’s all numbers and they’re all gray. I’m hunting for the one about to seal and I can’t find it without reading every line. Just group them — tell me what’s docked, what’s waiting.”

Marcus · Facility manager · Usability test, 05/26

“Percent loaded doesn’t help me at 2 a.m. Status does — is it docked, is it sealing, is it gone?” — Junie, load coordinator

WHAT CHANGED

The flat facility list became van lines grouped by bay, each van carrying a plain-language status (Docked, Sealing, Staging, Waiting) and a real departure time. The percent-loaded math moved to the background. The first thing you see now is what’s about to leave.

5

moderated sessions · ops + facility managers

50+

observations logged

3

redesigns shipped

~10

iterations per screen

OUTCOME & LEARNINGS

A real system, ready to build.

ParCel covers the operations manager’s night in 14 screens (10 canonical plus 4 ops flows), in both light and dark, on a component library and a three-tier token system, backed by three personas and their jobs-to-be-done. There’s enough here to hand to engineering and keep building.

01

A naked number is a bug.

At 2am, a number with nothing to compare it to is just noise. Every primary figure on screen earns a delta, a sparkline, or a goal it’s measured against.

02

Status is a color and a verb.

Lime and warn-red never work alone; they ride with ON PACE, AHEAD, and BEHIND. The color catches the eye, the word carries the meaning, and nobody reads the floor on color alone.

03

Meet the manager where they work.

The whole project hinged on one pivot: the fix wasn’t a sharper dashboard at the desk, it was real-time in the hand, out on the floor.