← All work
06

Case Study · BetterPlace · Attendance · 2021

Shift planning that took 48 minutes — made instant.

Role
Product Designer
Timeline
4 weeks
Users
Supervisors · Factories · Workers
Outcome
Clients returned from paper
48
Minutes to assign shifts for 50 employees for 2 weeks
2.5hr
Daily time supervisors spent managing rosters
6
Duties lost per week — direct production line impact
↩
Clients abandoned digital and returned to paper
Context
Roster illustration

Clients went back to paper.
That's the failure signal.

A roster is the tool supervisors use to plan, organise, and manage shifts for blue-collar workers. In Indian manufacturing, a machine that sits idle causes significant production loss — roster management is high-stakes, high-frequency.

BetterPlace's Attendance product had a roster feature. It worked with small data sets. As client roster sizes grew, the system struggled. What was manageable became so painful that clients abandoned the digital tool entirely and returned to manual Excel sheets and paper records.

When clients return to paper, the product has failed them fundamentally. Not as a feature gap — as a breakdown of trust.

Who was suffering

Supervisors
2.5 hours/day on roster management. Burned out. Returned to manual systems.
Blue-collar workers
Missed shifts. Lost pay. Missed overtime. No visibility into their own schedule.
Companies
6 duties lost per week. Direct production cost. Mismanagement cascading into losses.
System architecture flow diagram showing roster relationships between Location, Daily/Weekly view, Shift details, Add shifts, and Publish shift

Roster system architecture — mapping how location, shifts, OT, and publishing connect before designing any screen

Research

The supervisors weren't looking at the roster the way we designed it.

We ran in-person interviews with actual supervisors, board members, and production managers. Pandemic constrained access — it didn't eliminate the research.

"The supervisors were seeking at a different angle for roster. They want it handy. Fast. They should be able to get to the roster and know where is man."

Key insight 01

The primary question is spatial, not temporal

"Where is man?" — not "what is the schedule for Tuesday?" Supervisors think in positions (who is covering this machine, this zone) not time slots. The entire old roster UI was organised around the wrong axis.

Key insight 02

3–5 minutes to understand the current state and make one change

Supervisors navigated three separate screens to understand the current state, find the right employee, and make an edit. By the time they finished, the situation on the floor had changed.

Key insight 03

Backup planning was entirely in supervisors' heads

Workers call out sick, leave suddenly, miss shifts. Supervisors maintained a mental list of backups. The system had no concept of backup assignments. All contingency planning happened outside the product.

Key insight 04

12 hours/week reconciling roster against attendance

Two systems, no integration. Planned vs. actual comparison was done entirely by hand. 12 hours a week of manual cross-referencing.

Roster wireframe explorations showing different layout approaches with sticky notes and annotations about key design decisions

Wireframe explorations — iterating on layout, information density, and the shift assignment flow with active annotation

Key decisions

Fast to load. Faster to understand. Instant to edit.

Primary axis

Position-first view, not time-first

The new roster leads with positions (zones, machines, roles) as the primary structure. A supervisor opening the roster immediately sees "who is covering what." The answer to "where is man" is visible without navigation.

Speed

One-tap shift assignment — not three screens

Old flow: navigate → find day → click edit → select shift → confirm. Five steps, three screens. New flow: tap slot → select from filtered available workers → done. Two taps. One screen. Faster than paper for the first time.

Resilience

Backup as a first-class concept

Every shift slot supports a primary and backup assignment. Backup is visible in the roster view. When a primary worker is absent, the backup surfaces automatically. The contingency plan is now part of the plan.

Integration

Roster and attendance in one view

Planned vs. actual is visible on the same screen. The 12 hours/week of manual reconciliation is eliminated — the system does the comparison automatically.

Outcome

Clients came back from paper.

↓85%
Time to plan a 2-week roster for 50 employees
0
Manual reconciliation — roster and attendance now linked
↑
Client retention — adoption reversed after launch
✓
Backup planning now in the system, not in heads
What I learned

When clients return to paper, listen to the paper.

1
The manual process is the benchmark.Understanding exactly why the Excel sheet worked — what it did that the digital tool didn't — was the entire design brief. The paper was right. The product was wrong.
2
The primary question defines the primary axis."Where is man?" is a spatial question. Organising the roster around time was a design assumption never validated. One research session changed the entire architecture.
3
4 weeks with focus beats 4 months without it.Two designers, one PM, pandemic restrictions, four weeks. Constraints produced clarity. Every decision was ruthlessly prioritised. The shipped product was better for it.
← Back to all work