← Back to shift-keeper case
shift-keeper · deliverable

Text-based Storyboard — Shift Keeper

Kit 23 · Shift Keeper · deliverable · designed and built by I Wayan Usada · fictional employer, illustrative figures

Design premise

Absence is up and every meeting has a favourite fix. Revised after the build so it matches the 2D wall and the 3D week landscape; both read signals-data.js.

Step 1 · Read the wall (all nine, scored)

Each signal is a button. Select it, read it, name its root cause: workload design, management practice or personal circumstance. Each reading that matches the evidence scores 1. Interventions stay locked until all nine are read (the earlier build unlocked them after three, contradicting its own "read all nine" rule, and did not score the readings at all although it called them "the designed verdict").

SignalEvidence points to
Absence curve (Monday clustering)Workload design
Meeting load audit (23 h, 11 without decisions)Workload design
Handover gapsWorkload design
Tooling complaintsWorkload design
Exit interviews (late rosters)Management practice
Rota-swap black marketManagement practice
Appraisal skips in the highest-absence teamsManagement practice
The viral yoga pollManagement practice (a vacuum), and a decoy as an intervention
Caregiver clusterPersonal circumstance, made worse by policy

So the evidence is 4 workload, 4 management, 1 personal: eight of nine signals are organisational. The earlier text said "two-thirds … the third family", which did not match its own key.

Step 2 · Two interventions

Roster redesign, publish-early roster policy and meeting diet score 2 (structural); caregiver adjustment pathway 1 (honest, niche); handover protocol 0 (right fix, wrong size for two picks); resilience training and yoga 0 and count as wellbeing-washing.

Scoring and pass rule

Nine readings + two picks × 2 = 13. Pass = 10 or more, at least one structural fix, and no wellbeing-washing pick.

3D week landscape

One warehouse week as terrain (weekdays across, teams deep, height = absence). Each read signal pins onto the feature it explains, coloured by the cause the learner chose. Applying the two picks reshapes the week in an illustrative model: only workload-shaping fixes lower the Monday ridge; wellbeing-washing picks leave a wireframe ghost of the unchanged ridge.

Prefer the source? Download the raw markdown.