A standardized way to certify every sorter before they handle live packages — and the first time supervisors have data on which sorters are struggling, which belts cause the most errors, and which ZIP codes generate the most confusion.
There is no consistent method to verify knowledge before someone handles live packages on the floor. New hires absorb whatever the person standing next to them happens to know.
A misrouted package requires manual reprocessing, delays the recipient, and increases the chance of damage in transit. The cost compounds — not just per error, but per error multiplied by hours of downstream disruption.
No data on which sorters are struggling, which belts cause the most errors, or which ZIP codes generate the most confusion. Coaching was based on instinct rather than evidence — and rarely targeted the people who actually needed it.
One artifact. Three roles. Five non-negotiable properties — chosen to make the platform deployable today, not pending an IT roadmap.
Runs in Chrome or Edge. No installation, no app store, no IT ticket required.
No internet connection needed. All 41,000 ZIP routing records load locally in the browser.
Point the app at a network folder once. Every supervisor machine sees the same live data.
Sorter / Supervisor / Admin. Each sees only what they need — nothing more.
Built entirely in-house. No subscriptions, no cloud fees, no vendor dependency.
A randomized UPS shipping label appears. The sorter reads the routing code from the label's service bar — OH 443-03 for the ZIP 44303, where 443 is the first three digits of the ZIP and 03 is the last two — and chooses the correct destination belt from four gold tiles.
Distractors aren't random. They're drawn from confusable neighboring belts, so a correct answer reflects real routing knowledge rather than process of elimination.
Sender, recipient, address, weight, and tracking number all randomize per question. The sorter never sees the same label twice in a session.
Keys 1–4 select an answer. Space or Enter advances. No mouse required. The sorter can answer at floor pace — every interaction is keyboard-native.
end-of-session grade
Access is controlled by employee ID against the roster, with a shared password per role. Each role sees only what it needs — the sorter never sees analytics, the supervisor never sees the admin panel, and the admin doesn't waste their day on the quiz interface.
The supervisor's tier is where most of the operational value lives. Eight tabs, each answering a different question that used to require walking the belts. Data refreshes as sorters complete quiz sessions — and every answer ever submitted is stored and reviewable, per sorter, across a rolling 7-day accuracy trend.
Real-time team accuracy, top problem belts, and sorters needing attention — all in one tab. The first thing the supervisor opens at the start of shift.
Every belt ranked worst-to-best by error rate. Click any belt to drill in: which sorters miss it, how often, and which wrong belt they chose instead.
Individual drilldown: accuracy over time, per-belt performance table, top missort patterns, aisle rank. Every answer ever submitted, in one view.
Aisle-level leaderboards. Compare the four aisles head-to-head on accuracy, completion, and certification readiness.
A grid showing which belts get confused with which other belts. The diagonal is correct routing; off-diagonal cells reveal the actual error patterns.
Routing overlay management — apply temporary routing changes (peak season rebinds, belt closures) and audit when they were in effect.
Who is enrolled, who is active, who is certified, who has lapsed. Drives quiz eligibility and supervisor coaching priorities.
Sorter flag review and escalation — anomalies the system surfaced for human judgment. Resolve, escalate, or annotate; the trail persists.
Every sorter takes the same standardized test before handling live packages. Performance is measurable and comparable across all four aisles — no more "who trained whom" debates.
Supervisors identify struggling sorters before errors occur on the floor. Coaching is data-driven and targeted — not guesswork. The Belt Stats tab tells the supervisor exactly which conversations to have.
When routes change, admins update the truth table in the browser. Changes take effect immediately on all machines — no retraining sessions, no PDF reprints, no cohort regression.
The whole platform is a folder of static files. Drop it on a network drive every supervisor machine can reach, open index.html in a browser, point at the shared-data folder once — and the system is live across the facility.
Copy the project folder to a shared network drive that every supervisor machine can see.
Each supervisor machine opens index.html in Chrome or Edge. No install step.
On first launch, point the app at the shared-data/ folder. Once per machine.
All sort events, overlays, and roster data sync automatically across every machine.
Sorters log in with their employee ID. No account creation, no onboarding flow.
The 12-slide project document as presented internally. Includes mockups of the landing page, the quiz, the supervisor dashboard tabs, and the admin panel.