Warehouse Automation System

Identified an inefficiency on the warehouse floor and worked with a third-party WMS vendor to consolidate three fragmented inventory tables into a single intuitive view, reducing cognitive load and bridging the skill gap between experienced and new staff.

My Contributions

• Identified the inefficiency and pitched the consolidation to management and the WMS vendor (Thomax Technology)
• Authored the full technical specification — calculation logic for replenResidual, multi-condition filtering rules, edge case handling (negative → 0, bold positives), and an annotated example table — submitted directly to the vendor developer
• Caught a live-data bug mid-day after go-live: replenResidual values were static snapshots not recalculating on BulkBinCode updates; documented root cause and escalated for a dynamic fix
• Trained staff on the new workflow with practical tips and best practices

The warehouse relied on three separate tables that staff had to cross-reference simultaneously:

1. Inventory pallet contents (what items are where)
2. Collection status (items picked and ready)
3. Customer order tracking (item locations for the day)

This forced staff to mentally consolidate data across three views, increasing cognitive load and error risk during high-volume operations.

I authored a full technical specification and submitted it to the WMS vendor (Thomax Technology). The spec defined a new replenResidual column: ReplenQuantity − SUM(BulkBinCodes where location starts with “SUP” or equals “TABLE”). Entries where replenResidual ≤ 0 would be hidden; negative values clamped to 0; positives bolded. I included a full annotated example table covering all filter cases. The vendor developer (Dominic York) deployed the implementation and asked us to verify the output.

During live use I noticed entries showing 2.00 where 0.00 was expected. I traced this to static snapshot values: replenResidual was being calculated at query time but not recalculating when BulkBinCodes were updated, so the visibility filter was operating on stale data. I documented the root cause and escalated back to Thomax Technology for a dynamic recalculation fix.

• Mapped variables and dependencies to understand the full data landscape
• Identified essential vs. redundant fields to reduce cognitive load
• Coordinated tight deadlines via documentation, calls, and detailed PDFs

The spec I authored and submitted to the vendor defined the logic for a new replenResidual column:

Display Rules

• If replenResidual ≤ 0 — entry is hidden from the report
• If replenResidual is negative — clamp to 0 (never show negative)
• If replenResidual > 0 — entry is visible and the value is bolded
• Search input % shows all entries while applying the above rules

Annotated Example (Search input: TRA-FA-19%)

The BulkBinCodes column only shows non-search locations (TABLE, SUP prefixes). TRA-FA-19 is the current shelf location and is excluded from the SUM.

Submitted to the WMS vendor as part of support ticket THX-1373449. The vendor deployed the implementation on 11 December 2023 and requested verification of output accuracy — which is when the stale-data bug was caught.