4,041 pieces of gear, one sticker each, full history on a phone scan.

The problem: an inventory that couldn't be trusted

• The data set was old, denormalized, and dirty — years of FileMaker exports had interleaved item records with audit history, carried orphan rows, and let bad values sit unnoticed.
• Every lookup meant opening a spreadsheet. Walk up to a mic or a rack unit and there was no fast way to know its serial, cost, location, or repair history.
• Service history lived nowhere consistent — repairs happened, money got spent, and none of it was tracked against the specific piece of gear.
• Replacement decisions had no evidence — the case to replace an aging item came from memory, not a record of what had been sunk into fixing it.

What we built

A tracking system rebuilt from the messy export up, plus a custom mobile scanner app that puts every item's history one scan away.

• A cleaned, normalized system of record. The FileMaker export rebuilt into a clean workbook and a mirrored SQLite database — item records split from history, orphan rows removed, one row per tag.
• A unique barcode for every item — 4,041 of them. Every piece of gear gets its own printable QR label.
• A custom mobile scanner app that reads anything. Point a phone at a tag and the item's full card appears instantly — manufacturer, model, serial, location, cost, complete service history. Reads QR plus Code-128 and every common format; type-the-tag entry works when a sticker's worn.
• Service and repair tracking per tag — staff log service dates and repair costs against the specific item, building a maintenance history that never existed before.
• Automatic replacement flagging — the system auto-flags any item where repair spend has passed half its purchase price.

The results

• 4,041 devices, each individually tagged, scannable, and reconciled into one clean record — migrated up from the old FileMaker export, not bolted onto it.
• A worn, untrusted database became a system of record staff actually use — deduplicated, with the orphan and bad-data rows stripped out.
• Full gear history in seconds, from a phone, anywhere on the floor — no spreadsheet, no guessing which export was current.
• Replacement pitches now carry evidence — items where repairs exceeded half the acquisition cost surface automatically.

Why it worked

The program didn't need a new inventory — it needed the one it had to become trustworthy and reachable. The win wasn't collecting data; it was cleaning the data that existed and putting it where the work happens: on a phone, at the gear, one scan away. And it was built to grow up — the mirrored database is the backbone for a future integration that writes repairs automatically as service tickets close.