Reducing picking errors by 27% for a parts manufacturer
We implemented a simple barcode scanning module at a plant in Gliwice. Before our visit, every fourth order was returned for correction due to packing mistakes.
The plant in Gliwice struggled with massive errors when packing small metal parts. We introduced a scanning system that eliminated human error at the goods issue stage.
The challenge
The manufacturer sent an average of 83 shipments per day to automotive industry clients. The problem was a huge number of nearly identical sleeves and washers. In March 2024, an audit showed that as many as 24% of shipments contained at least one incorrect element.
Every mistake generated a cost of about 340 PLN — including return transport, re-shipment, and office worker time spent clarifying the matter. Warehousemen worked under time pressure and relied solely on sight and paper packing lists, which with 47 different types of products was a recipe for failure.
Our approach
Our team of 3 implementers spent 6 business days directly on the floor in Gliwice. Instead of theoretical assumptions, we observed the work of Marek and Grzegorz, who were responsible for packing.
We identified 14 critical points where assortment swapping most frequently occurred. We chose a small-steps method — instead of replacing the entire ERP system, we decided to build a lightweight verification module that works with the company's existing database. We focused on making the shop floor earn, rather than wasting time correcting the same order twice.
The solution
We provided 4 rugged barcode readers integrated with a simple tablet interface. The system works in a binary mode: a green light means the product is correct, a red light blocks the ability to print a shipping label.
We created a database of 114 unique codes for every part, even those that hadn't been marked before. The employee no longer needs to read complex symbols on cartons. A single scan click is enough. The entire software was written to make work easier, not add redundant menu clicking.
Results
After system implementation in August 2024, complaints fell by over a quarter. The company regained warehouse flow and stopped losing money on free returns.
Timeline
-
March 2024On-site visit and error counting at packing stations.
-
May 2024Code database preparation for 114 product types.
-
June 2024Module testing at 2 stations for 11 days.
-
August 2024System launch across the warehouse and team training.
"Silesia Growth Hub didn't promise miracles; they showed us where our cash was leaking. The scanners work, and I finally don't have to answer calls from angry recipients in the morning."