Construction

Recovering 3,400 PLN monthly in a construction warehouse

We found errors in supplier discount calculations using a simple Big Data analysis algorithm. The process took us 12 days.

+3,421 PLN savings/mo
ClientBud-Max Lublin
IndustryConstruction
TimelineApril – May 2024

At Bud-Max warehouse, money was leaking through small errors in invoices from large suppliers. We showed that 12 days of data work is enough to stop this cash drain without hiring new people.

Big Data AnalysisInvoice AutomationPythonAnomaly DetectionPower BI Reports

The challenge

The warehouse owner, Mr. Marek, handles 12 main suppliers of bulk materials and construction chemicals. On average, 156 invoices enter the company every month. Supplier systems often made mistakes when calculating discounts for early payment or volume thresholds. The accountant spent 14 hours a month manually checking papers, and even then, she only caught the biggest mistakes. The company was losing money because no one had the energy to count every bag of cement individually.

Our approach

Our team (3 people) entered the company for a two-day document audit. We downloaded data from their warehouse system for the last 7 months. Instead of looking for everything at once, we focused on the 4 most expensive product categories. We wrote a simple script that automatically compared prices from signed contracts with what actually appeared on PDF invoices. We didn't promise miracles, just clean facts extracted from tables.

The solution

Lublin Logic Services implemented a tool that scans the folder with new invoices every morning. The system automatically calculates how much a given product should cost according to the wholesale price list. If the difference is more than 3 groszy, the accountant gets a red alert on the screen. We also added a module for predicting warehouse shortages, which suggests when to order goods to hit a better price threshold with the manufacturer.

Results

In the first full month of operation, the algorithm pointed out 43 erroneous invoices that would have gone unnoticed. Bud-Max recovered exactly 3,421 PLN, and the employee's invoice checking time dropped from 14 hours to just 18 minutes a day.

3,421 PLN
Recovered amount in 1st month
18 min
Daily invoice control time
43 units
Detected settlement errors
7 mo
Historical data analysis

Timeline

  1. April 12, 2024
    Workshop with accounting and synchronization of data from the warehouse system.
  2. April 19, 2024
    Algorithm tests on historical data from the last 7 months.
  3. April 26, 2024
    Launch of the control panel for Mrs. Anna from the office.
  4. May 08, 2024
    Summary of first recovered amounts and invoice corrections.

"I thought that at this scale small errors were inevitable and not worth looking for. Lublin Logic Services proved to me that these 'pennies' amount to a specific delivery van on an annual scale."

Marek Wiśniewski Owner, Bud-Max Lublin June 2024