How to start managing purchases and increase net profit by 10%

Client

Well known Russian resource exporting company. Annually exports around 200 billion cubic meters of gas.

Client’s Problem

Requirements of the branches and divisions of the company are continuously supplied to the only procurement department in unprocessed form of a simple Excel file which contains basic information. It’s sorting, comparison with suppliers' prices, rates, coordination and purchase required enormous labor costs and any mistake could affect the work of entire structural divisions and could mean financial losses for the company.

Goal

• Develop an optimal pricing and coordination procedure for various types of goods
• Streamline organizational structure of the procurement department
• Automate product categorization
• Simplify the maintenance and updating for the price database
• Automate the search for the best rates in suppliers' prices
• Provide convenient way of working with large amounts of data
• Make the procurement process transparent and provide management with reporting

Solution

Together with the client’s analysts we came up with a new way of procurement data processing, thought out new working process for employees and redistributed responsibilities and it’s areas between them. We agreed on the composition and work plan and began to implement it.

What’s been done

Developed a high load procurement processing center and 6 functional services for:

• processing price lists
• goods categorization
• search for favorable prices for products based on machine learning
• handling non conventional needs
• search for similar product items
• supplier relationship management

Result

Significantly increased speed and efficiency of the procurement department through the optimization of business processes, automation and the introduction of new tools.

• Ensured mass batch processing and various operations with millions of commodity items. For example, exporting 9 million rows to an Excel spreadsheet in just 1 minute.
• Distributing 100,000 products into categories takes from 10 to 30 seconds.
• Time to search for the best prices for one list of goods was reduced from 4 days to just 30 to 60 seconds.