About two years ago, one Borbakis data scientist finished an extensive analysis for an international machine production company (client). Now, years after the analysis was created and handed over the results, are markable.
65% yearly revenue increase – two years after the analysis was completed.
The case was simple but with a considerable impact. The client wanted answers to the following questions:
“How can we increase our sales to customers who have already brought a production machine, and how can we improve our customer lifetime value in a timeline of 5 years (First year of buying the product machine + 4 years after)?”
Industry and available data
The client is designing and building production machines for different production companies and has over +1000 product machines in their portfolio. Besides producing the product machines, the company offers service agreements for the machines and machine reparations.
Because of Contractual rules, the client company cannot be mentioned in this article.
The analysis was based exclusively on ten years of Point-of-sales data (POS-data). The POS data contained the following information:
- Individual product purchased and product amount
- Product price and total sales price
- Time and date for purchase
- Customer information: Name, company, address, mail
The analysis had an impact on different areas of the companies’ departments. When you want to apply data science analysis to your company, it is essential to remember: The results will affect your business areas; therefore, you must hold a good dialog with all leaders and employees. Ensure that they understand how they implement the results and understand why the new working way creates value for the business and the individual department.
The analysis result is not creating value itself – it needs to be implemented by employees and leaders.
The results from the analysis will not change your employee’s work. It will change the way they prioritize which product to sell to the right customer. So they can do their work easier and with better results.
In this case, it was the client sales department, and the business executers there had to implement the analysis result in their work.
The analysis defined what and when a specific customer needs specific spare parts for the product machine.
The insight gave the sales department a new approach to additional sales to already existing clients.
Understanding when a particular client needs a specific spare part allows the sales department to contact the customer and offer the product right before they need it or buy it somewhere else.
Understanding the customers’ buying behavior also helps the sales department spot customers who didn’t follow the standard approaches. By reaching out to the customers there didn’t adhere to the typical buying process, the sales team started identifying product areas there needed improvements. By talking to customers there didn’t follow the regular buying approaches, the company got a new market understanding and the knowledge to improve the areas where their competitors were doing better.
Not only did the insight increased sales to already existing customers. It also created a closer customer relation, which increased the average customer lifetime value and a more elastic price when the sales teams offered spare parts.
Knowing what kind of machines or machine parts there is most likely to get worn, the analysis gave the business executives the insight to more easily set the right price for service agreements and reparation. Before the analysis, the client didn’t know if they were making money or losing money on their service arrangements.
By implementing the analysis results when setting the services price, the business executers made individual prices on the service packets. The new prices created value not only for the client but also for the customers.
Because the customer no longer needed to pay for a service packet containing services, the machine would never need. The client was sure that the service packeted they sold was generating actual revenue.
The result from the analysis also optimized other areas as:
- Inventory management
- Schedule planning for service technicians
- Marketing department
- Economy department
How a pattern algortihm works
The pattern analysis was created using a machine-learning algorithm in Python and ten years of historical POS sales data.
By comparing customers’ purchase history and lifetime value, the algorithm identified patterns in the customers buying behavior, which could be translated to predictions, with a surgent amount of probability, of how new customers would behave after purchasing a production machine in the same category.
The analysis takes historical data and finds the patterns, example:
After XX% time, customers who have brought product machine A will purchase spare part B.
The chances that ‘look-a-like’ customers will follow the same patterns in the future are big. By identifying and understanding the patterns the customer leaves, is it possible to predict how your new customers will interact in the future.
Are you curious about knowing how Borbaki can help your company? Do not hesitate to contact us, and let us take a casual talk about your company’s opportunities.