Process Mining for Manufacturing | QPR

QPR Blog

You are here

11 September 2017

Process Mining for Manufacturing

Digitalization provides a great opportunity for manufacturing companies that are able to utilize the benefits before competition. In this blog, I discuss and summarize QPR's learnings from conducting years of process mining projects for manufacturing companies.

Let me start with a story...

Once upon a time, there was a manufacturing company, X Inc, with one factory producing steel pipes. It served customers relatively close to the factory, produced 10 different products and shipped the products from its stock warehouse (Make-to-Stock). Over the time, customers started to ask for different sizes of pipes and eventually X Inc decided to start a new business model where the production of goods only started when customers made an order (Build-to-Order). Customers were happy with the product quality for many years but they were not so satisfied with the long lead times for these Build-to-Order products. Then, a competitor arrived and started to sell more pipes directly from stock and provided shorter lead times for customer orders. Our company X Inc decided to also increase the amount of Make-to-Stock products to keep its customers. However, as the amount of Make-to-Stock products increases, the accurate forecasting of all product variants becomes much more difficult, and eventually the company X Inc ended up having high stock levels on products that nobody was buying anymore, as well as stock-out shortages for hot selling products… Eventually X Inc started to think: "How could we increase the amount of Build-to-Order products in order to make our business manageable and profitable again?".

The majority of our manufacturing customers tell a similar story as above: forecasting for Make-to-Stock products is getting more difficult due rapidly changing customer needs, and on the other hand, the customer’s expectation for a short delivery time is driving down the delivery accuracy of Build-to-Order products. So what is the solution? Good news! Process mining is becoming widely adopted by the most competitive manufacturing industries and the methodology is now ready for all manufacturing companies to support their business. Concrete benefits for manufacturing companies from process mining projects include:

  • Shorter lead times
  • Better delivery accuracy
  • More efficient logistics
  • Less customer complaints
  • Lower capital costs

The great innovation in applying process mining to manufacturing business is to cover as large end-to-end process as possible, with as much detail as possible. Advanced analysis and filtering functionality allows the human analyst to hide distracting details in order to "see the forest for the trees" and later on use all the details effectively to find Root Causes for identified problems.

What activities are typically included in an end-to-end Order-to-Cash process for a manufacturing company for Build-To-Order business model? Here is my top-10 list:

  • Receive customer order
  • Order handling
  • Production planning
  • Delivery planning
  • Order confirmation
  • Purchase materials
  • Actual Production
  • Quality control & Packing
  • Delivery
  • Invoicing

In order to achieve the end-to-end visibility we collect data from all IT systems used during the process including: CRM (Customer Relationship Management), OE (Order Entry), ERP (Enterprise Resource Planning), MES (Manufacturing Execution System), SCM (Supply Chain Management) and Finance systems. Once the transaction-level event data is loaded to the process mining tool, the analysis will immediately point out suspicious long lead times, unwanted process variations, loops, order changes and re-work. Process mining results are very beneficial when the volumes are high and the complexity of an end-to-end process is high. The complexity is often caused by the company having multiple production sites in multiple countries, having a large amount of products served using different fulfillment models (Make-to-Stock, Assemble-to-Order, Build-to-Order, Engineer-to-Order), serving many customers, supporting customer-specific processes for some customers and offering flexibility by having a large amount of order changes.

A high-level process flowchart generated automatically based on the actual event data is typically also compared against the As-Design view to processes. This comparison gives Business Process Management professionals the understanding of how well the actual operations match with the intended operations design. For process harmonization initiatives the ability to benchmark and compare business areas is a powerful tool.

Once the main issues are found, the analysis moves to the second phase which is identifying the root causes. In this step, we utilize all the details extracted from source systems like CRM, OE, ERP, MES, SCM and Fina. The objective is to find those most common repeating problems that have the biggest negative impact to business results. We show these Root Cause candidates in a prioritized list, prepare a concrete action plan for the selected areas, and tackle root causes one-by-one. Since the process mining approach is based on actual data, the results of improvement activities become visible as soon as the actual business operations change - which makes it possible to monitor changes and ensure long lasting results.

And last... what happens when the world changes? Well... having the process mining analysis connected to your IT systems will produce: 

  • A top-10 list of manufacturing process and business problems for an end-to-end Order-to-Cash process
  • A top-10 list of the most critical Root Causes

that empower you to react to sudden big changes as well as to small changes that have a radical impact on your business in the long run.

Welcome to the process mining journey where real data drives process improvements!

Click here to send a comment or schedule a 30 minutes meeting to discuss more with the author,

Teemu

Teemu Lehto
Vice President, Process Mining