Logistics Process Improvement - Root Cause Analysis by Process Mining | QPR

QPR Blog

You are here

30 January 2019

Logistics Process Improvement - Root Cause Analysis by Process Mining

Do you really know why your products are returned? Did you know that process mining can help you to locate the root causes of reverse logistics and late deliveries?

By digging into the logistics process data with process mining, you get to discover bottlenecks that would otherwise stay hidden. We have listed quite a few questions and challenges on our web page, that have been relevant in our network of logistics professionals. Which ones do you find relatable to your situation? 

Would you be able to tell, what reasons are behind product returns or late deliveries just by looking into your ERP? If you say, “Yes”, I’m inclined to ask, "Are you absolutely sure?". Have a look at the following examples and think if they would apply to your case as well. Also think, what are the products we are working with in this process. I’ll tell you the correct answer at the end of this blog post. 

 

Transparency to as-is end-to-end logistics process 

One of the first things process mining can shed light on in logistics is the complete process you want to investigate. You might expect your process to look like something on the left panel in the picture below – fairly straightforward with clear options of how it should be executed. However, the reality is actually something like on the right. It includes the steps you expected to be there, but also many more. There are also numerous flows that should never happen. Good examples of these are loops where a process step is repeated several times, or backward flows where you basically return to a previous step in a process. 

logistics process flow chart

When asked to draw logistics process, most people visualize something like the flowchart on the left. A straightforward execution of cases with maybe some options or decision points. When the process is visualized based on data, the reality is closer to the flowchart on the right. You would not call that "straightforward", would you?

 

Product returns and reverse logistics

One of the common findings from a logistics process through process mining, is a process step that should not be there. For example, in the picture below Product Return is selected as an unwanted process step. Perfectly understandable, this is a very costly part of a process. You would probably be able to tell from your ERP that there are product returns (there typically is even an order or document type for returns), but would you be able to tell what kind of cases typically end up being returned?

In the picture below, I have analyzed in more detail the cases that go through process step Product Returned. What I can tell about these cases is that typically:

1) product category in these orders is Women, and

2) the product group is Road or BMX. We can also see that

3) cases where product category is Children are rarely returned as well as

4) Fixed-gear products. 

logistics process influence root cause analysis

Influence Analysis easily reveals what are the characteristics related to certain cases. Select a step in the process flowchart on the left and see, what are the typical characteristics related to cases going through this step. In this picture, the attributes defining the selected cases are Category - Women as well as Product - Road where contribution is highlighted in red. On the other hand, we can see that Category - Children and Category - Men are typically not so related to cases going through Product return. These attributes are highlighted in blue. 

 

Sometimes it is easier to see the differences if you focus your attention only on one attribute. Below, you can see the difference first among various categories and then among various product types. 

logistics process influence root cause analysis categorization

Comparison within attribute Category. Product return is more common when the Category is Women than when it is Children or Men. 

 

logistics process influence root cause analysis grouping

Comparison within attribute Product. Product return is more common when the Product is Road or BMX than when it is Fixed-gear. 

 

Dependency between process steps and successful process execution 

There are several angles on how to investigate problematic cases in logistics using process mining. We just looked into the characteristics of orders and cases flowing through a process, but another very relevant angle is also, how the process execution and the process steps actually are dependent on each other. In the following picture, you see that I have selected Late delivery on the left-hand side flowchart to investigate it more thoroughly. On the right-hand side, you can see that red process steps are more commonly related to late deliveries than the steps highlighted in blue. It seems that, whenever products are shipped in Essen and by train, a late delivery is more likely than, for example, when the goods are shipped in Essen by truck. 

logistics process influence root cause analysis benchmarking

Red process steps are more commonly related to late deliveries than the steps highlighted in blue. It seems that, whenever products are shipped in Essen and by train, a late delivery is more likely than for example when the goods are shipped in Essen by truck. 

 

What is the product? 

Did you figure it out? What are the products that flow through this logistics process? 

Bicycles:) 

 

Want to learn more about how to analyze your Logistics process? Check out our on-demand webinar:

Watch webinar

 

Marja Rajamäki's picture
Marja Rajamäki
Business Development Manager