Process Mapping: Best ROI for Improvement Initiative(Part 2)

Let's start this article with one goal in mind, what should we look for in good process mapping. 

The first step that we need to do in process mapping is to get the real flow of the current process. Although we might spot some improvement pocket immediately, but don't stop there to take care of the problem, you need to keep mapping the process until it reaches the end of the process defined by the bookend. The reason is sometimes there are tasks that need to be done that serve as an input or hook for something downstream. Things that we assume as a wasted task may actually be critical to the related task down the line.


In collecting all of this information, don't sit in a room and think it through yourself or with the team, but actually being out there, asking questions, patience, and taking the time to understand every single detail in the process is crucial to map the process as close to the real life as possible.

When creating the diagram, start with the Level1 view. This is typically a high level collection of the whole process. Usually it is under 8 objects and mostly are all rectangles.

Then the Level2 view will have a more detail and refined task listed. This will have diamonds/gateways, rectangles, start and end point, etc.

Once Level2 information are marked, it may be necessary to create a Level3 view, which is more detail. By the time we are at the level 3, chances are we can spot wastes better.

Below is the list of areas that you want to look for:

  • Automation Opportunities
  • Duplication
  • Redundancy
  • Decisions
  • Delays
  • Rework Loops
  • Non-Value Added Steps vs Value-Added Steps

These represent opportunity for waste elimination. 

Always consider Automation over manual processes. If there are no decision making in the step, it's always better to automate it. Once a step is automated is is harder to maintain, because it is fixed and constant. However with the manual processes handled by people, it is more flexible because human are adaptable. The best practice in the newly created processes is to have the task done manually first, once it reach the consistency desired, then we can start automating that task. If we can't get a consistent result in the manual process, there's going to be a lot of effort to adjust to what we want in the automation which will cost more resource.

In duplication and redundancy, those are usually easy to spot because of the similarities in the task description. Although it might be in a different branch after the gateway/diamond split, but if it can be done prior to the diamonds/gateways, you just create yourself a consistent output.

Some decisions (diamonds/gateways) are unavoidable, but pay attention to some that may not even needs to be there. Sometimes just by moving things around you can eliminate decisions from the process. Rules of thumb are, the less decision you have in the process, the better output quality you will get.

Try to spot some delays. You may mark delays with ("D") shape and color it different so it is easy to spot and investigate each of the D Shape manually. By observing D in the process, you may reduce some bottleneck that slows down the whole process.

Rework Loops. Anything with the arrow pointing backward is worth to be investigated.

Non-Value Added Steps vs Value-Added Steps. Just by the word itself, it describes waste. Why would we want anything that doesn't add value to the process in the first place? There are some steps that performed because of the legacy setup, but once the process stabilizes itself, it doesn't need to be there anymore, and people are doing it just because of tribal habit. By revisiting the process and re-mapping process, we may identify those that aren't relevant anymore.