DMAIC Project: Accelerating the Operation of the Software Component

Baseline (Define)

The project focused on a software component that was running too slowly at the start. Based on usage data, it was seen that users who used the component with large data sets used the component only once and stopped.
The performance, i.e. the turnaround time, had to be significantly accelerated in order to use the component.

Process performance at the beginning (Measure)

The aim was to get the lead time down to less than six seconds. At baseline, this was not achieved in any of the performance runs. Based on the measured data, it was found that the performance time measured by the machine and experienced by the user are not the same. The user-perceived performance was measured manually and found to be worse than expected.

Action (Analyse & Improve)

The analysis showed that several of the presumed root causes did not cause the problem. The real problem was unearthed using the Design of Experiment (DoE) method, which revealed several aspects of the process that were not previously known and would not have been detected in practice without the use of DoE.

Results (Control)

It was decided to implement some of the components using a different technique. The lead time target for the new component was achieved in more than 60% of the performance runs. Performance improved by more than 70%.
Black Belt's experience with the project:

• A truly enlightening experience to make. It was especially helpful to be able to work on a project that was so familiar to me. I was able to focus on the LSS methods without losing focus on a new substantive issue.
• If I had done this project before the course, I would have skipped both MSAs (Measurement System Analysis). When I did them now, I was shocked. The MSA showed me that the target should be halved. The MSA also showed that the meter could not be used at all.
• The change of the meter made it slightly more difficult to assess the results. I'm still surprised how much you can infer about a process by measuring results without even knowing the process.
• Doing the Design of Experiment (DOE) was an incredible experience. There would be no way to detect aftersort & big data behavior change without DoE.
• Before the project, the 12 steps of the Lean Six Sigma project were happily mixed up in my head. After the project, they have (at least for now) become clearly organized.

This project was a Black Belt training project, which the participant did as part of the training and received a Black Belt certificate at the end of the project.

If you are interested in learning how to improve your processes using Lean and Six Sigma, and earn your Belt Belt certification at the same time, check out the training at the link: Black Belt training.