In this article series we started with the conclusion that there is no workaround solution for master data maintenance. Then we realized that master data maintenance is too complex to be done manually. In the last article we talked about the fact that gut feeling and experience are not sufficient to obtain economically effective master data settings.

So the question arises which possibilities for qualified master data maintenance we still have left. Basically, the solution is obvious. To find economically reasonable master data settings, we need an instrument that enables us to compare the economic effects of alternative master data settings. For many years we are using a self-developed and continuously optimized simulation system for this purpose, which is also available on the market under the name DISKOVER SCO from the company SCT Supply Chain Technologies.

We can store so-called rule sets in the DISKOVER system. Rule sets are cascades of decision tables that define which parameter settings are to be made for which article groups under which conditions. Based on the extensive ocean of data in each ERP system, DISKOVER simulates how these parameter settings would have worked in the past. The rule sets must therefore prove themselves on the real past and for each individual material number and show how well they would have coped with the unpredictable reality; which is why we speak of empirical simulation.

No alt text provided for this image

Sometimes possible master data corrections are tested in the ERP system by making the settings and analyzing the results in the MRP list. This only works for material requirements planning, not e.g. for reorder point control. Furthermore, the results have only limited significance since the ERP system always calculates the future in an idealized way. The only thing we can say for sure when looking at the MRP list is that the future will be different from what the list tells us.

Not every single planning or replenishment decision will be ideal, no matter how you set the parameters, but it is important that the resulting replenishment proposals of the ERP system are on average as profitable as possible.

We evaluate how economical the parameter settings are by means of KPI, such as achieved average stock, achieved statistical readiness to deliver, storage costs, order costs or forecast deviations. The simulation system allows us to compare different rule sets at the level of individual material numbers and to optimize them iteratively. What’s more, we can instruct the simulation system to independently find the most economically suitable alternative from a list of parameter settings.

Crashing the logistics in the computer to find the best possible rule sets

Following the example of crash simulation in car body construction, we speak of crashing the logistics in the computer to find the best possible rule sets. As explained above the rule sets define which parameter settings should be made for which article groups under which conditions. The rule sets must therefore be applied regularly and automatically in daily business, which is also feasible with DISKOVER SCO.

With this engineering approach, we avoid the risk of incorrect master data settings and can therefore implement results quickly.

What do we learn from this?

It is possible to determine economically reasonable rules for parameter settings for the planning and control processes of an ERP system and apply them automatically and continuously. This offers better economic results, reduces the planning and control effort and puts an end to the muddling along.

…and the story goes on…

Since we have mapped the complete dynamic value stream at individual item level with the data from the ERP system (experts also refer to this as building a „digital twin“), we can not only determine the most economical parameter settings possible, but also optimize logistical business models and uncover economic improvement potential in the entire supply chain; as supply chain engineers we simply call this „situation analysis„. This is my cliffhanger for a new series of articles.