Select Page

A response to the theses of Jörg W. Fischer

Prof. Jörg W. Fischer has dealt intensively with the current development of automation and has formulated very valuable theses in two articles that build on each other, on what effects this development has on the large standard IT systems with which industry controls and organizes its processes.

This is a really big topic. No wonder that – from my point of view – it is not possible to cover all aspects satisfactorily at first go. Nor does Jörg W. Fischer succeed in making a prophecy in this large field that I can endorse without hesitation. I try to summarize his theses. And, at the same time, I develop a critique of the perspective he has sketched out.

After five and a half years as Lead Principal PLM Architect at Siemens Digital Industries Software, Prof. Jörg W. Fischer has been a professor at the Karlsruhe University of Applied Sciences since 2013, where he focuses on production management, digitalization and PLM at the Faculty of Mechanical Engineering & Mechatronics.

Thesis 1:

PLM instead of ERP at the top of the automation pyramid

The first thesis, in my words, is that ERP’s relatively stable leadership position in manufacturing IT for more than 30 years is beginning to crumble.

This could lead to PLM moving to the top of the automation pyramid instead.

According to Fischer, this is not happening because ERP software is no longer doing its job. Rather, for some years now, industry has been driven by its customers to supply increasingly individual products at the price and pace of serial products, known as the “batch size 1” issue.

“Changing the automation pyramid”, Figure 4 in “ERP, PLM MES and CRM – a new perspective – a new universe – Part 1”, Jörg W. Fischer, April 2023.

This is precisely what requires different processes than before. And these processes need different data than the ERP can provide with its data model.

At the same time, the industry, like the entire economy, is fundamentally being shaken to an unprecedented degree. “The recent leaps in technology and global crises have created a potential for multidimensional disruption that has probably never existed in the form it has. This creates the need for companies to question themselves in all areas, to embed agility in their processes, to be able to respond quickly and effectively to changing customer demands, and also to deal with data-based business models.” (Jörg W. Fischer, “ERP, PLM, MES and CRM – a new perspective – a new universe – part 1,” April 2023)

This development has a noticeable impact on the role of ERP, which for decades was under the impression that almost all corporate functions could be controlled via it. Production in particular, as the most important cost block in industry, seemed to run optimally when the bill of materials came from the ERP and could be used for order processing and related transactions. For some years now, this has been increasingly evidently no longer the case.

If it is only during an order that a final decision has to be made as to which standard products are to be supplemented or modified by which components on a customer-specific basis, then it must be possible to access product data very dynamically throughout their entire life cycle. A variety of variants and product families that is difficult to keep track of, as is familiar above all from the automotive industry, is becoming the norm in more and more sectors. And this means that companies must have a 150% bill of material (EBOM) filled directly from engineering, which can become a 100% bill of material for manufacturing (MBOM) in the actual order.

ERP systems are fundamentally overwhelmed by this, because it contradicts their formal, stable data model, which assumes that product data is unchangeable after the end of the design process. PLM systems, on the other hand, which originated in engineering and are familiar with the dynamic growth of product data, have in many cases already been adapted to the management of EBOM and MBOM, according to Fischer.

Fischer’s comprehensible conclusion: In PLM, an EBOM is created that can be transformed into an MBOM in the event of an order. The product development process is based neither on pure series nor on pure order development. Instead, the product is largely configured. From the configuration, manufacturing gets its MBOM, which the ERP system can use to start order processing. Configure to Order (CTO) becomes the common process pattern.

In the previous world and the old automation pyramid, the manufacturing execution system (MES) for production control was below the ERP. In the future, Fischer says, it will be below PLM and in turn supply the ERP with the MBOM. This is indeed a perspective that, to my knowledge, no one has yet formulated in this way.

“Interaction of information creation and use in different process patterns”, Figure 3 in “ERP, PLM, MES and CRM – a new perspective – a new universe – Part 1”, Jörg W. Fischer, April 2023.

Thesis 2:

New layer “Digital Information Architecture” above order processing

With the second thesis, Fischer enters the field of prediction in the second article. Because more and more products are being developed as systems among systems in the course of digitalization, traditional PLM, which comes from mechanics, often reaches its limits.

In the future, says Fischer, “the creation and lifecycle management (xLM) of information is separate from the use of it in the execution systems. The xLM layer provides the necessary, version- and time-correct master data to the execution systems.”

He uses xLM here in much the same way as the term SysLM, which I coined in 2012: as the management of data from all engineering disciplines over its lifecycle.


This management is currently organized via PLM or also via Application Lifecycle Management (ALM) for software or via even other systems, depending on the specialist area.

Fischer speaks of a layer and not a system, because: “As expected, the new xLM layer will not be a classic monolithic IT system. Rather, it can be assumed that different, distributed, microservice-based solution approaches will emerge in the future to implement this layer.” (Jörg W. Fischer, “ERP, PLM, MES and CRM – a look into the future – Part 2,” June 2023)

Fischer introduces yet another term in parallel to include all of a company’s non-structured data related to product and production. He calls the layer of this data the Industrial Data Science Layer (IDSL). It also currently appears as “Digital Core, Data Mesh, Data Lake, Knowledge Graph, etc.” according to his definition. Fischer writes, “Its task is to provide the platform or components and generate entirely new contexts, insights and values from companies’ data treasures.” And finally, “Consistently thought through, a new intelligent control layer for companies emerges.” (Jörg W. Fischer, “ERP, PLM, MES and CRM – a look into the future – Part 2,” June 2023)

Correct analysis, but disputed forecasts

In my opinion, Fischer has correctly recognized that the general trend toward increasingly individualized products – and thus toward the dominance of customer requirements – in the processes leads to an equally general trend of configure to order, which overtaxes most ERP systems and leads to a renewed increase in the role of PLM.

It also seems correct to me to analyze that the historically very mechanics-oriented functionality of PLM is also reaching its limits here. In order to be able to extract the value of all data from product and production, functionalities in the form of microservice-based solutions will have to be added, because the integration of all data in the conventional PLM data models is unrealistic.

At two points, however, the theses seem to me to be misleading.


“The future position of the major system classes”, Figure 3 in “ERP, PLM, MES and CRM – a look into the future – Part 2”, Jörg W. Fischer, June 2023

  1. In the automation pyramid, the ranking of the major IT systems does not simply change. In the smart factory of the future, this pyramid can no longer reflect reality.
  2. It is not important whether the layer for data acquisition and evaluation, which is usually still missing today, is given its own name. It emerges very specifically in each company as a microservice-based solution.

Re point 1:

As early as 2011, when the Industrie 4.0 initiative was launched and the standardization of a reference architecture and ultimately the asset administration shell (AAS) was tackled, it was clear: A central aspect of this digital transformation is the decomposition of the automation pyramid and the decentralization of decisions in industrial processes. And to such an extent that ultimately decisions at the level of individual machines are made more or less automatically by themselves. Today, one could say that cloud technology and standardized containers as microservices have proven to be the toolbox for this over the years.

So the role of PLM and ERP functionality, which is changing right now, is no longer taking place within a hierarchical pyramid, but in this agile, rapidly changing environment of increasingly composable software components.

Incidentally, in addition to this general development, there is another possibility that does not even appear in Fischer’s theses. Bluestar PLM is an interesting example of this. This system is fully embedded in the Microsoft Dynamics ERP. And with it, exactly the CTO process described by Fischer can be mapped, as I was recently able to learn at a customer’s site.

Within Bluestar PLM, you can turn a 150% EBOM into a 100% MBOM, and only then does the data go to the ERP, which is now used to organize the further processing of the order. Here, PLM and CRM appear as functions of the same overall solution. So also this embedded version can solve the named problem.

Re point 2:

Coining a new term for new, intelligent IT layers seems unnecessary to me. If you look at current smart factories, the old-fashioned systems are just as irrelevant as such terminology.

For example, in Rittal’s showcase factory for compact control cabinets in Haiger, data from product creation and from production planning and control are linked with data from sensors in production and even from energy supply. The end result is that everyone can see in near real time on the large dashboards in the production hall which robot line is currently producing how well and where maintenance may be required. I was able to see this in practice recently as well.

There is no talk of a new layer in PLM and ERP there. The focus is on the microservice-based, cloud-native production software ONCITE Digital Production System from sister company German Edge Cloud, which is supplied with all the data it needs. It would be exciting to hear from Rittal and German Edge Cloud, respectively, how they classify this novel solution with regard to ERP, PLM and MES.

My conclusion:

Yes, the individualization of products and numerous other challenges are changing the way industry plans and controls automation in particular. Traditional ERP, MES and PLM are reaching their limits and changing their relationship to each other.

But a new data processing layer that works for all industries is unlikely. It is becoming apparent that in the future, companies will use both old and new software, presumably predominantly microservice-based, with many individual solutions to tackle the countless use cases that are opening up.