These digital services work in practice

Digital business models are increasingly becoming part of corporate strategy. From the perspective of Dr. Thomas Heller, managing director of the Fraunhofer Smart Maintenance Community, the focus must go beyond mere monetization. "Companies should also consider how services and business models can help avoid or reduce waste," says Heller.

"With regard to the necessary data and digital capabilities, pay-per-use is the highest level of development - but in practice, it is still not very widespread," reports Jan Rodig, who leads the Competence Center Strategy & Digital at Struktur Management Partner, a consulting firm specializing in transformation and turnaround in medium-sized businesses.

From his perspective, the most common practice today is to see complementary digital services around the core product, machine, or system. About two-thirds of machine builders already offer digital services as a complement to their product, estimates Rodig. These include OEE benchmarking and optimization, online condition monitoring, and energy management.

Predictive and prescriptive maintenance

The most well-known model is predictive maintenance. However, in practice, implementation is complex depending on the data situation. For Thomas Heller, it is important to clearly distinguish terms that are often confused: Condition monitoring describes the monitoring of systems to detect problems such as increased temperature or changed vibration. Predictive maintenance aims to determine the remaining life of systems and components and draw conclusions for production programs and the ideal time for component replacement.

Going one step further, prescriptive maintenance describes the appropriate measures through which the production operation is then impacted as little as possible. These approaches are about customers achieving fewer disruptions in their production through appropriate investments in technologies and new processes.

It is now clear to most companies that data is the fuel for digital business models and services. Nevertheless, in practice, there is a very large discrepancy between the technological possibilities and their use, according to Heller.

Often, the problem is not the amount of data, but its availability: "The data is in different silos, such as in ERP or MES systems, up to notes that operators have stuck to the machine. The problem lies in making the data available in such a way that it can actually be processed. This is clearly where the greatest effort for AI and predictive maintenance usually lies," says Heller.

Digital twins open up more optimization possibilities

For Thomas Heller, the digital twin is one of the most exciting digital business models for plant manufacturers. "With services based on the digital twin, customers have the opportunity, for example, to see in detail how this plant will function. Ideally, they can simulate how new machines and components fit into their brownfield overall structure," explains the Fraunhofer expert.

This trend will certainly intensify. The data model of the digital twin is particularly suitable for predictive maintenance, as the virtual representation of the machine lays the foundation for simulation, planning, and real-time monitoring. This enables services that contribute to faster commissioning, error prevention, and better training. Faster commissioning, in particular, is increasingly becoming a key differentiator in competition. Many data as a service offerings aimed at optimization can be consistently set up on a DT basis.

New services: Classical AI is complemented by generative AI

Large language models like ChatGPT will change many processes due to their ability to process human language. From Rodig's perspective, AI and generative AI offer the opportunity to develop rapid value-added services, for example, by applying a large language model to user manuals and product documentation to provide quick answers. Visual AI in conjunction with industrial cameras also offers great potential, especially for automating quality tests.

However, behaviors must also change with AI and digital twin-based services. “We often have more of a communication problem than a data problem: Even if, for example, predictive services signal that a component needs to be replaced, production, following the principle of hope, says in case of doubt: We must continue to operate and absolutely fulfill the next customer orders,” reports Heller from practice. If things go wrong, it is up to maintenance to get the system up and running again as quickly as possible.

Remote: Remote maintenance has become established

Remote maintenance is also one of the most widely used services in practice. One of the main advantages is that problems can often be solved faster with less personnel effort. "It is also important that equipment manufacturers can connect with customers' production and not only provide repair assistance but also recommendations for optimal setting parameters during operation," advises Heller. This allows production to be optimized better than the operator could achieve alone. The further the topic of Digital Twin progresses, the better the results will be.

Thomas Heller sees obsolescence as a "huge issue" for plant operators. In an industry of long-lasting products, the parts needed to operate an existing plant are often either no longer available or not available in the required configuration on the market. Of course, it is also a question of sustainability that machines and plants that are actually functioning do not have to be replaced just because individual spare parts can no longer be procured.

Even though there are now a number of companies that focus on reproducing no longer available, especially electronic components, a comprehensive solution is still far off, says the Fraunhofer expert: "The issue of obsolescence will increase significantly because product life cycles have become ever shorter.

Therefore, the demand for remanufactured parts is increasing, as is possible with 3D printing, for example. Spare parts as a service is also a potential business model. The question from Heller's perspective is: 'Have all plant manufacturers already recognized that this is a big issue for their customers and that they want such a spare parts service?'

A deep understanding of the customer situation is important

So-called TCO models (total cost of ownership) or pay-per-use have been in discussion for some time. In this case, it is not the plant itself that is sold, but rather the quantity produced with it. Examples of this can be found in turbine and compressor manufacturing. According to Rodig, the biggest challenge for the supreme discipline 'pay-per-use' for companies is to develop a deep understanding of the customer situation. 'The development of these models is often engineering-driven, with a focus on technical implementation. However, it must be understood what added value specifically means for the customer. This usually requires a longer process with exploratory interviews,' says the digitalization consultant.

Thomas Heller also views the approach skeptically: 'There are many examples where the concept has not worked. This is because it is often very difficult to determine who is responsible for plant downtime. Here, conflict is pre-programmed: does the plant not work as planned, or did the team do something wrong?' states the Fraunhofer expert.

Typically, TCO models are driven by purchasing, where it initially seems lucrative to reduce investment costs. However, the needs of production and maintenance are often not adequately considered, and it often becomes expensive when adjustments need to be made during ongoing operations.

Pay-per-use: The challenges are high

The issue of pay-per-use is particularly problematic where raw materials and base products are not always of identical quality, as is the case in the food industry. Clarifying the question of blame here is almost impossible. According to Rodig, the challenges also include value-based pricing - finding the right price - as well as the issue of financing.

Pre-financing the machines is often difficult, especially for medium-sized companies. However, Rodig sees the necessary cultural change as the main hurdle. "With pay per use, the machine can no longer be the sole pride; a cultural shift is crucial. This paradigm shift is extremely difficult for many companies. The fact that the machine becomes just a 'means to an end' is a hurdle that many in a culture shaped by engineering cannot overcome," states the consultant. Pay-per-use models work particularly well where customers have a financing problem, the consultant knows.

The core problem is monetization

Many companies make a big mistake when it comes to monetization, Rodig knows from 15 years of experience with the topic. Although the corresponding software margins of 25 or 30 are tempting - many managers in the industry thought they could generate significant additional revenue with them. However, the success of monetization depends heavily on the competitive situation and the overall positioning of the company.

The same service can be very monetizable for one company, but not for another. “The willingness to pay for digital offerings is high where they generate high added value: where, for example, machines are operating at full capacity, it is interesting if machine operators can pull together the data and use OEE optimization to get more out of the existing machine inventory,” explains Jan Rodig.

Monetization is usually more successful indirectly: digital services are often an important opportunity to differentiate in tough competition, enhance the core product - and thus defend one's own business model, says Rodig: “In most cases, beyond direct monetization, two levers are at play: first, more machines are sold, and second, the price is higher compared to the list price.”

If, for example, due to pressure from Chinese competitors, only 70 percent of the list price could be achieved, a package of software and hardware could increase this result to around 95 percent, the expert cites an example from his own practice. The resulting possibilities for benchmarking and energy management significantly increase customer value. However, the benefit of a digital offering often remains difficult to measure.