With digital twin it is wise to start small with feasible processes that can have an immediate impact and to tackle more complex processes later on.
Digital twin has recently become a buzzword and major investment opportunity. In fact, data from ABI Research predicts: spending on industrial digital twins is expected to grow from $4.6 billion in 2022 to $33.9 billion by 2030. Major players, including Amazon, Microsoft and Google, have all recently launched their own digital twin solutions.
Despite these developments, adopting a digital twin solution is not always an easy process for businesses. While the “holy grail” would be to have a complete end-to-end digital view of the entire business, this isn’t a practical place to start for many businesses, especially those starting from scratch.
What should companies consider when adopting a digital twin strategy?
Digital twins can model anything from a single machine to an entire company. As defined by IBM“A digital twin is a virtual representation of an object or system spanning its life cycle, updated from real-time data and using simulation, machine learning and reasoning to facilitate decision-making.”
Imagine a digital representation of an engine, then the assembly line that makes the engine, the supply chain that supplies the parts, and even the hiring process and staffing model to ensure the right skilled workers are available when and where they are needed to make the line buzz.
In the long run, an end-to-end system promises incremental improvements in the way a business works. For example, many companies engage in scenario planning. Often this is done on an annual or semi-annual basis, with manual work by analysts to build some selected models in spreadsheets, followed by wargaming around a table or in front of a whiteboard.
Compare that to hundreds or thousands of scenarios that are constantly generated, where many decisions are automated as some become more or less likely, while algorithmic criteria have the most impact for proactive human attention.
The good news is that the entire company does not need to be modeled from day one to realize value from a digital twin strategy. In fact, it is wise to start small with feasible processes that can have an immediate impact and tackle more complex processes later.
The digitization of personnel management is a good example. Every company has processes for recruiting and hiring, selling to customers and – in the case of a service company – allocating talent to projects. Imagine the status quo is that each business unit lead reviews his or her business book and upcoming marketing campaigns on a quarterly basis, then inputs their priorities for the roles and skills HR needs to hire into a spreadsheet.
Compare this to a smart system – a digital twin – that generates information about what is likely to be needed in terms of skills, using data such as past results of similar marketing campaigns, actual time to hire for specific skills, attrition and economic inputs that affect demand. Something as simple as automatically adjusting the employee referral bonus for different roles or changing the queue priority for recruiters without human intervention can provide a competitive advantage.
Whatever part of the business a company lands on as a starting point, they also need to make sure they build a high-fidelity model. The speed at which data must flow into and through digital twins of processes or products can vary: but the bottom line is that life is happening in real time.
Businesses need to take the first step to verify that their infrastructure is capable of handling and processing data in real time. If not, their digital twin journey will likely be short. Having a solid foundation is crucial.
The implications of bad data
What is the worst-case scenario if a company doesn’t follow the advice above and make sure its data is reliable in the real world?
A possible outcome would be that they fall below the model. The intricacies of the data required for accuracy vary depending on what is being modeled. For example, the weather-relevant data is on a quarter-mile scale; data measured by sensors on farm equipment can be down to the quarter inch. These details are important and using data not measured correctly for the given scenario can seriously mess up a model, rendering it ineffective in its purpose as a digital twin.
Companies also need to be aware of entering data that is not important into a model. Without understanding the machine or the business process, it will be difficult to distinguish the important variables from the irrelevant variables. Adding factors that don’t model the real world will hinder the output. For this reason, any digital twin process should start with a thorough audit of the process to understand what is relevant and what is not.
One way to ensure that the digital twin is effective and contains only the most relevant and accurate data is to staff cross-functional teams. The industry responsible for the process or machine being modeled must be in charge and have data science personnel assigned to their team. This will reduce the amount of learning and education required when starting the digital twin process as the engineering team is already embedded and has a deeper understanding of the business process.
Using a digital twin to get ahead
The shift to digital twins is already underway. The vulnerability exposed by the pandemic makes it clear that moving the business continuity process to the digital world can help prevent major supply chain or workforce issues before they spiral out of control.
And – beyond the crises – companies can also reduce waste by creating an automated digital solution for processes they already perform manually. Workers could be freed up to leverage insights from the digital twin instead of spending hours compiling reports that could be out of date almost immediately, thanks to outdated data.
As companies in various sectors start investing serious dollars in digital twin solutions, it is also becoming an important differentiator in the competition. The early adopters get an edge in achieving the end-to-end “holy grail” business model, while those who wait are banned from catching up.
The time has come for digital twins. How companies embrace the technology will have repercussions in the coming years.