

Is it time to find your digital twin?
Although digital twins have existed for a number of years supporting activities such as weather forecasting, use of the technology has accelerated in the past 18 months. By helping businesses to simulate and model their operations in a virtual context, digital twins are ideal for organisations forced to pivot in the wake of disruption triggered by the Coronavirus pandemic.
Common scenarios include manufacturing, autonomous cars, energy and healthcare. Such organisations use digital twins to process massive volumes of data when designing new products and automating processes in factories, plants and hospitals. In the case of healthcare, there are exciting possibilities that even include creating the digital twin of a patient where it becomes possible to anticipate and prevent disease before it threatens the wellbeing of the individual.
What are digital twins and why are they so important?
While they come in many shapes and sizes, most digital twins comprise these three features:
- Enable virtual modelling: digital twins take data from the real world and simulate this environment in a digital setting, frequently using machine learning to optimise the simulation.
- Support decision making: rather than offer a plurality of solutions, digital twins are often used to test a hypothesis or simplify the decision making process by testing scenarios and removing uncertainty.
- Ingest real world data: most digital twins are connected by a ‘thread’ that receives a constant stream of data from the real world. This includes a copy of the physical world’s properties and states, such as shape, position, gesture, status and motion.
First things first: What problems are you trying to solve?
Digital twin technology is increasingly useful for smaller and medium-sized organisations, helping them to survive and then flourish during digital disruption. But to realise the benefits, you need an unbiased view of your project, especially the decisions that you plan to support.
Most decisions fall into three categories:
- Research: Including the critical strategic decisions that make or break a company. Here, the greatest challenge is that data is often lacking, especially when launching a new product, acquiring a competitor, or during an unforeseen ‘black swan’ event. A good example is the Coronavirus pandemic where early decisions about lockdown were made based on relatively small amounts of real world data.
- Development: Having decided to move forward with a new product, where do you build the new plant or assembly line? If you move forward with the merger, how does the acquired organisation merge best with the parent business?
- Operations: In theory this is the stage where human beings leave the stage and the digital twin, supported by machine learning, takes over. A good example would be an autonomous vehicle where we are confident that the computer has a good enough picture of the outside world to take control of the steering wheel. Another example is predictive maintenance that optimises the performance of the manufacturing plant.
Can you explain the results to your audience?
As our examples show, the explainability of your digital twin modelling is also crucial and should take into account all the audiences impacted by the decision making process.
In the case of a pandemic, for example, what is the best way to present outputs to medical experts so that they can determine whether to accelerate vaccinations in different demographics? How about policy-makers hoping to reopen schools, or employers setting a work from home policy? The right blend of information be it a report, spreadsheet or digital dashboard can be the difference between success and failure.
What should you do next, and who can help?
What next? The latest advances in digital twinning enable you to launch a project based on simple modelling and a limited number of parameters. But to get value from any exercise you need to be sure that it is unclouded by bias and internal politics.
This matters enormously because digital twins, by their very nature, shine light on failings in the current system, as well as identifying opportunities for improvement. It therefore requires a certain amount of courage to expose different business segments to this high level of scrutiny.
Objectives and objectivity
Another important question is where does your digital twinning exercise fit alongside other transformation projects? As our examples show, machine learning, analytics, reporting and cloud computing are all important elements of digital twinning. Should you begin with a light-touch standalone project, or should you integrate it with your overall digital strategy to avoid silos and disconnects in the future?
At Upstart we recommend working with an objective third party who will strip bias and politics out of the process and enable all stakeholders to realise the benefits of a digital twinning program. Such an organisation will also have your best interests at heart when asking challenging questions and helping to establish the scope of the exercise.
Having achieved consensus and initiated the project, all parties can then move to the next stage, including planning and execution. But asking the right questions from the outset is critical. Finding the right organisation to pose them should be the next step on your digital twinning journey.
Want to talk?
To find out how your organisation can better collaborate