The Rise of the Digital Twin
I am a non-fun twin. Despite being born within minutes of each other, my brother and I look, act, and sound nothing like one another. Fraternal twins can't pretend to be their sibling and fool teachers, parents, or friends. And so, we are brothers who just happened to be born on the same day. But identical twins, aside from the hoodwinking benefits, also have the benefit of being an unintentional control group. In other words, how does the introduction of an outside stimulus affect one sibling, and how does that compare to the unaffected twin? (Ex. What do the lungs of a twin who has smoked for 20 years look like compared to the other twin who never touched a cigarette?)
But everyone can, in fact, have a twin. And not just people, but cities, systems, and businesses; duplicates you can test things out on without fear of negative recourse. And ultimately, these carbon copies can create something that is a continual improvement on the original form factor. Welcome to the age of the digital twin.
The term Digital Twin was originally defined as “a virtual representation of what has been manufactured, through various lifecycle phases” - Violeta Damjanovic-Behrendt, Salzburg Research
The idea surrounding digital twins has been around since the early 2000s but came into widescale acceptance around 2010 when it was adopted by NASA to help improve designs for their spacecraft. By throwing as much data into simulations and models as possible, scientists and engineers were able to create near-life-like digital representations of objects and systems that they could then test against. They were cheaper, reusable, and in some ways, more accurate than producing exact tangible copies of objects to perform shakedown tests on. Now, rather than taking a newly-designed car to the Arctic Circle to perform cold weather tests, engineers can run simulations against the car's digital twin to see how it performs, reacts, and fails, in a similar computer-based environment.
My favorite videogame growing up was SimCity. And then SimCity2000. And then SimCity3000. And then SimCity 4. And then, well, SimCity fell on its face, and I moved over to Cities: Skylines. But each iteration became more and more accurate to the inner workings of a major metropolis. Where once you had to solely worry about traffic and electrical grids, you're now faced with getting fishing fleets to salmon breeding grounds while scraping together funds to build a new dormitory at the city-run college. These city simulations became packed full of micro-data wakes that users could essentially, on a macro level, study the path a single citizen takes throughout the day and see how changes in the city's infrastructure affect their daily lives. CityZenith has taken the core idea of these city simulation games and turned it into a viable digital twin for building infrastructure management.
Now, by using CityZenith, builders, developers, and commercial real estate investors can run real-time simulations and stress tests on their properties to see how they function in real-world scenarios. What happens if a hurricane blows into town? What are the local regulations and what effects can these have on the property if they change over time? Will the building be better served with an extra elevator or is that just a wasted expense? All of these forecasts, produced by a digital twin, provide an all-angle data output that a decade ago wasn't possible.
Just as CityZenith has proven digital twins can work in the building management sector, it's time the investment world brought digital twins into the financial mainstream.
At Arcspring, we are constantly seeking ways to not only create operational efficiencies for our portfolio companies but to make it easier for our deal team to make judgments on potential investments and acquisitions based on their financial returns. By leveraging digital twins, Arcspring could - theoretically - create near-perfect forecasts for our companies utilizing the data we uncover during due diligence.
For example, we could better understand the ripples should one layer of a supply chain for a portfolio company break down. We'd accurately know how certain types of personalities would function as CEOs. And, because a digital twin's timeline can stretch into decades, by plugging in a series of data sets, we'd understand how long it would take for a successful exit.
I'd be hard-pressed to believe there aren't software developers out there already building software that can create digital twins of portfolio companies. (As of this writing, I found a few consulting companies that perform digital twin-like modeling for private equity firms, but not a purely software-based digital twin solution.) From a value creation perspective, such a tool would provide insights that would be crucial to increasing EBITDA.
As with anything transformative, there are a few downsides. For instance, no amount of data will be able to forecast a force majeure that causes worldwide disruption. (See: COVID-19) However, a digital twin can predict how a company would react to parts of their business being knocked offline for an extended period of time and how to best mitigate that type of crisis. As with all things digital, there is a very slight chance of IP theft, and having a complete representation of your company in digital form would make a tempting acquisition to unscrupulous competitors. The final risk is information integrity. A digital twin is only as accurate as the data that builds it. An error or lack of oversight can create models that appear near-perfect at first glance, but are rife with errors internally and provide outcomes that are unknowingly erroneous. All of these risks can be properly mitigated through close monitoring, extra security protocols, and periodic audits to maintain the integrity of the model.
Digital twins are already hard at work in the automotive, healthcare, and manufacturing industries. It's time to create a space for them in the financial world, too. Building big data models is at the heart of any successful private equity firm, and having the ability to knock it around and poke holes in it in real-time would provide a level of understanding that we simply haven't had before.
So, who wants to build it?
