The art of decision-making using simulation

These are the types of questions any agile business is asking every day. 

However, the conversation is changing. "If we introduce this new service, what happens to our carbon footprint and  sustainability commitments to clients, staff, and shareholders?  Could we predict and quantify that before we spend all of the efforts? Just imagine if that was possible…." 

Now let's jump to a Formula 1 design team meeting. It's a very different type of conversation. An F1 car is in a constant state of development and new ideas are being tabled every day. But before any change is made to the car, the potential change is modelled to ensure the new parts are safe and simulations run to check it would add meaningful performance to the car. Often the effect of the change is loaded into a full-motion simulator so that a human (driver) can experience it. It's not just about the numbers but also what it feels like. The cost of change is calculated and compared to the benefits case of other competing activities. There's no time to build and test these concepts in 'real life' so everything is simulated. In minutes, not weeks. There's no crystal ball but data and simulation are constantly used to evaluate the probable outcome before any change commences.  The good news is you don't need to be a Formula 1 team to think like this, you just need a fresh approach to decision making.

A good simulation is adept at looking at future scenarios, whether that be considering strategic options (months and years ahead) or more operational issues (tomorrow, today, now).  In both cases, making these decisions may involve a complex interplay of factors and useful historical data may not exist, making other decision-making methods less reliable.  

The events of 2020 have seen a drastic change in human behaviours. Data from the past, and the algorithms that rely on it, are no longer representative, presenting a major hurdle for those that develop and use AI.¹

Simulation differentiates itself from AI by including physical modelling of systems to aid understanding of real-world systems - sometimes called 'Digital Twin'. Its birth was in the development and management of physical assets and machines to understand how they might behave under different operating conditions, to predict reliability and schedule maintenance. The next innovation in digital twins is already with us and we're applying simulation to model new operational processes, markets and people. In their own way, these things are more complex and unpredictable than machines and are filled with external effects (a global pandemic anyone?), emotions, new competitors and multiple priorities to balance.

Simulation is for everyone!

To many in the wider business, traditional simulations of manufacturing processes and warehouse logistics are owned by operations or engineering teams. Simulation offers increasing potential to affect the bottom line of organisations through revenue growth and cost saving initiatives but aren’t consistently well understood by CFOs and their teams..  However, decision making is no longer just about the bottom line and this is where we need to broaden the conversation and involve commercial teams, marketing, risk etc. because these teams have additional metrics businesses need to consider.

Increasingly we're being asked to model and predict changes in metrics such as market demand, customer sentiment, cost-to-serve and sustainability impact as well the traditional financial metrics. This would be impossible without a simulation. Some of this we control, we have the data, we can pull levers and act. But what about external factors like customer behaviours, regulation and those annoying competitors? We know they will influence our business but there is more ambiguity around what and when. Again, simulation can help facilitate the right kind of boardroom conversations - informed by data, not gut feeling. 

Changing behaviours

For example, as a result of the pandemic, grocers have seen very different shopping patterns with an accelerated move to on-line and click and collect style buying. This change is still unfolding but will undoubtedly impact the retailer's cost base and will require a reaction. But what to do? More dark stores? Different delivery patterns? A new distribution centre? Partnerships with couriers? 

Are shoppers entirely motivated by price and convenience or would some of them compromise a little bit and embrace a more sustainable service?

Each potential action the grocer could take is like a new widget in an F1 designer's head. It might work but  wouldn't it be better to run the model, look and scenarios and quantify the outcomes before we act?

_____________________________________________________________________________________________

¹ https://www2.deloitte.com/uk/en/blog/experience-analytics/2020/repowering-the-economy-with-artificial-intelligence.html