#23 How to get a grip on the future with simulations

Who could have seen it coming: COVID, the war between Ukraine and Russia, inflation, and a shortage of many critical resources? Nobody? Or is there a crystal ball out there that can help you prepare your company for different scenarios in time? There is! Just not in the form of a glowing crystal ball with a fortune teller behind it.

In a business context, simulations help you play through scenarios, analyze them, and train the people involved in making decisions. What's behind all of this, how to set up and run a simulation successfully, and what your company stands to gain from it in the future: that's what this episode is about.

SHOW NOTES:

Christian UnderwoodProf. Jürgen Weigandandhttps://www.juergenweigand.comUnderwood GmbHHope Is Not a StrategyWHUSimulations

Pre-order the book Hope Is Not a Strategyhere.

Detailed episode description:

Table of contents:

- Simulations in a strategy context- Business wargaming- Collecting and using data- Intelligent algorithms- Training decision-making with simulations- Building awareness through scenarios- Simulations as preparation for the future- Learning and thinking beyond your own limits- Tips for putting it into practice - Contact- Show notes

Today's episode centers on one question: can simulations help you make better decisions?

The short answer is yes, they can. What matters most, though, is how you define a simulation and how you then put it to use.

So let's start by clarifying what simulations mean in a strategy context.

Simulations in a strategy context

Simulations originally came out of the military, as an application of strategy. Movies often show the famous sandbox exercises: older gentlemen standing around a sandbox, moving troops, simulating attacks and defenses. That's the basic idea behind a simulation. You play through various predefined scenarios to see how they develop through interaction. The technical term for this is wargaming, or business wargaming in a corporate context.

Business wargaming

Business wargaming is first and foremost a role play in which you work through different scenarios from the perspectives of different parties. The company, its customers, its competitors, and members of the organization each take on a role. Outsiders are often invited in as well. The goal is to test the conclusions you reached earlier through logic and reason. The role play also brings in emotion, because every player is trying to win from the position they currently hold. That makes it a good way to test what goes beyond ordinary rationality and to see which options and outcomes might emerge.

Collecting and using data

You need data as the basis for examining different scenarios. The key question is whether you want to analyze something at the highest level, in very general terms, or whether you also want to derive quantitative insights. The more quantitative the analysis, the better and the more numerous the data needs to be. For the company, that means it needs data from inside the organization. And in scenarios where competitors or market conditions play a role, you also need data on them. If, on the other hand, you only want to look at scenarios at a high level, a general question is often enough, for example when an existing product is to be repositioned against competitors. In that case you can simply ask how customers will react if the company shifts its market position slightly up or down.

Aldi and Lidl are a good example. They're currently pushing various parts of their stores upmarket, toward Rewe and Edeka. A wargame would be a good way to test how customers react at certain points and which obstacles might come up. You can also see how competitors move in different scenarios and then derive the best decision from that. All the instruments you already have can be used here. If you want to generate data on customer behavior, for instance, you can run traditional market research with surveys, test markets, and so on. Those results then feed into the wargame so you can define the scenarios as precisely as possible. At this point it may sound like a lot of complicated programming work. It isn't, as long as you know where the data sits in your company and how it can be used where it's needed. In other words, you have to figure out where the interfaces are. But cost data, at least, is usually available in every company. Otherwise something is fundamentally wrong. On the market demand side, you can run market research. Often, though, smart people inside your own company, for example those running sales and therefore close to the customer, can supply information too. They can answer questions that may be quantitative and are certainly qualitative.

Intelligent algorithms

In the past you'd have said the results are only ever as good as what you programmed. By now, though, we have self-learning algorithms. That gives you far more options, and if these algorithms are formulated well, they can generate scenarios on their own. They may produce solutions that are completely different from what you had in mind at the start. Google, for example, managed to beat the world's best Go player with its simulation. Of course you have to consider how far that really transfers to an ordinary company. But a lot is possible in that direction, whether programmed into apps, driven by artificial intelligence, or kept purely at a qualitative level.

In practice, this is mainly used today by companies that make heavy use of big data. At Amazon, for instance, the entire background of pricing is simulated automatically. If you want to use big data, you always have to automate as well. No single person can take in that much variety and spot patterns at a glance. Automated processes fed with intelligent algorithms absolutely can. Some caution is warranted here, though, because for strategy development you need to know whether you're looking at symptoms or at consequences that have causes. To change something, you first have to establish a cause-and-effect relationship. With big data, as Amazon and others use it, correlation is what counts. That's an excellent concept. It means that when a customer wants to buy a particular product on Amazon, they're immediately shown similar products that other customers with similar interests bought as well. That's correlation analysis. And it works exceptionally well. But if you want to develop and execute a strategy for your company, you have to be clear about what is cause and what is effect.

Training decision-making with simulations

Now back to running business wargames and how simulations can help you make better decisions. Simulation also works extremely well in teaching. You only get better at things by applying and practicing them yourself. The same goes for decision-making. When you train decision-making, you also sharpen your awareness of where you might be making your own mistakes. This is where topics like cognitive bias and bounded rationality come in. All of it can be practiced in a simulation, in a protected setting. Think of Lufthansa's flight simulator. Pilots train for every conceivable situation there, even a crash, without putting themselves in any danger at all.

The insights that simulations bring to light in companies are often scenarios people already know. Understanding how interdependent the different elements of a company are, for example. Everyone knows that in principle, but actually feeling it in a business wargame, noticing that connections were recognized too late and that your own targets may have been missed, sharpens the awareness. What matters about simulations is that you can make even the dumbest decisions in a protected setting and review them afterwards. That creates the chance to test the "what if." That in particular can really move you forward. When decisions produce good results, people tend to be satisfied, but the crucial question should be whether even better results would have been possible. And whether the results only turned out well because others made decisions that happened to benefit your own company, which means the success can't be traced back to your own strategy or tactical moves at all. This is where a simulation lets you see connections clearly and practice taking in as many things as possible at the same time.

Strengthening awareness through scenarios

Experience shows that teams in companies often work across disciplines and on their own silo topics. Everyone thinks for themselves, frequently only for their own organizational unit. But in that context, when a simulation requires concrete decisions under time pressure, an exciting realization often emerges: close collaboration and communication with other colleagues and departments plays an enormously important role. That factor should also come up in simulations and be practiced, because in reality there is time pressure. You can't think about decisions indefinitely and keep fine-tuning them. At some point decisions have to be made, and they have consequences. If they turn out positive as expected, wonderful. But often the consequences are negative, and then you have to find the causes. Some executives think they can make decisions intuitively based on their extensive experience, but there is a risk of extrapolating something from the current situation into the future that won't necessarily play out that way at all. And it's always helpful to look beyond your own convictions, check the decision once more and play through a scenario for the future that may look very different from what you expect.

Simulations as preparation for the future

In these volatile times, the time horizons keep getting shorter. Nobody can foresee the changes, disruptions and factors like the Corona crisis or the war in Ukraine. That makes it all the more important to be prepared for things you don't necessarily expect. Thinking ahead and writing down and sketching out the various scenarios can really help here. That's also how simulations came about in a business context. Shell was the first company to introduce them, back in the 1960s. An entire unit specialized purely in thinking ahead and exploring possible scenarios. At the end of the 1960s there was one scenario, the so-called oil price shock, which then actually occurred in the early 1970s. So mentally Shell was already prepared for it. That's exactly what using simulations means. Whether they are qualitative or quantitative, driven by artificial intelligence or not. You try to make educated guesses. The future is uncertain, it's always a matter of guessing how things will really turn out, but that guessing can be improved and, to a certain degree, tested.

Learning and thinking beyond your own boundariesYou can learn how to react in unexpected situations. That learning never stops. To come back briefly to the book "Hope Is Not a Strategy," where the topic is taken a bit further: moving on from the simulation to learning along the workflow, at the moment it's actually applied, getting exactly the piece of knowledge you need to solve the problem that has come up at that point, and simulating in very small increments too. This does require a certain amount of thinking and possibly resources as well. Strategy simulations don't always have to be big ones. Plenty of simulations are also used in logistics and operations, where the people in charge want to explore the consequences of, say, reconfiguring or replacing a machine. What's crucial, as mentioned, is to step out of silo thinking and think beyond your own boundaries. That's what moves a company forward. And you should see the connection here: if simulations are used in different forms and in different places across the company, then there should also be someone who thinks about the bigger picture and can spell out the implications and repercussions for other areas.

Tips for putting it into practice Finally, here is some advice on how to take the first steps toward successful simulations so your company can benefit too.The simplest way to simulate something on the basis of existing data is Excel: you load in the data and experiment, change parameters and analyze how the results develop. You can also look at supply and demand based on data. The simplest form is to observe what effects occur when you change the data. Since Excel is the only tool involved, no investment is needed.If you want to go further and examine a specific problem, there are problem solvers. For example the software Jürgen Weigand developed himself, which can be used for this. You just have to be clear that a simulation or software like this only works well if the data you enter is really the data you need. So if you want to make pricing decisions, you need data on the cost side and on the market, plus a price expectation. Then you can test scenarios and make simulations as complex as you like. But the rule here is: the more complex, the more varied the results can be. Hence Jürgen's recommendation at this point: keep it short and simple. What comes out is a forecast, a recommendation, which can then be reviewed internally once more in a business wargame. The truth is out on the field.ContactIf you're interested in running your own simulation with your team, Jürgen Weigand offers various simulation software solutions. You'll find all the key information in the show notes. You're also welcome to get in touch with us, and we'll help you make better decisions with simulations.

SHOW NOTES:

Christian UnderwoodProf. Jürgen Weigandandhttps://www.juergenweigand.comUnderwood GmbHHope Is Not a StrategyWHUSimulations

Preorder the book Hope Is Not a Strategy here.