Not sure what a Monte Carlo simulation is? Take a look at a great breakdown of what Monte Carlo simulations are.
Monte Carlo simulation methods make it possible to account for uncertainty in the complex and varied decisions you make in your business. I don’t suggest that one man startups with $1,000 a month in revenue start by using Monte Carlo simulations to analyze decisions. It’s a lot of complexity to add all at once and it’s important to really understand the methodology you’re using to make decisions. It’s probably more important that you understand the methods you use than that the method itself is incrementally more accurate.
If you’re familiar with 538 or Nate Silver from the last couple election cycles, you’re probably familiar with models that offer results similar to what you see from a Monte Carlo simulation. Results are in the form of chances a scenario plays out like building this new technology has a 65% chance to be profitable given these assumptions.
There are some big benefits to using this process in my experience. I have used the method for project management, scheduling, revenue projection, traffic projection, and predicting the impact of pricing changes.
Monte Carlo simulation benefits
Durable, reusable models
Once you understand the method and construct a model for a decision, it’s a highly reusable tool. If you take the time to build a decent equation or series of equations for a given decision then your model should stand the test of time. You need not recreate it to analyze future decisions. You will make some minor changes to a series of assumptions that govern the model and see what the impact is.
The durability of these models helps everyone reason about the impact of the inputs and it also lends itself to incremental improvement when you discover that you could model something more accurately or you need to add a variable. Even when you need to make adjustments, doing so tends to be cheap after the initial setup. My experience with high-low models or other discrete estimation sheets is that they are rarely reused.
Impact on team reasoning
Most people don’t reason about the world in a probabilistic way. We may think in terms of binary outcomes, but few people label most events with a likelihood of occurrence much less see all factors in that light.
Most of the time this is fine, but complex decisions and complex situations benefit from clearer thinking. Helping people working in teams see that there are a range of possibilities for completing each milestone and how those uncertainties sum into a project scheduling estimate can really change the way they see the process and even the rest of the world. Probabilistic thinking is a key critical thinking skill in a data driven world.
Like any model you would choose to use, there is an increase in the accuracy of your projections independent of the impact on thinking. Part of that is the shift to thinking in probabilistic terms, but part of it is due to a better method of combining the numbers to find estimates. The common estimation method of taking an average or looking at a high and a low case come with some classic errors in statistical reasoning. Models based on these methods are quicker to create and may aid speed of decision making (a good thing) but they are inaccurate and will provide you with poor decisions in the end. Only you can decide when the tradeoff is worth the cost for speed relative to accuracy, but I would note that once you understand the method and have a model under your belt you won’t find the difference in time to create a model to be very great. The time cost is upfront in building your first model and understanding the method.