Run a Monte Carlo simulation of a startup while conceiving it? Has it been done?

As an entrepreneur you try to find a model of your business. If you simulate this via Monte Carlo you could see sensitive areas in e.g. the marketing strategy and how it affects e.g. sales. I like to think of it as a prediction device. My intuition is that a simulation allows for "what if" questions in a way shorter time. A strategy change that materializes to consequences in months can be examined in days or hours. Also in real life there is no "undo" button. My second intuition is that a startup is small enough to allow for gradually building the simulation in parallel to the real business. The validity of the simulation can be easly examined with the real world data. Literally simulate your startup. Has it been done before?

• Stochastic variation is almost never an important concern of a new business. Far more important is the question of whether the basic assumptions about the demand for the product, the intent of the competition, and expertise in producing that product are valid. – Douglas Zare Jun 20 '12 at 10:43
• I think it is also a quantitative problem especially if you burn money and run against time – Roland Kofler Jun 20 '12 at 12:59
• Yes, and it's called Risk Management. ;) – Lucas Reis Jun 20 '12 at 13:13
• Some nice applications of Monte Carlo simulations in business can be seen here: amazon.com/The-Failure-Risk-Management-Broken/dp/0470387955/… . – Lucas Reis Jun 20 '12 at 13:15
• @RolandKofler: I started to think about a way to estimate users in a startup of a mobile app... Even posted a question here: stats.stackexchange.com/questions/27803/… – Lucas Reis Jun 20 '12 at 14:23

Real options valuation(1, 2, 3) can be used to estimate the uncertainty in business decisions such as whether to start a business, or building a mine, or undertake a research project. These options can be valued analytically e.g. Black-Scholes or more realistically with Monte Carlo algorithms. One Monte Carlo approach that I find intuitive is Binomial options pricing, which models a binary decision that occurs with some probability $p$ and can change the value $x$ of the asset (start up, corn futures etc.) by some $dx$. While simplistic, this model has the nice property of being equivalent to Black-Scholes; however, it can be modified to allow for scenarios that cannot be approached analytically--as one would expect in a start up environment.