Which software would you buy: @
Budgeting and Forecasting with Monte Carlo simulations
Answers
This is an interesting site for your question: http://www.vertex42.com/ExcelArticles/mc/. Looks a bit dated, but someone went to the trouble a few years back to build it. Actually, this looks like a pretty interesting
I used @Risk for a few years and it was simple and clear and plugged right into Excel.
I'm curious, if you want to share, about why you are using Monte Carlo simulation. I used it in my early days in finance a couple of times but gave up when I realized that although outcomes may be random, the best way to prepare didn't seem to be by running simulations of random factors through a model, but rather to take best guesses for a range of outcomes that you can then convert into working models, only a few of which you can reasonably process at any time. But I'm open to the idea that I may be missing something here and would love to hear why others use randomized modeling.
3
Our business is highly difficult to predict as there are parts of the business that are procyclical, other parts that are countercyclical, and yet others that appear to not have any predictability. I'm building a simple annual budget and multi-year forecast for our company, but given the highly uncertain nature of the top line components, and a long and deep list of projects in the coming years, I wanted to make everyone more cognizant of the business' different risks, and the potential magnitude of those risks. By illustrating and depicting the various risks in a graphic and easy to digest manner, I hope to incite some thoughtful discussion and planning for the coming years. If I'm successful, it'll be worth the $1500 and 20 hours I'll spend on modifying our static budget into a simulation model.
Michael, look at this blog posting, http://whatifyourstrategy.com/2010/05/14/the-how-likely-case/.
Chussil's got a number of interesting blog postings relating to uncertainty in forecasting.
Hi Michael, I did corporate FP&A for about 25 years and found Monte Carlo simulation very useful (we used an earlier version of Crystal Ball). Let me make a few points which I hope you'll find pertinent.
Simulations can be programmed with many types of variable including those with a limited range of outcomes (50/50 probability of two different outcomes as a simple example) - simulation can still be very useful if there are a large number of variable to help understand the range and distribution of possible outcomes.
Forecasting tends to be very optimistic especially with respect to sales and profit margins (as opposed to operating expenses) and my observation is that this is because most variability comes from things that are not controlled by the forecaster (i.e, a volcano in Iceland grounds planes), so introducing simulation can be helpful in understanding that +/- 5% may not be a realistic best case/worst case range even if it seems very reasonable on the surface.
That is the tip of the iceberg so to speak, and there are certainly trade-offs involved because as I'm sure you know, people really just want to know the answer not why it is hard to forecast accurately!!
Filed Under:
Cash Management