# Sensitivity Analysis: How the right idea can get spoiled by the problems of the implementation

Concerning the basic underlying idea behind “Sensitivity Analysis”, I cannot find a single negative thing to say about it. In order to understand the problem that exists in the implementation, consider the following line of thought:

Forecasts     +     Calculation Method     =     Result

±A%                                 ±B%                           ±(A+B)%

We begin with the creation of some forecasts, which we process thru some Calculation Method, in order to arrive at a result. Now, when we say that forecasted sales for May 2011 are going to be 10 mil USD, we know that it is highly unlikely that this is going to be the absolutely accurate actual number. Some level of inaccuracy is incorporated into that figure, depending on the quality of the forecaster’s work. Generally speaking, let’s say that it is ±A% inaccurate. If the calculation method we use has built-in inaccuracies, like we have already seen in the “Net Present Value” method (let’s say that it is ±B% inaccurate), the end result is going to be ±(A+B)% inaccurate. In other words, one level of inaccuracy gets piled up on top of the other, to create an even bigger combined inaccuracy level.

If anyone hopes that A and B in some scenarios might cancel each other out (for example A= +8.00% and B= -7.50%), then please note that:

After the Financial Analysis work has finished, and you have established a result, any experienced professional will tell you that this is just the beginning of the real and the most meaningful part of the work, which is the “What if” scenario storm. What will the profitability look like:

• if we make only 92% of the sales target
• if the days of credit to the customers are 19 more than we expected
• if the Interest rates do this