Sensitivity Analysis basic concept and how to calculate sensitivity analysis are described here.

Sensitivity Analysis

Sensitivity analysis is basically a tool that detects how much an outside variable has an impact on your company. Are your teeth sensitive to cold or heat, that's the kind of question which we ask ourselves when we do sensitivity analysis, not about our teeth but about the assumptions that are used in various business forecasts. So sensitivity analysis is all about challenging and analyzing the effect of changes in assumptions used in forecasts and as we know there are various different places in business where we may need to forecast some information.

  Cash Flow Forecast

The most important forecast in business is the cash flow forecast so we make assumptions about when cash will come into the business? And how much? When will cash go out, in our investment appraisal we clearly make assumptions about what the project cash flows are going to be? And when they arise? How likely is it that they will arise at certain times in the future? What will be the initial investment, could that change? Business forecasts are full of assumptions and sensitivity analysis allows us to challenge those assumptions it asks questions like how reliable are the assumptions made what happens if things turn out significantly differently and also which assumptions are most significant? Which are the ones that we need to forecast? So sensitivity analysis helps us answer these kinds of questions.

Calculating Sensitivity Analysis

In sensitivity analysis, we calculate the risk of only one variable at a time while keeping other variables at normal data. For the variable whose effect we are calculating, we will only take its low or high value while keeping other variables at normal value. We will calculate two NPVs, one will be based on original data which you have predicted and the second NPV will predict that for one variable how much low and high the value can go.   Suppose we have four variables (external factors) as follows:

A

B

C

D

Market Demand

Technology

Exchange rate

Logistic

  

If you are calculating “A” then you will take only high (optimistic) and low (pessimistic) values of A while keeping B, C, and D at normal values. One by one you will have to calculate NPVs of all four variables and you will get to know which variable is the most sensitive one. Let’s suppose the NPVs of these variables are A= 240, B= 120, C= 30, D= 90  

The variable A has the highest NPV and is the most sensitive variable. The variable B has the least NPV and is the least sensitive variable. Our main concern and first priority should be the most sensitive one.

The drawback of Sensitivity Analysis

One of the drawbacks of sensitivity analysis is that you're only as good as the forecasts that you make and the assumptions you make. You will be able to test one assumption at a time or you will calculate the risk of only one variable at a time so is that a drawback? I'm not sure it's too much of a drawback, I think it is good it's a robust/strong, and valid way of challenging a forecast mode in business. You might argue that sensitivity analysis is a complicated concept. I'm not sure that's entirely valid either but nevertheless, there are some potential drawbacks to the use of sensitivity analysis.