Uncertainty and sensitivity analysis in organizations // use of probability in decision making

The administrators of organizations usually have to plan, control and make decisions, based on uncertain calculations of projections of the future. They may know some values, such as prices supplied by their suppliers, about their products and price behavior in the future, but even there, since other price and amount data feel unsafe. Therefore, when adverse events cause harmful effects for companies, their managers require and need the range and possible probability of the results.

One of the ways to deal with this problem is with the use of a sensitivity analysis, which consists of making a single calculation, "best calculation", that is, the most probable for each concept. The optimistic calculation and another pessimistic calculation can also be made for each one. Knowing that the discrepancies will not be symmetric, since the capacity limitations do not allow an organization to exceed the estimated sales volume, while during a recession a very pronounced decrease could occur.

It could also be done, vary the most probable calculation of each concept at the same time, in order to observe how much the overall result can affect. Therefore it is necessary to identify the critical items, since a percentage, however small, would mean a great difference in the overall result.

However, it is not very useful to change each concept in a fixed percentage, since in some of them the percentage could vary much more, while it is not likely that in another it will change to the same extent. It is feasible, and could be, to define an optimistic calculation of income, where only one probability is estimated, such as a 10 surplus. And even then there would only be one chance in 10 of not reaching the purpose in the lousy calculation on income.

However, it is difficult to calculate the interdependence of the concepts, but it is necessary to express them. To combine only the pessimistic calculations of the most delicate concepts would be to give a too gloomy projection.
In summary the sensitivity analysis helps answer the question what if ... ?, but does not give any indication of the likelihood that the "yes" will occur.