BUSINESS MODELING

Business Modeling for Decision Makers

Business Modeling for Decision Makers

For decades we have used analytical tools such as probabilistic modeling for dealing with uncertainty. However, outside the financial services industry only a few of companies have embraced them. Instead, executives often make big strategic decisions just from the gut. Some of the reasons for this include the perceived complexity of modeling techniques, lack of transparency, and the “garbage in, garbage out” syndrome. Indeed, numerous examples, including the current credit crisis, suggest decision makers rely too much on the seemingly sophisticated outputs of probabilistic models, while reality behaves quite differently.

Regression analysis is used when we want to predict a continuous dependent variable from a number of independent variables. If the dependent variable is dichotomous, then logistic regression should be used. (If the split between the two levels of the dependent variable is close to 50-50, then both logistic and linear regression will end up giving we similar results.) The independent variables used in regression can be either continuous or dichotomous. Independent variables with more than two levels can also be used in regression analyses, but they first must be converted into variables that have only two levels. This is called dummy coding and will be discussed later. Usually, regression analysis is used with naturally-occurring variables, as opposed to experimentally manipulated variables, although we can use regression with experimentally manipulated variables. One point to keep in mind with regression analysis is that causal relationships among the variables cannot be determined. While the terminology is such that we say that X "predicts" Y, we cannot say that X causes Y.

In our opinion, the problem is not with the tools, but rather in a misunderstanding about their purpose and use. Due to the complex dynamics of real strategic business problems, expectations that the likelihood of all possible outcomes can be accurately estimated (or even imagined) are obviously unrealistic. Claims that a strategic decision is safe or optimal to some highly quantified confidence level (e.g., 95 percent) or that it carries a specific 95 or 99 percent value at risk, are usually inappropriate. Such claims might be meaningful where risk arises solely from the aggregate of a long sequence of micro-risks whose behavior can be predicted adequately from extensive historical data (for instance - and with caveats - risks of default in a truly diversified credit portfolio). But they are rarely meaningful when managers face one-of-kind business decisions involving strategic, execution, or macroeconomic risks.( Pannell, 1997 158)

Even in these cases, however, and perhaps especially in these cases, probabilistic and similar modeling methods can be tremendously useful as a structuring device to organize and combine all available insights about the relevant uncertainties and their impact. Used as an exploratory decision making tool, they improve decision makers' understanding of the key business value drivers, the importance and interdependencies of the most relevant uncertainties, and how sensitive a decision is to the assumptions that have gone into making it. Used correctly, the methods offer essential support to the ...