![[NASA Logo]](images/meatball.gif)
Why is there variability in the transaction price of a given commodity? For example, all oranges are not the same. Some look better than others. Some are sweeter than others. If one's purpose is to [purchase,orange], then the orange has qualities [purchase,orange,number_of_blemishes], [purchase,orange,sweetness], [purchase,orange,acidity], and [purchase,orange,price]. Whether an orange purchaser will purchase a lot of oranges or not can be treated as a probability distribution over these qualities. Hence, the qualities of a commodity affect the parameters of the choice distribution, such as mean and variance.
In design for competitive advantage, value is defined as the measure of choice. Thus, the probability distribution just described is a measure of value. It is defined by quality (the qualitites) and price, which is itself a qualitie. A slightly different perspective leads to an economic demand function. To accomplish this, the distribution provides the probability that a given number of oranges will be purchased over a set of qualitie points {([purchase,orange,number_of_blemishes], [purchase,orange,sweetness], [purchase,orange,acidity], [purchase,orange,price])}.
Note that price is a variable within the value distribution. In fact, the marginal distribution on price of the value distribution provides the probability that a commodity will be purchased within a given price range. But price is a money currency scale. Suppose we now imbed all of the money currency scales within the the value distribution. Integrating over all nonmonetary qualities leaves us with the money distribution for a given commodity.
Given this model, one can conclude that money is an information marker modeled by a probability distribution over a qualitie subspace affecting, and affected by, the value of a purpose.
To Index
How can we measure value? Given the above model, we must first identify the qualities which affect individual choice. Next we must understand and model their effect on individual choice as these qualities vary.
Conjoint analysis is such a quantitative measurement process.
The combined effect, through simulation, of many individual choices approximates the value distribution for a market characterized by the individuals and the qualities used to define the individual choice.
Note that individuals may make different choices with the same level of the qualities because the chosen qualities did not include all causal qualities. The result is noise, and hence a probability distribution, even for an individual.
Note, further, that we should consider robust design for market value using robust design techniques (Unal and Dean, 1991b) similar to those of Taguchi methods.
To Index