One's immediate response when asked what constitutes a decision is usually the choice among a series of alternative actions. While we would not want to argue with this as far as it goes, it clearly does not go far enough. Perhaps we should ask what constitutes a 'good' decision. Well, what does 'good' mean? How is this measured?

This introduces the notion of a criterion. The set of options must be evaluated against a clearly stated criterion so that we can easily see which are 'good' and which not so good when appraised in this way. For example, if our criterion is 'profitability' then this is a relatively easy measure, and with the right data we would be able to rate the various options accordingly. When there is only one criterion, then we can treat it as the decision 'objective' which then becomes to maximize profitability.

So far so good. We have a series of options, we rate them against a stated criterion, and the one that comes out best (i.e. highest in the case of profitability) will be deemed the most attractive. However, most important decisions arising in the corporate world are more complex than this. How does this complexity manifest itself?

Firstly, there may be multiple criteria. Perhaps it is not sufficient to consider profitability only because perhaps the option that is suggested according to this criterion is also the most risky, the most socially irresponsible, or the one likely to bring a bad name to the company. Suddenly things are not so clear.

One response may be for us to apply the same method we used when only one criterion was considered. That is, rate the options against each of the criteria separately. This assumes that we can come up with equally quantifiable measures for the new criteria and this is an important aspect. Even if the measures are not immediately quantifiable, we need nonetheless to be able to rate them on some sort of a preference scale against each of the criteria.

Now it may turn out that option 1 scores best against profitability, but option 2 is most favoured when it comes to risk (ie. appears the safest approach), while option 3 might score highest in the social responsibility stakes. We arrive at a crucial point. How do we resolve this?

The answer is to rate each of the criteria against each other, pretty much the way we rated the options against them. These options must therefore be weighed according to how well they serve the overall objective which now becomes some combination of these criteria. Therefore we now have a set of preferences for the options, one set for each of the stated criteria. We also have a set of preferences describing the relative importance of the criteria with respect to the overall objective. A weighted average of the options preferences is now taken, with the weights being provided by the relative preferences of the criteria. It may not sound like it when described in words, but this is a simple calculation when seen on paper or on the computer screen. It is easily conducted in a spreadsheet or database.

The calculations make up what is widely known as the Analytical Hierarchy Process (known as AHP), a famous tool for analyzing the relative attractiveness of decision options. We will discuss this method in much greater detail in forthcoming articles.