One of our early discoveries with DecisionPad was its importance as a decision communication tool. The design has always emphasized accessibility, transparency and clarity. Hence the easily understandable tradeoff matrix format with lots of graphical support. Since changes in one place update everywhere immediately, the system lends itself to “what-if” exploration.
Some decision tools behave as “black boxes”; you put in your ratings and data and get an answer back just like magic. These are useful for marketing tools like website “product model suggestions” or “better health tips” but they are not effective for serious decisions. People will not trust an analytical method where the process is hidden and mysterious.
Watching people use DecisionPad, we found that one of the first things they do in examining a model is make a few small changes to see if the results go the direction they expect. They are able to validate the system by direct interaction. Since they can see all the ratings, weights and scores, and since changes take effect instantly, this is a fast credibility builder. It is essential for acceptance by people who did not construct the model.
All of the entries in the DecisionPad matrix are in plain language, no cryptic coding required: specs are numbers or features, prices are dollars, subjective ratings are agreement or excellence. That makes it easy for everyone to understand the evaluation and check for areas of controversy.
However that means the matrix is full of apples and oranges! How to analyze it? By assigning a “utility” or “usefulness” value to each of the entries using a conversion scale you specify. Utility numbers can be in whatever range the audience is comfortable with: -3 to +3, 1 to 10 or 0 to 1000. The scale is a reusable rule to say what utility number is assigned to “excellent” or “$550”. These scaling rules are easy to set up and many common ones are preloaded. They are easily reviewed or graphed – they are as transparent as the matrix itself.
A major aid to clarity in utility analysis is separating the relative importance of criteria (weights) from the utility values across the cells. Unaided decision discussions can get bogged down when these two issues get muddled together verbally. DecisionPad shows a column of weights next to the criteria and provides a graphical way to set and review which is easy to relate to. People can sort the criteria by weight. Agreement on the order of criteria importance is often the first step to forming a solid recommendation.
Graphs are available for a variety of decision elements, which helps speed understanding. One of the most useful ones has been matrix views with small graphs in the cells, implemented in the first DecisionPad. These have grown more elegant as computer displays have improved. These little graphs are now called “sparklines”, a term coined by Edward Tufte who has written extensively on various interesting uses. They pack a lot of understanding into a small space.
Notes can be attached to any item as backup and definition. These become a part of the package and are used to document the thinking.
The net result is that DecisionPad is very approachable. No scary technology visible, in fact it feels natural and intuitive. It requires sophisticated technology inside to make that happen of course, but it does not need to be in your face about it. This makes it useful for meetings, building consensus, creating buy-in to the final decision, and for re-justifying the decision when next year’s budget comes around: the matrix and reports will clearly show why the decision was made in a way everyone can understand.