Useful software for decision-making: ChoiceAnalyst
I used a DOS version of ChoiceAnalyst, called Best Choice 3. It turned out to be useful for the project team. For example, we were working to develop a complex software product, and early on I asked the team to put themselves in the mindset that the project had failed, and we brainstormed (on the assumption that the project failed) what the reasons might be. (BTW, it’s much easier to brainstorm what might go wrong if you can dissipate mindless optimism by asking that people assume that the project HAS failed, and we’re figuring out why.)
Once we had a list of the possible failure causes, I wanted to look at the causes in terms of both severity (if it did happen) and likelihood (that it would happen). So those became the two criteria. The decision makers were the team, and I weighted them according to experience—the newbie coder got a lower weight than the experienced database guys.
The program scrambles the choices (reasons for failure) and puts up a number of random pairs for each decision maker—the pairings are not ALL possible pairings, but enough to ensure that if we HAD done all pairs, we’d have come to much the same conclusion.
Each team member then looks at each pair from the point of view of severity (assuming that the things HAD happened) and decide which of each pair would be more severe. This ranks choices on severity. Then they go through another list of pairs (all automatically generated by the program) and decides which of each pair is more likely. This ranks choices on likelihood.
I had the program combine all the rankings, weighting the two criteria (severity and likelihood) equally in this case, and weighting the decision makers’ contributions by the weight assigned to the individual decisionmakers.
Final output is a consolidated list of the causes, ranked by the combination of likelihood and severity, in the view of the entire team.
This description is sort of laborious, but using the program is totally easy and actually sort of fun. We also used it to evaluate new hires based on résumé and interview, with the decision makers being the team, but the hiring manager’s contribution weighted more heavily. I’ve also used it on the personal level to pick out vacation spots (each family member participating) and such. The example on the site is a couple buying a house: the choices are the houses they’ve looked at, the criteria are quality of construction, desirability of neighborhood, schools, closets and storage space, and “wishlist items”. Obviously, the criteria are weighted differently.
BTW, since the program delivers rankings on a scale that allows both ties and large jumps and each ranking is given in the form 17.3 and the like, one trick is to use as choices the criteria you plan to use, and use a single criterion (“importance”) and then have the team make the decisions, comparing criteria pairwise to determine which in each pair is more important. The final result are the numeric rankings, and these rankings can then be used as appropriate weights for the criteria for the next stage of the decision.
UPDATE: The Wife just had the brilliant idea of another preliminary run-through of the program: the choices being the list of decision-makers for the actual decision, the sole criterion being “Worth paying attention to,” and the decision-makers for this run being the team itself. The rankings of the choices—which reflect the degree the team as a whole feels about who on the team is most worth listening to—would then be the weights assigned to each decision-maker—which might make it desirable to have a program setting that hides those weights under password protection.
UPDATE2: It occurs to me that the weights assigned to the decision-makers should be dynamic, not permanent: different weights for different matters under consideration, so that the preliminary run to determine decision-maker weights using “worth listening to” would be done with respect to a particular problem. For example, looking at choices regarding an interface design problem would give different weights to individual team members than looking at choices regarding a database problem.
You can see an interactive video demonstration that is much easier to follow than the description above.
I have not used the Windows version, but I would buy it in a heartbeat if I were still working with a team.
Here’s the scoop, from the first page of their Web site:
We are all faced with making decisions like:
- Which expenses can we eliminate
- Conducting a staff performance review
- Prioritizing your Things to do List
- Choosing mutual funds/stocks to buy
- Deciding which job applicant to hire
- Selecting the right marketing options
- Picking the best job offer to accept
- Deciding which car to purchase
- Analyzing Risk Management options
- Selecting features in a software project
- Deciding which college to attend
Assume you have an important decision to make. It consists of several alternatives with multiple criteria and you must pick the best choice.
Imagine having a software program that organizes all of your alternatives and criteria then reduces the most complex decision down to a series of the simplest decisions you could make, choosing between only two choices while considering one criterion, and then repeating that process for the other alternatives. Once you have stepped through this transparent process for each criteria, the overall best choice is mathematically calculated for you.
ChoiceAnalyst assists your decision making process in many ways. Here are just a few:
- Complex decisions are simplified
- Alternatives with multiple criteria are more easily analyzed
- Multiple decision makers become organized for clarity
- Consensus is more readily attained
- Decision models can be shared by email, WAN or LAN
- The subjectivity in decision making is removed
- Weighting for each criterion reflects its relative importance
- Weighting for each decision maker reflects individual knowledge or experience
- A method for determining Criteria weighting when you aren’t sure
- Reports can be generated to document the decision process
- “What-If” analysis is available for analyzing the decision outcome
- And much more…
Let’s look at just how simply a decision model can be built and analyzed with ChoiceAnalyst…