A good analysis of options requires a good set of options to analyse.
Desirable factors (to make a good set) are:
Sufficient options to cover the space of choice. The “traditional” approach of three options – especially if one is obviously not sufficient, one is obviously beyond our means and the middle one is the one the project team wants – is not a worthwhile exercise, as there is effectively only one credible option. All options have to be at least credible.
The Decisive book (referred to under What makes one decision better than another? ) has an extensive coverage of this under the heading “Widening the options”
Not too many options: Typically there are many different choices associated with a project or course of action. If every combination of alternatives were to be considered the range of options can get very large. Investing time analysing such large numbers of options is very rarely effective and a smaller number of representative options suffices.
Distinct options Having several very similar options and one quite different one is not a balanced consideration. Ideally the many different options should feel “evenly spread” across the space of possibilities.
Include the “do nothing”. Arguably this is always a choice – albeit an unpalatable one. In the extreme case that it really is not even a option that could be contemplated, then a “do minimal” option may need to be created.
Options consistent with intent: Sometimes analysis of options is done on the presumption that one of the options considered will be the path chosen. An example here might be appointment of a new Chief Executive from a set of candidates. Other times the examination of options may be in order to gain insight but the actual recommendation that results could well be a combination of the assessed options. An example here could be implementation and migration plans for the software that forms the core platform of a business, where different aspects of multiple options might usefully be combined. In the second case (analysis for insight rather than selection) there are effectively unlimited possible outcomes, so finding an appropriate number of options and ensuring they are usefully distinct is a core judgement.