When using a reference model or framework it is helpful to have a rational basis to determine how useful the framework is (considering that all models are wrong but some are useful).
The diagnostic power of a framework is the degree to which each datapoint fits well into one just one category (fits well into that category and does not fit all other categories).
The predictive power of a framework is the extent to which the implications from being in a particular category are distinct and different from the implications that would flow from being in any other category.
For a framework or reference model to be truly useful, it should have high diagnostic and predictive power. It is not enough that the results be true – there needs to be real insight.
Consider the following 2-by-2 diagnostic and predictive framework. It is clearly nonsense but with a thankfully modest number of exceptions, all conclusions are true and the framework is MECE. But the framework has zero diagnostic power and zero predictive power.
| Framework to determine actions for an I&T[1] Organisation | If CIO has 2 arms | If CIO has 2 legs |
| If CIO has one mouth | Move applications to the cloud | Renovate core applications |
| If CIO has one stomach | Gain informed consent to the I&T Strategy | Outsource routine activity |
[1] I&T – Information and Technology a term embracing Information Systems, Information Technology, Digital etc.