The use of “sensitivity analysis as a thought process” should be undertaken in the planning and design stages of modelling.

For example, if one needed to build a model to forecast Sales Revenue, this could be done in several ways, each result in a model with different variables and granularity:

Model 1: Volume multiplied by Price.

Model 2: Market size multiplied by market share.

Model 3: Sum of the Sales Revenue per customer (or per product, geographic region,…).

The process can inform the choice of the appropriate model. For example, in the above, Model 2 may be the favoured approach if itis believed that the final analysis used in decision support will require running a sensitivity to market size or share (since these are input variables in Model 2, but not in Model 1 nor Model 3).