What-if analysis

Another common method of trying to understand uncertainty is what is known as a ‘what-if’ analysis, or scenario analysis. This is slightly different from a sensitivity analysis, although the premise is the same; a scenario analysis seeks to understand the change in our output based on a reasonable scenario of change in all of our variables. Remember that a sensitivity analysis only looks at the change in one variable and is more specific.

Typically the following three scenarios will be constructed:

• Best case (not all variables need to be ‘high’)

• Base case (here we might only change 2 or 3 out of 5 variables)

• Worst case (not all variables need to be low)

We need to ensure that our scenarios are consistent across variables; for example, if we change the price to low, then it makes logical sense to change cost to low as well.

Unfortunately, a what-if analysis is not a foolproof method of dealing with, or trying to understand, uncertainty. Many uncertainties can arise that may be a factor to our output that we are not aware of, furthermore the what-if analysis depends on our inputs (e.g. our best guest estimates on certain variables, or what the worst case might be). This will mean that the what-if analysis is susceptible to human biases, and as such we must be aware of that when we are interpreting the outputs.

Watch the following video on how to undertake a what-if analysis:

Watch: (Optional)Excel what-if analysis: How to use the scenario manager (13:09) [1]

References

1. Excel what-if analysis: How to use the scenario manager [Video]. Envato Tuts+; 2016 May 16. Available from: https://www.youtube.com/watch?v=Cl3Xvv6pDnA