How to Make a Decision in the Presence of Uncertainty

Uncertainty is all around us. In the project world, uncertanty comes in two forms:

1. Aleatory Uncertainty - the naturally occurring variances due to the underlying statistical processes of the project. These can be schedule variances, cost variances, and technical variances - all driven by a stochastic process with a known or unknown statistical distribution. If you don't know what the distribution is, the Triangle Distribution is a good place to start. For example: The statistical processes of testing our code ranges from 2 to 4 days for a full cyber security scan. Planning on a specific duration has to consider this range and provide the needed margin. Aleatory uncertainty is irreducible. Only margin can protect the project from this uncertainty.
2. Epistemic Uncertainty - the probability that something will happen in the future. The something we're interested in is usually unfavorable. For example: The probability that the server capacity we have selected will not meet the demands of the user when we go live. Epistemic uncertainty, being probabilistic, can be addressed with redundancy, extra capacity, experiments, surge capacity and other direction actions to buy down the risk that results from this uncertanty before the risk turns into an issue.

When we hear you can make decisions without estimates, this is physically not possible if you accept the fundamental principle that uncertanty is present on all projects. If there is no uncertanty - no aleatory or epistemic uncertainties - then there will be no probabilistic or statistical processes driving the project's outcomes. If that is the case, then decision have no probabilistic or statistical impact and whatever decision you make with the information you have is Deterministic.

So if you want to learn how and why estimating is needed to make decisions in the presence of uncertainty start here:

• Making Hard Decisions: An Introduction to Decision Analysis, Robert Clemen
• Making Multi-Objective Decisions, Mansooreh Mollaghasemi and Julia Pet-Edwards
• Decision Analysis for the Professional, Pete McNamee and John Celona

And then when you hear about a conjecture that decisions can be made without estimating you'll know better, and consider anyone making that conjecture as uninformed about how probabilistic and stochastic processes in the project world actually work - especially when spending other people's money.