When will it be done - Probabilistic Predictions
Prateek Singh

The very first question a customer asks us when we start work is - When will it be done? Traditional methods of answering this question are fraught with errors. The most common errors include heavy reliance on estimates and use of averages to give one deterministic answer. We are all aware that our world is not deterministic and each prediction has a probability of being right and a complementary probability of being wrong. In this session, we will use examples and a simple exercise to demonstrate a much easier method which can help make probabilistic predictions. These predictions can help teams have more informed conversations with their customers about their probability of completing a project on time and around the risk profiles of their projects. The audience will learn how with very little estimation and simple measurements they can better inform and equip teams, managers and customers with information about possible completion dates of the project. We will show how these techniques are actively being used to predict the completion of single items and a set of multiple items in the real world.

Learning Objectives
Audience will learn how collecting two simple pieces of data can help them answer the customer's most common question - When will it be done?
Session Type
Planning, Metrics, Forecasting, Process