One of the 4 summary root causes of project performance failures is Unrealistic Cost and Schedule Estimates based on inadequate risk-adjusted growth models.
This assessment comes from research at the Institute for Defense Analyses (www.ida.org), where I sometimes work as a contractor in support of the Office of Secretary of Defense Performance Assessment and Root Cause Analysis.
I'm always puzzled when I hear that estimating is a waste (which they're not), estimates are hard (which they are), estimates are not needed to make decisions in the presence of uncertainty (any decision in the presence of uncertainty requires making estimates), estimates are misused by management as commitments (which they are in bad management domains), and estimates aren't needed to make decisions (simply not true). Those making these claims may have their own personal purposes, their own agenda, their own reasons for sell this idea that estimates are somehow no longer needed when we're spending other people's money to produce value in exchange for that money.
One notion in a governance paradigm is It's Not Your Money
I'd like to provide some background on how proper governance leads to proper estimating and how that's done when those asking for and those providing the estimates have a mutual understanding the reasons for why estimates are needed for any non-trivial project work.
- Managing in the presence of uncertainty is standard business practice. If you have no uncertainty, then estimates provide no value. It is only when there is uncertainty - Aleatory and Epistemic uncertainty - do estimates provide actionable information to the decision makers.
Risk Management is how adults manage projects - Tim Lister. Be an adult, manage other people's money using a risk management process.
- Microeconomics of software development - in the presence of uncertainty - means making decisions with scarce resources for probabilistic outcomes. This is how business works. The Value produced by the expenditure of money (cost) is recorded on the Balance Sheet. We cannot know the Monetary Value of any product or service without knowing the Cost to produce that Value. The naive notion of focusing on Value, while ignoring Cost is just that - naive.
Microeconomics of decision making with Managerial Finance and Behavioral Economics are all part of credible business management processes when spending other people's money. This can only be ignored on de minimus projects.
- Since the future is always emerging, and past performance is probabilistic, a model is needed for what this future may bring, given the past and most importantly, what uncertainties are currently in the future and what uncertainties are likely to emerge in the future. The notion of empirical evidence used to make decisions about the future must address the variance in that empirical evidence. Without knowing and using information about these variances, any forecasts or estimates of the future have low confidence.
Models of future performance come in many shapes and sizes. Monte Carlo Simulation, Method of Moments, Parametric Models are a few used in our domain. Tools are available for each of these. Databases of past performance are available for a wide variety of IT projects. Government and commercial firms record past performance and use Reference Class Forecasting to develop risk-informed estimates for future performance.
- There is a wide range of resources for learning how to estimate, assessing the credibility of an estimating process.
Estimates are not necessarily useful to those spending the money to produce the value. Estimates are critically important to those providing the money to produce the Value in exchange for that money. Saying estimates are waste and not saying for whom they are a waste is a fallacy. Estimating and making decisions with estimates is a Closed Loop control system.
Here are some resources you may find useful
- Analytical Method for Cost and Schedule Analysis
- Analytical Method for Probabilistic Cost and Schedule Analysis - Final Report
- A longer bibliography of Risk Management on Agile Projects
