Stage Gates and Stopping Rules by the Numbers

About this book

Development projects fail in two opposite ways. Some commit too early: tooling is ordered, launch dates are announced and suppliers are booked before the business knows whether the product works or whether anyone will buy it, so that stopping later costs far more than it needed to. Others continue too long: each review finds a reason to spend a little more, money already spent becomes the argument for spending the rest, and a project that should have been stopped or changed runs on to launch. Both failures come from the same place. Commitments get out of step with what is actually known.

This book treats that problem with numbers. It starts with premature commitment and the difference between spending money and committing it, then builds a decision tree for a staged programme and puts a dollar value on the right to stop at a gate. It goes on to the value of information, perfect and sample, which says how much a test or a trial is worth before it is run, and to waiting as a simple real option. The second half turns to the forces that keep projects going: escalation of commitment, the sunk cost effect, and the stopping rules and kill criteria that work only if they are written before the evidence arrives. It ends with a choice between continuing, pivoting and stopping in favour of a competing investment, and with a review that judges the decisions separately from how they turned out. Each chapter carries one running case, includes a plausible error and how it was found, and ends with an exercise and its worked answer.

It is for owners, general and engineering managers, product and project managers and finance staff in small and medium Australian businesses. It assumes spreadsheet arithmetic and simple probability: expected values, conditional probabilities and Bayes’ rule are explained where they are first used. Stage gates that make real decisions describes how a gate process is set up and run; this book does the arithmetic that tells a gate what to decide.

The running case

Every chapter follows one fictional programme. This is an illustration. Corella Plains Implements is a family-owned maker of trailed boomsprays and chaser bins in the Wimmera region of western Victoria, with about 70 staff. In February 2025 it starts developing the SB36, a 36 m trailed boomspray with cameras that detect weeds in fallow paddocks and spray only the weeds, rather than the whole paddock. The programme runs through five stages separated by gates: scoping (A$40,000), concept and a bench detection trial (A$180,000), development of two prototype booms (A$560,000), a summer field-validation season on growers’ farms (A$320,000) and a launch costing A$1.8 million, all excluding GST. The development stages total A$1.1 million and the whole programme to launch A$2.9 million.

The programme meets most of the problems the book describes. A tooling order is placed too early. The development stage runs over through a series of small top-ups. The field season returns a weak result, and at the launch gate the team must choose between launching anyway, waiting a season, pivoting to a cheaper retrofit kit for growers’ existing booms, or stopping, while a robotic welding cell for the chaser-bin line competes for the same capital.

The business, its people, its products, its records and every number in the case are invented. None describes KEVOS® work, a real business or a real product.

How to use the book

Read Chapters 1 and 2 first: the distinction between spending and committing, and the decision tree built at the development gate, are used in every later chapter. Chapters 3 to 5 value flexibility (the right to stop, information and waiting) and are best read in order. Chapters 6 to 8 deal with the behaviour that keeps projects going and the rules that counter it, and Chapters 9 and 10 cover the launch-gate decision and the review that follows. Every calculation in the book was checked by running it, and the figures are rounded only for display.

The book is one of a series of Library books on delivering projects, written to be read together. Reference Class Forecasting by the Numbers supplies the outside-view estimates of cost and time that the trees in this book need as inputs. Front-End Planning by the Numbers covers the work done before the first gate that decides what the options are. Demand and Benefit Forecasts by the Numbers deals with the market estimates behind the payoffs at each branch, and Prototypes and Simulation by the Numbers with designing the tests whose value this book calculates.

Source limits and professional responsibility

The methods are established and the sources are listed with what each was used for. Where the book quotes published research findings, they are attributed and describe the studies that produced them, mostly laboratory experiments and studies of larger organisations; how far they apply to a small business is discussed where they are used. The case’s probabilities, payoffs and costs are teaching values, not market data, and a real appraisal needs its own estimates, checked by the people who will be accountable for them. The figures are before tax and ignore any R&D tax offset; accounting and tax treatment needs advice from the business’s accountant. This book is general information about project and investment decisions. It is not financial, accounting, tax or investment advice.