Stage-Gate Process

The Stage-Gate process structures product development into successive phases (stages) separated by decision points (gates). At each gate, a committee reviews the project against defined criteria and decides whether it continues, is reworked, or is stopped — resources flow only into projects that pass the review.

Origin

Robert G. Cooper derived the model from empirical studies of success and failure factors in new products and first described it systematically in Stage-Gate Systems: A New Tool for Managing New Products (Business Horizons, 1990). The mature version, with flexible process variants, is documented in Winning at New Products (4th ed., Basic Books, 2011). Stage-Gate has since become the most widely adopted formal new-product process in manufacturing companies.

Typical use

The process is used wherever many development projects compete for limited resources: in new-product development, portfolio management, and the annual budgeting of R&D. Its strength is discipline — it forces assumptions about market, technology and economics to be made explicit early, and weak projects to be terminated before they become expensive. For ventures with high uncertainty it is increasingly combined with iterative approaches such as Lean Startup.

Procedure

Define the stages
The development path is divided into clearly bounded stages — typically scoping, business case, development, testing and validation, launch. Each stage bundles the activities that reduce the project’s uncertainty most effectively and ends with defined deliverables.
Set the gate criteria
For each gate, must-meet and should-meet criteria are agreed: strategic fit, market attractiveness, technical feasibility, expected margin, evidence of customer demand. The criteria must be fixed before the review — otherwise gates degenerate into negotiations over decisions already taken.
Staff the gates and decide
A committee with authority over resources takes one of four decisions at each gate: go, kill, hold, or recycle. What matters is the willingness to kill — a process in which no project ever dies allocates resources by inertia rather than by evidence.
Adapt and maintain the process
The process is tailored to the risk profile: shortened variants for small projects, additional gates for large ones, regular revision of the criteria. Without this maintenance, Stage-Gate ossifies into bureaucracy that lengthens lead times instead of managing risk.

Limits and typical mistakes

Stage-Gate is built for ventures whose market and technology are essentially known — that is, for incremental and sustaining innovation. Under radical uncertainty its instruments fail: a robust business case cannot be calculated for a market that does not yet exist. The most common mistake in practice is the erosion of gates into mere milestone reviews, where projects are waved through because much has already been invested. Equally widespread is false precision: financial figures with several decimal places suggest a forecasting quality the underlying assumptions can never support. Finally, the sequential logic tempts teams to postpone learning to the end of a stage, although the critical assumptions could often be tested early and cheaply.

Relation to the Innovator’s Dilemma

The Stage-Gate process is the institutionalised form of the mechanism Christensen describes: gate criteria such as market size, margin and proven customer demand are exactly the filters at which disruptive ventures fail — their markets are initially small, their margins low, their customers not yet your own. A well-run Stage-Gate process therefore screens out disruptive projects not despite its quality but because of it. The consequence is not to abolish the process but to limit its reach: disruptive ventures need different gates — criteria such as learning speed and assumption validation — or a separate process with its own resource logic. How both can be anchored organisationally is the subject of the consulting services.