Draft:Stage-Gate Process

• Comment: In accordance with the Wikimedia Foundation's Terms of Use, I disclose that I have been paid by my employer for my contributions to this article. MessyCreative (talk) 14:07, 19 November 2025 (UTC)

---

Stage-Gate Process is a widely employed product development process that divides the effort into distinct time-sequenced stages separated by management decision gates. Multifunctional teams must successfully complete a prescribed set of related cross-functional tasks in each stage prior to obtaining management approval to proceed to the next stage of product development. The framework of the Stage-Gate™ process includes workflow and decision-flow paths and defines the supporting systems and practices necessary to ensure the process’s ongoing smooth operation,” according to the 2023 PDMA Handbook.

Overview

Further information: New product development

The Stage-Gate system breaks the product innovation process into a predetermined set of stages, as in the figure below. Each stage defines prescribed, cross-functional, and parallel tasks to be undertaken by the project team, incorporating established best practices and critical success factors. The entrance to each stage are gates, which are go/kill and investment decision points. This stage-and-gate format led to the “Stage-Gate system” designation, which first appeared in academic literature in 1988.

Stage-Gate has since become a popular process, especially for physical-product firms. “The 2010 APQC benchmarking study reveals that 88% of U.S. firms employ a Stage-Gate system to manage new products from idea to launch. In return, the companies that adopt this system are reported to receive benefits such as improved teamwork, improved success rates, earlier detection of failure, a better launch, and even shorter cycle times—reduced by about 30%. These findings highlight the importance of the Stage-Gate model in the area of new product development.”

History

The term ”stage-gate” first appeared in literature in 1988, when Robert G. Cooper introduced it in an article published in the Journal of Marketing Management.^[] Cooper was a pioneer in research in new product management, and had conducted significant research in the area of success drivers of NPD since the 1970s. The Stage-Gate model was based on his study of highly successful NPD project teams and also on his research into successful versus unsuccessful product development initiatives. The original Stage Gate model was introduced in Canada in the early 1980s, where Cooper was a professor at McGill University, and then adopted by a few leading Canadian and US firms, after which its use spread rapidly.

The ISBM (Institute for the Study of Business Markets, Pennsylvania State University) was instrumental to the adoption of Stage-Gate in the USA by introducing it to its member firms. Dr. Ralph Oliva from DuPont was ISBM’s first Managing Director and supported this work; Cooper was made an ISBM Distinguished Research Fellow. Member firms and firms located near ISBM were among the first to adopt Stage-Gate in the late 1980s and 90s: DuPont, PPG (Pittsburgh Paint & Glass), Rohm and Haas (now part of Dow Chemical), ExxonMobil Chemical (then in NJ), Hercules Chemicals, and Procter & Gamble.

Initial performance results were promising, and thus the model gained popularity, particularly in the United States. The 2010 APQC benchmarking study revealed that 88% of U.S. businesses employed a Stage-Gate system to manage new products, from idea to launch. In return, the companies that adopted this system were reported to receive benefits such as improved teamwork, improved success rates, earlier detection of failure, a better launch, and even shorter cycle times, reduced by about 30%. A more recent global PDMA study revealed that about 54% of firms globally use Stage-Gate for NPD. This study concluded that “…faster cycle times result from the use of a formal, structured NPD process such as functional phase review, cross-functional stage-gate, or agile,” with Stage-Gate being the more popular.

Predecessors to Stage-Gate: The Waterfall Process originated much earlier (1940s-1950s) for large-scale engineering projects, later adopted for software development. Like water flowing down a series of rock ledges in a waterfall, the process moves through distinct phases in a level-by-level approach.

Waterfall is a linear and sequential project management approach, and tends to be fairly rigid, much like Stage-Gate was in its early days: Each phase must be completed before moving to the next, with little overlap or iteration. Additionally, Waterfall is a technical process, not a business process, and does not have explicit decision points or gates between phases.

The Phase Review Process: This early process is very similar to Waterfall and was developed by NASA for engineering projects. Unlike Stage-Gate, it is not cross-functional; rather projects are passed between functions typically at the end of each phase^.

Phase Review does have review points between phases, however. These are not investment or go/no go decision points as in Stage-Gate, however. NASA’s process uses TRLs (technological readiness levels) at these review points. These TLRs include criteria to gauge the project’s technical maturity and readiness to move forward.

No business criteria are used, however, nor do the TRL criteria judge the project’s value to the business. Recently, the TRL has been broadened to include business as well as technical readiness criteria, and some firms are beginning to use that new system.

The Phase-Gate process is a term sometimes used by a few authors to capture one or more of the NPD processes above.However, the term “phase gate” does not appear in the PDMA Handbook glossary; it may be a name used by some firms or authors to avoid using the terms above, which are strongly associated with a particular organization, like NASA, or are trademarked.

The “Inventor” of the Stage-Gate Process: Dr. Robert G. Cooper is ISBM Distinguished Research Fellow at Penn State University’s Smeal College of Business Administration, Professor Emeritus at McMaster University’s DeGroote School of Business, an Honorary Advisor at the Snyder Innovation Management Center, Syracuse University, and a Crawford Fellow of the Product Development and Management Association. He is a co-founder of Stage-Gate International.

Cooper is a thought leader in the field of New-Product Development and creator of the widely deployed Stage-Gate® system. He has published 12 books and more than 175 articles on the management of new products. He has won the IRI’s (Innovation Research Interchange) prestigious Maurice Holland Award three times. He has consulted for leading firms in the area of product innovation management and has helped dozens of Fortune 500 companies adopt best practices to enhance their new-product development efforts.

Cooper was a Professor at McMaster’s De Groote School of Business for 15 years; prior to that, he was at McGill University’s Bronfman School of Business for 10 years (where he was Associate Dean and Director of the MBA Program for his last 5 years there). Cooper holds Bachelor’s and Master’s degrees in Engineering, an MBA, and a PhD in Business.

How the system works

The Stages: Stage-Gate breaks the new product project into discrete and identifiable stages, typically 4–6 in number, as shown in the figure below. These stages are where the project team members execute prescribed or suggested actions or tasks. Each stage is designed to gather information and undertake tasks needed to progress the project to the next gate or decision point:

• Each stage contains important best practice key tasks to be done. These best practices have been uncovered from many research studies into what makes a new product a success, and what distinguishes very successful new products. Such tasks include, for example, Voice of the Customer (VoC) research to better understand user needs, iterative testing in real time with both users and technically to ensure product viability, and robust front-end homework to better define the product and project.
• The tasks within each stage are designed to gather critical information and to reduce the project’s unknowns and uncertainties. Each stage costs more than the preceding one: The system is an incremental commitment one. But with each step-wise increase in project cost, the unknowns and uncertainties are driven down, so that risk is effectively managed. This is a risk-mitigating model!
• The project is undertaken by a cross-functional team with members from various departments in the company: R&D, Engineering, Marketing, Sales, and Operations. The team starts small in the earlier stages, but team members are added by Gate 3. The team is led by a team leader (sometimes a project manager), and the team, not just the leader, is ultimately accountable for the success of the project.
• Each stage in the process is cross-functional: There is no “R&D stage” or “Marketing stage”; rather, every stage is Marketing, R&D, Operations, Engineering, and Sales. Each stage consists of a set of parallel tasks – that is, tasks within a stage are done concurrently and across functional areas. The project is not handed off from function to function at the completion of each stage—the cross-functional team moves the project from beginning to end.

The flow of the typical Stage-Gate system,is the most current version and is for major new product projects. Its five stages plus ideation are:

Discovery and Ideation: pre-work designed to uncover opportunities and to generate new-product ideas. Ideas come from many sources: from customers, technical research projects (see Stage-Gate-TD below), open innovation, and internally.

Concept (or Scoping): a quick, preliminary investigation and scoping of the idea or proposed project, shaping it into a concept—largely information and AI search, done mostly by desk research.

Build the Business Case: a much more detailed investigation involving primary research—both market and technical—leading to a business case, including product and project definition, the project justification (financial evaluation and risk assessment), and a project plan for the next stage or stages.

Development: the detailed design and development of the new product, including initial testing: alpha tests, POCs (proof of concept tests), tests of physical models, and simulation model testing, done internally; also, the design of the operations or production process.

Testing and Validation: includes more extensive tests to verify and validate the proposed new product, its marketing, and its production or operations process. Tests and trials occur in the lab or engineering department with prototypes; in the marketplace, such as beta tests or pilots with users; and in operations, such trial or limited production.

Launch: commercialization—beginning of full operations or production, marketing, and selling.

The post launch period ensues and entails monitoring performance and use, fixing issues, and improving the product. The team remains accountable until all is working well and the final Post Launch Review occurs.

The Gates: Preceding each stage is a gate or a go/kill investment decision point. Gates are not just about project approval; they are resource commitment and investment decision meetings. At these gate meetings, senior management meets with the project team to scrutinize the project: They review the progress of the project, determine that the project is ready to move forward, and either approve the proposed action plan and resources for the next stage (Go), ask for more information (Recycle), or stop the project (Kill or Hold).

Gates consist of three elements:

1. A set of suggested gate deliverables: what the project leader and team should bring to the decision point (e.g., the results of a set of completed tasks). These deliverables are fairly standard for each gate, are visible, and are agreed as the output of the previous gate. Expectations and next-stage objectives for the project team are thus very clear.
2. Gate criteria against which the project is judged include both project readiness and project value. Readiness is gauged on readiness criteria such as TLRs. To measure project value, the net present value (NPV) metric is often used; however, the financial data are frequently uncertain or wrong, rendering the NPV metric unreliable. Thus, scorecards that gauge more qualitative metrics can be used in conjunction with financial metrics to provide more insight and improve reliability.
3. Gate outputs: a decision (Go/Kill/Hold/Recycle), and in the event of Go, an approved action plan for the next stage, with resources committed (money, people required, and person-days), an agreed timeline, and a list of agreed deliverables and a date for the next gate.

Gates are usually staffed by senior managers from different functional areas, who control the resources required by the project leader and team for the next stage. They are called the gatekeepers and are a predefined group for each of the five gates. For larger projects requiring more resources, Gates 3, 4, and 5 are often staffed by the leadership team of the business. The entire project team should be at the gate meeting too.

New versions

Leading users of Stage-Gate have made many improvements to the process in order to accelerate projects, increase success rates, improve NPD productivity, and make the model suitable for different applications.

Customized Stage-Gate for Different Project Types: Not every R&D project is a major new-product project. Some are smaller, lower risk, such as product improvements, cost reductions, or “one-offs” for a single customer. Collectively, such smaller projects usually consume considerable resources, and thus, they still need astute management and guidance. Tailored versions of Stage-Gate—Lite and XPress with three or even two stages instead of five—have been developed for such low-risk projects.

Another important category of R&D projects is those whose emphasis is on the “R” in R&D, namely, science or more fundamental research projects. The deliverable is not a new product, but a new insight or understanding, a new capability, or a discovery that might yield a series of new products or a new manufacturing process: a “technology platform”.

The first technology development (TD) Stage-Gate model was pioneered by Thamhain in 1997, and others followed. These are typically two- or three-stage models; for example, ExxonMobil Chemical’s research model features two stages that precede its usual 5-stage new-product model.

Additionally, process development Stage-Gate models have also been designed and deployed to handle process developments. These are R&D projects for a new or improved manufacturing process, such as in the brewing industry or chemical industry.  

A Stage-Gate Model for Deploying Artificial Intelligence (AI): Although AI promises great efficiencies and improved effectiveness, deploying AI within the business has been a challenge. Firms currently face very high failures rates of AI adoption and deployment projects—as high as 80–95% according to reliable studies from the Rand Corporation, MIT, and an S&P Global study.

An AI adoption project is similar to a new product project, except that the source of the technology is external—typically an AI vendor—and the “customer” is an internal user group. The reasons for AI adoption project failure have been studied, and many are similar to those that cause new product failures. As a result, leading firms are modifying their new-product Stage-Gate process to accommodate AI adoption and deployment projects.

An IBM report recommends a seven-step “stage-gating” process for AI adoption in order to validate business impact before scaling. Similarly, an MIT study of European firms emphasizes the need for incremental funding and evidence-based decision gates to minimize waste and risk: “Norway’s postal and logistics group, Posten, and Spanish energy giant Repsol, introduced a Stage-Gate process for funding digital innovations. The approach makes it very easy for an initiative to receive the limited resources needed to get started. As the project progresses, it needs to prove itself at several stages to receive more resources.”

Stage-Gate is appropriate to undertake AI adoption-and-deployment projects within the business, but it is a significantly tailored AI process. This AI-adoption process is outlined in a Research-Technology Management article, showing a four-stage Stage-Gate process for these challenging AI adoption and deployment projects; its stages include building the business case, AI acquisition and alpha testing, pilots with users, and scaling to full production.

Parallel Processing: Simultaneous, concurrent execution or parallel processing helps to accelerate NPD. Re-designing the Stage-Gate process to overlap tasks—starting one task or stage before the preceding one is complete—can save time. Key tasks and even entire stages overlap, allowing projects to move ahead once information becomes reliable and stable—that is, not waiting for perfect information and the completion of the previous task or stage, as in the figure above.

In this way, the project can be accelerated—a rugby approach with multiple parallel tasks, rather than a relay race, with tasks strung out in series. It is even acceptable to move tasks from one stage to an earlier stage and, in effect, to overlap stages: starting one stage before the previous stage is finished. In so doing, key decision points and long lead-time items can be moved forward, such as the purchase of production equipment before field trials are complete.

An example is the rapid development of Covid vaccines: A traditional model for drug development is shown across the top of the Covid vaccine process figure above. The development of vaccines for COVID-19, however, followed the process across the bottom of the figure, employing parallel processing—tasks and stages done in parallel rather than sequentially—and also a rolling approval process (FDA approval moved along as clinical trial data became available). Additionally, manufacturing decision points were moved forward, resulting in “rapid manufacturing”. The new model reduced COVID-19 vaccine development from an expected 10–15 years to less than a year—a modern miracle of product development.

Iterative Stage-Gate: In a traditional Stage-Gate model, the product is defined before the project moves into the Development stage—a fairly early “frozen product definition”.Indeed, “sharp, early and fact-based product definition” was a fundamental principle of the original Stage-Gate approach, the research showing that a clear definition (technical objectives) would speed technical development work in Stage 3.

For some businesses, markets, and projects, however, the product definition must evolve as Development in Stage 3 proceeds: Often, customers are not clear on what they need: As Steve Jobs famously said, “People don’t know what they want until you show it to them”. Additionally, things change: Partway through development, a new customer need is identified, a new competitive product appears, or a new technical solution emerges, rendering the original product definition no longer valid.

Leading firms have thus made the idea-to-launch system much more adaptive—the product may be less than 50% defined as the Development stage begins. This process allows the product’s design and definition to change as the project moves through Stage 3: to adapt to new information, customer feedback, and changing conditions. That is, the product definition evolves and solidifies over the course of Development, and is not fixed or frozen at the beginning of Development.

Firms have thus built in multiple iterations that permit experimentation with users and technically. Each iteration consists of:

• Build. Build something to show the customer and to test technically—a virtual prototype, a model, a digital twin, or an MVP(minimally viable product).
• Test. Test the product version with users or customers, and also test the product technically.
• Feedback. Determine customer reaction and assess the technical test results.
• Revise. Reset thinking about the value proposition, benefits sought, the product’s design, and the technical solution.

Each iteration validates the product and moves the project closer to the final product design. This iterative approach promotes experimentation, encouraging project teams to fail often, fail early, and fail cheaply, a principle that Jobs applied at Apple (early iterative product versions often get negative feedback and need considerable revision). The median interval between iterative demos to customers is 17–24 weeks, according to one major European study, but is likely shorter today with the advent of digital tools.

The Agile-Stage-Gate Hybrid Model: Agile Development methods were developed in the 1990s in the software world to deal with projects facing uncertain and changing information.Agile Development consists of a series of very fast build-and-test iterations called sprints; it is adaptive and evolutionary—both the product definition and the project plan change as the project moves forward; it emphasizes frequent and fast delivery of results (product versions or increments) in weeks, not months, and in rhythmic takt time; and it is based on self-managed, empowered, and project-dedicated project teams.

Although Agile Development has its roots in the software industry, firms that produce physical products have also benefited from Agile methods. What leading manufacturers have done is simply to build Agile’s sprinting and iteration methods into the stages of their traditional gating process, replacing the more rigid plan-based project management tools such as Gantt charts, timelines, and milestones. By integrating Agile methods into traditional gating models, leading firms, such as Honeywell, GE, and LEGO, have reduced time to market as well as being able to respond quickly to changing customer requirements.

Adopting Agile doesn’t mean abandoning Stage-Gate. Indeed, the Stage-Gate framework provides important support for Agile development processes. Boehm and Turner, discussing the contrasts between plan-driven software development and Agile approaches, conclude that future projects will need both agility and discipline, which can be achieved by containing the Agile development methodology within the gate model.

Specific benefits when Agile is applied to physical-product projects are in the bar-chart figure above based on three large-sample studies. The greatest impacts are increased flexibility to react to changes, improved team morale, and higher customer satisfaction, as well as shorter development times and more adherence to the time schedule.

Agile-Stage-Gate is particularly suited to hybrid hardware-software projects—for example, for developing smart physical products with embedded software—where both software and hardware development people must work together. A study of six global firms using Agile methods within Stage-Gate showed that all six firms—from LEGO to Honeywell—had adopted the new model largely to accommodate combined software-hardware projects.

Eco-Stage-Gate: The goal of sustainability creates many opportunities for product innovation, including gaining competitive advantage and developing superior new products. It also creates challenges, such as regulatory compliance. Thus, companies are revisiting their NPD processes to better integrate sustainability principles and tools, but frameworks designed to create sustainable new products are lacking.

Eco-Stage-Gate, a modified version of the traditional Stage-Gate model, is specifically designed to prioritize environmental goals. By embedding eco-design tools, lifecycle thinking, and sustainability criteria in each stage and gate, Eco-Stage-Gate enables companies to develop eco-friendly products from inception to market launch. Key features include cross-functional collaboration, supply chain integration, and the use of green scorecards for decision-making at gates.

AI in Stage-Gate: Artificial Intelligence (AI) is poised to revolutionize all aspects of business, particularly new-product development (NPD). Leading and large early-adopter firms demonstrate that AI not only finds many applications in NPD but also offers substantial payoffs, such as 50% reductions in development times. The topic of AI in NPD and Stage-Gate is huge and rapidly changing, however, and beyond the scope of this Wikipedia section, but here is a quick overview.

Currently, more than 40 unique AI applications exist for AI in NPD. The varied nature of the applications—everything from doing market research to undertaking engineering design iterations—is exciting and somewhat overwhelming. A positioning map for “AI in NPD”, showing many different AI applications in NPD, gives a visual presentation of this complex landscape (see positioning map below). Across the map is where the AI application occurs in the Stage-Gate process from idea to launch. The vertical axis is the role of AI as an originator versus a facilitator: the creative and inventive side of AI versus AI doing what is done now, only faster and better.

Many of the best-practice activities embedded within the stages of Stage-Gate can be executed better, faster, or cheaper with AI. For example, tasks in the front end of the process in the positioning map—generating ideas, screening ideas, undertaking market and competitive analyses, and building the business case—can be undertaken by either readily available off-the-shelf AI tools or, with the right prompts, by LLMs such as Perplexity or ChatGPT.

Similarly, AI tools exist for concept generation, product design, and product testing. For example, digital twins can be used during the Development stage to simulate the product in use, thus providing initial user and technical testing feedback without having a physical prototype. During Stage 4, Testing and Validation, digital twins can be used to monitor test customers using the product and how it performs in a real use environment. AI can even be used to help plan the launch—sales-force planning, marketing communications, and pricing—and later to monitor the product in use after launch.

In spite of the reported benefits, AI adoption for NPD has been low: Only 23% of US and EU firms were using AI for any task in NPD by early 2024, which had risen modestly to 28% by 2025. The reasons for hesitation have been identified: A major reason is that AI installations do not yield the promised results— the AI failure rate for business in general was noted above to be about 80%. Firms are thus adopting more professional, structured, and gated approaches to implementing AI for NPD and in the business generally, as noted in the section above, “A Stage-Gate Model for Deploying Artificial Intelligence (AI)”.