Stage-Gate Framework

Most companies have no shortage of product ideas. What they have a shortage of is a reliable process for turning those ideas into products that actually ship and make money. The Stage-Gate framework is that process, and it's been used by over 3,000 companies worldwide since the late 1980s. P&G, 3M, LEGO, Siemens, Corning — the companies with the best new product track records almost all run some version of Stage-Gate.

I'll be upfront about my bias here: I think Stage-Gate gets a bad reputation from people who've only experienced rigid, bureaucratic implementations. Done well, it's not a creativity killer — it's a framework that forces honest evaluation at every step so you don't waste two years and millions of dollars on a product nobody wants.

Origins: Robert Cooper and the Science of New Product Success

Dr. Robert G. Cooper developed the Stage-Gate framework based on studying over 3,000 new product launches to understand why some succeeded and most failed. Cooper, a professor at McMaster University's DeGroote School of Business in Hamilton, Ontario, coined the term in a 1988 publication and published the landmark paper in Business Horizons in 1990.

Cooper's research identified a consistent pattern: companies with structured development processes had dramatically higher success rates than those using ad hoc approaches. But the key word was "structured," not "rigid" — Cooper always emphasized that the gates needed to be decision points, not rubber stamps.

Cooper was named the "World's Top Innovation Management Scholar" and received the PDMA's highest honors. He's spent the last 35 years refining the framework through multiple generations, adapting it to agile methodologies, digital products, and AI-driven development.

The Six Stages and Five Gates

The classic Stage-Gate model has six stages separated by five decision gates:

Stage 0: Discovery (Ideation)

Before the formal process begins, you need ideas worth evaluating. Discovery encompasses all the activities that generate new product concepts: brainstorming sessions, customer research, competitive analysis, technology scouting, voice-of-customer studies, and strategic planning sessions.

The best companies don't treat ideation as a one-time event. They build systematic idea-generation processes: innovation portals where employees submit concepts, regular customer advisory board meetings, technology watch programs, and "kill your own product" exercises where teams imagine what a competitor would build to displace them.

Gate 1: Idea Screen

Decision: Go / Kill / Recycle

The first gate is a quick, inexpensive filter. Does this idea align with our strategy? Is the market potentially large enough? Do we have the technical capability to build it? Is there a clear customer need?

Gate 1 should be fast and low-ceremony. You're not asking for a business case — you're asking whether this idea is worth spending time investigating further. Most ideas should die here, and that's healthy.

Stage 1: Scoping

A quick, inexpensive assessment of the idea's technical merits and market prospects. This is desk research, not fieldwork: market size estimates, competitive landscape overview, preliminary technical feasibility, and rough financial projections.

The goal is to gather enough information to make a more informed Go/Kill decision at Gate 2 without investing significant resources.

Gate 2: Second Screen

Decision: Go / Kill / Hold / Recycle

A more rigorous evaluation based on Scoping results. The criteria are tighter: is the market attractive enough? Is the competitive position defensible? Are the technical risks manageable? Does the rough financial case make sense?

"Hold" and "Recycle" are important options here. Hold means the idea is good but the timing is wrong (maybe a technology dependency isn't ready yet). Recycle means the core insight is valuable but the concept needs rethinking.

Stage 2: Build Business Case

This is the heavyweight stage before major resource commitment. It includes:

• Detailed market research: customer interviews, surveys, conjoint analysis, competitive deep-dive
• Technical assessment: feasibility studies, prototype concepts, architecture decisions
• Financial analysis: detailed revenue projections, cost estimates, NPV/IRR calculations, sensitivity analysis
• Project plan: development timeline, resource requirements, risk identification

The business case is the document that justifies committing significant development resources. It should be honest about assumptions and risks, not a sales pitch designed to get through the gate.

Gate 3: Go to Development

Decision: Go / Kill / Hold / Recycle

This is the most consequential gate. Saying "Go" here commits the organization to full development investment. Gatekeepers (typically senior leadership) evaluate the business case quality, market attractiveness, competitive advantage, technical feasibility, and financial projections.

Cooper emphasizes that Gate 3 needs tough gatekeepers. The most common failure in Stage-Gate is rubber-stamping projects through Gate 3 because the organization has already invested time in Stages 1-2 and doesn't want to "waste" that investment. Sunk cost fallacy kills more products than bad markets do.

Stage 3: Development

The actual product gets built. Design, engineering, coding, manufacturing process development, regulatory compliance, quality assurance. This is where the majority of the budget gets spent.

Modern implementations increasingly integrate agile sprints within Stage 3, breaking the development work into iterative cycles with regular check-ins rather than one long waterfall process.

Gate 4: Go to Testing

Decision: Go / Kill / Hold / Recycle

Is the product functionally complete? Does it meet the specifications from the business case? Are there showstopper issues? Is the test plan ready?

Stage 4: Testing and Validation

In-house testing, alpha/beta testing with real customers, pilot production runs, test marketing in limited geographies or segments. The goal is to validate three things: the product works as designed, customers actually want it, and the production/delivery process scales.

Testing often reveals that the product works but the value proposition needs adjustment, or that the target market is slightly different than assumed. This is why Stage-Gate needs to accommodate learning and iteration, not just binary pass/fail.

Gate 5: Go to Launch

Decision: Go / Kill / Hold / Recycle

The final gate before full commercial launch. All test results reviewed. Launch plan approved. Manufacturing ready. Sales and marketing resources committed.

Stage 5: Launch

Full commercial launch: production ramp-up, marketing campaign execution, sales force activation, distribution rollout. Post-launch monitoring tracks actual performance against the business case projections.

The Gate Decision Framework

Each gate uses a consistent decision structure:

Cooper's research shows that the Kill decision is the hardest but most important. Companies that are reluctant to kill projects end up with bloated portfolios of mediocre products competing for the same limited resources. The best companies kill early and kill often.

What's Changed: Agile-Stage-Gate and Beyond

The original Stage-Gate was designed for physical product development in the 1980s-90s. Several generations of evolution have adapted it:

The Agile-Stage-Gate Hybrid

The most significant evolution. Rather than treating each stage as a waterfall block, modern implementations embed agile sprints within stages. Stage 3 (Development) might consist of two-week sprints with regular demos and retrospectives, while the gate structure provides the strategic go/kill decisions that agile lacks.

Companies like Honeywell, LEGO, Danfoss, and Tetra Pak have implemented Agile-Stage-Gate hybrids. The results are compelling: Cooper's research shows 30% reduction in time-to-market, improved team morale, and better alignment between technical development and business objectives.

The key insight: agile is great for managing how to build something, but terrible at deciding what to build and whether to build it. Stage-Gate handles the what and whether; agile handles the how.

NexGen Stage-Gate

Cooper's next-generation model added several capabilities:

• Scalable process: Stage-Gate Lite for lower-risk projects (modifications, extensions) and Stage-Gate XPress for very simple projects. Not every product needs five gates and six stages.
• Flexible gates: Gates can be conditional ("Go with conditions"), and stages can overlap when risk is manageable.
• Built-in iteration: Stages explicitly include build-test-feedback-revise loops rather than linear progression.
• Portfolio-level gates: Gates evaluate projects not just individually but against the full portfolio — is this the best use of our development resources compared to everything else in the pipeline?

AI Integration (2024-2025)

AI is transforming Stage-Gate at every stage:

• Discovery: AI tools analyze market trends, patent filings, and customer feedback to generate and screen ideas at scale
• Scoping: AI-powered market sizing and competitive analysis reduce Stage 1 from weeks to days
• Business Case: Predictive models assess commercial viability with greater accuracy, incorporating more data points than human analysts can process
• Development: AI-assisted design, code generation, and testing compress Stage 3 timelines
• Testing: AI analyzes test results and customer feedback patterns faster, identifying issues human reviewers miss

Cooper has noted that AI adoption in NPD processes rose from 23% in early 2024 to 28% by 2025, with the trajectory suggesting 40%+ within two years. He's identified over 40 specific AI applications across the Stage-Gate process.

Real-World Implementations

P&G: The Poster Child

Procter & Gamble is perhaps the most famous Stage-Gate practitioner. Their "SIMPL" (Successful Initiative Management and Product Launch) process is a customized Stage-Gate implementation that has helped P&G maintain one of the highest new product success rates in consumer goods. P&G's innovation pipeline generates billions in annual revenue from products launched within the past five years.

3M: NPI/NTI Systems

3M runs parallel Stage-Gate systems: NPI (New Product Introduction) for product innovations and NTI (New Technology Introduction) for platform technologies that enable multiple future products. This dual-track approach lets 3M invest in both near-term product launches and longer-term technology platforms.

LEGO: Speed Through Structure

LEGO's implementation demonstrates that Stage-Gate doesn't have to be slow. Their Story Starter product went from concept to market in 12 months using an Agile-Stage-Gate hybrid. The gate structure ensured strategic alignment and commercial viability while agile sprints within stages kept development velocity high.

Corning: Technology-Stage-Gate

Corning adapted Stage-Gate specifically for materials science innovation, where development timelines are measured in years and technical uncertainty is extremely high. Their version emphasizes Stage 2 (Build Business Case) heavily, because a bad bet on materials R&D is far more expensive than a bad bet on consumer software.

Stage-Gate Variations

Criticisms and Honest Assessment

Stage-Gate isn't perfect, and pretending it is does the framework a disservice:

Bureaucracy risk is real. Organizations that add documentation requirements at every gate, require committee meetings for every decision, and treat gates as presentations rather than decisions turn Stage-Gate into a paperwork exercise that slows innovation without improving outcomes.

It can bias toward incremental innovation. The gate criteria naturally favor projects with clear business cases, proven markets, and quantifiable returns. Genuinely disruptive innovations often can't pass Gate 3 because the market doesn't exist yet and the business case requires too many assumptions. This is why companies like 3M run separate processes for technology exploration.

Kill discipline is hard to maintain. Despite Cooper's emphasis on tough gates, most organizations struggle to kill projects that have momentum, political support, or sunk costs. Research consistently shows that companies let too many projects through gates and spread resources too thin across too many initiatives.

Sequential thinking can slow response. In markets that move fast (software, digital, consumer tech), waiting until Stage 4 to test with real customers can mean arriving after competitors who shipped earlier and iterated. The Agile-Stage-Gate hybrid addresses this, but not all organizations have made the transition.

The framework is most powerful when leaders treat gates as genuine decision points and are willing to kill projects. It's weakest when gates become status updates and "Go" is the only decision anyone's willing to make.

Thought Leaders and Key Figures

Organizations and Resources

• Stage-Gate International — Cooper's consulting firm; resources, training, and certification
• Product Development and Management Association (PDMA) — Founded 1976; 3,500+ members; publishes the Journal of Product Innovation Management; annual conferences and best-practice research
• Innovation Research Interchange (IRI) — R&D management community using Stage-Gate methodologies
• McKinsey & Company — Publishes innovation process research building on Stage-Gate principles
• Harvard Business School — Case studies featuring Stage-Gate implementations across industries
• Cooper's research page at McMaster University — Academic publications and working papers

Frequently Asked Questions

How is Stage-Gate different from Waterfall?

Waterfall is a sequential development methodology — it prescribes how to build a product (requirements, then design, then build, then test). Stage-Gate is a business decision framework — it prescribes when and how to evaluate whether a product should continue getting investment. You can run agile development within Stage-Gate stages. The gates are business decisions; the stages contain whatever development methodology makes sense for your product.

When should we kill a project vs. recycle it?

Kill when the fundamental market need is insufficient, the competitive window has closed, or the technical barriers are insurmountable. Recycle when the core customer insight is valid but the product concept, business model, or target market needs rethinking. The distinction matters: killing a project means the idea is dead; recycling means the insight lives but needs a new execution approach.

How much time and money does a Stage-Gate process add?

Good Stage-Gate actually saves time and money by killing bad projects early (before significant development investment) and focusing resources on winners. Cooper's data shows that companies with effective Stage-Gate processes have 30% faster time-to-market and 63-78% new product success rates, versus ~25% success rates for companies without structured processes. The overhead of gates is far less than the cost of failed products that should have been killed earlier.

Can Stage-Gate work for software and digital products?

Yes, through the Agile-Stage-Gate hybrid. Software companies keep the gate structure for strategic go/kill decisions but use agile sprints within stages for iterative development. The gates ensure business alignment; the sprints ensure development velocity. Companies like Honeywell and Danfoss have implemented this hybrid successfully across both hardware and software product lines.

Which Stage-Gate variation should we use?

Match the rigor to the risk. Full Stage-Gate for major new products with significant investment and market uncertainty. Stage-Gate Lite for modifications and extensions where the market is known. Stage-Gate XPress for minor updates. Stage-Gate TD for early-stage technology research that isn't ready for commercial evaluation yet. Most companies run multiple variations simultaneously for different project types.

What makes a good gatekeeper?

Gatekeepers need three things: authority to allocate resources, willingness to kill projects, and enough domain knowledge to evaluate the business case. The best gate meetings include cross-functional senior leaders (marketing, R&D, finance, operations) who can assess both commercial viability and technical feasibility. The worst gate meetings are rubber stamps where the gatekeeper has already decided "Go" before the presentation starts.

Why do most project kills happen too late?

Sunk cost fallacy and organizational politics. Once a project has consumed months of work and attracted a champion, killing it feels like waste. But the research is clear: projects killed at Gate 2 or Gate 3 save dramatically more resources than projects killed at Gate 4 or Gate 5. Building a culture where early kills are celebrated ("we saved $2M by killing this at Gate 2") rather than mourned is essential.

How is AI changing the Stage-Gate process?

AI is accelerating every stage: automated idea screening in Discovery, AI-powered market analysis in Scoping, predictive commercial models in Business Case, AI-assisted development in Stage 3, and pattern recognition in Testing. Cooper's 2025 research identifies 40+ AI applications across the process. The most impactful near-term application is using AI in Stage 2 to build better business cases with more data and fewer assumptions, improving the quality of Gate 3 decisions.

Sources & References

1. Cooper, R. G. (1990). "Stage-Gate Systems: A New Tool for Managing New Products." Business Horizons, 33(3), 44-54.
2. Cooper, R. G. (2017). Winning at New Products: Creating Value Through Innovation, 5th Edition. Basic Books. Amazon
3. Cooper, R. G. (2014). "What's Next? After Stage-Gate." Research-Technology Management, 57(1), 20-31.
4. Stage-Gate International. stage-gate.com
5. Product Development and Management Association. pdma.org
6. Cooper, R. G. & Sommer, A. F. (2016). "The Agile-Stage-Gate Hybrid Model." Research-Technology Management.
7. Christensen, C. M. (2015). "What Is Disruptive Innovation?" Harvard Business Review. hbr.org

Written by Conan Pesci | Created: April 3, 2026 | Last Updated: April 3, 2026