# Gate Design for New Businesses: Stopping What Should Stop Without Killing What Could Grow

- Author: Shuichiro Ogawa (Design Researcher / Design Consultant), https://jamjet-iqbal.com/
- Published: 2026-11-11; updated: 2026-11-11
- Canonical (HTML): https://jamjet-iqbal.com/en/notes/new-business-gate-design/
- Japanese version: https://jamjet-iqbal.com/notes/new-business-gate-design/ (Markdown: https://jamjet-iqbal.com/notes/new-business-gate-design.md)
- Language: en
- Keywords: new-business-development, stage-gate, gate-review, metered-funding, kill-criteria, escalation-of-commitment, portfolio-management, corporate-venturing, innovation-governance, business-design

> This note examines gates that decide, stage by stage, whether to keep investing in a new business. It draws on 11 public programs and evaluations, 5 materials from research firms and consultancies, 20 first-hand materials by practitioners who built and run gates, and 14 academic works, and it checks figures and quotations against the full texts. Gate failures fall into three types: projects that should stop do not stop, projects that could grow are stopped too early, and the gate becomes an approval ritual. For the first type, studies of the mechanism are consistent. Executives showed a strong bias toward continuing failing new products; precommitment to self-specified decision rules helped little, whereas a predetermined decision rule or a new decision maker helped relatively more (Boulding et al. 1997). Managers were more reluctant to stop really new products (Schmidt and Calantone 1998), and in teams, hope was more strongly tied to continuing than fear was tied to terminating (Huang et al. 2019, 66 teams). Practitioner remedies separate the people who decide from those who started the project (cross-department review at CyberAgent, independent reviewers in the UK Gateway Review) and release money by validation stage (McGrath, Ries, and Recruit's four stages). However, in a survey of 888 people involved in new businesses at large Japanese firms, 28.5% of written exit criteria were not followed, and criteria written before launch were not followed in 38.1% of cases, compared with 17.2% for criteria written after launch (a survey by a support firm, showing association only). Criteria therefore need to come with a decision on who reviews them and when. There is also evidence for the second and third types. Repeatedly applying strict and objective gate criteria made projects inflexible and lowered the performance of novel products (Sethi and Iqbal 2008, 120 projects), and the transition-rate standards of the US SBIR program were judged to risk limiting exploratory work (National Academies 2026). Since 2022, average investment before a new venture breaks even has fallen from about 125 million to about 77 million US dollars, and the leash has become shorter (McKinsey 2025). Cooper, who developed Stage-Gate, proposes shifting the purpose of gates from optimizing investment decisions to maximizing learning velocity and avoiding the premature killing of promising projects (Cooper and Zhao 2026, a proposal rather than evidence).

## What are gates meant to prevent?

A **gate** in new-business development is a checkpoint at which, at predetermined stages, a decision is made on whether to keep investing in a business idea.
According to Stage-Gate International, which promotes the Stage-Gate method, cross-functional business leaders at each gate measure every project against the same set of success criteria and decide to go, kill, hold, or recycle it.
When the decision is to go, resources for the next stage are committed.

Gates are meant to prevent more than one kind of failure.
The first is that a project that should stop does not stop and keeps consuming resources.
The second is that a project that could have grown is stopped too early.
Cooper and Zhao (2026) write that early Stage-Gate systems in the US were biased toward avoiding the first type (false positives), and that dynamic industries should prioritize avoiding the second type (false negatives).
The third is that the gate becomes a formality.
Jones of Stage-Gate International asks whether a company's gate meetings are rubber-stamping rituals that approve most projects with little or no pushback.

Which of these failures occurs, then, and why?
This note draws on 11 public programs and evaluations, 5 materials from research firms and consultancies, 20 first-hand materials by practitioners who built and run gates, and 14 academic works, and it checks figures and quotations against the full texts.
For works whose full texts could not be obtained and which are therefore described from their abstracts, "abstract" is added after the year.

## Why do projects that should stop not stop?

Studies that examined the mechanism behind projects that do not stop are consistent with one another.
Boulding et al. (1997, abstract) conducted controlled experiments with senior executives and found a strong bias toward continued commitment to failing new products.
According to the same abstract, improving the quality of information did not greatly reduce the bias, nor did precommitment at launch to self-specified decision rules.
The methods that worked relatively better were precommitment to a predetermined decision rule and the introduction of a new decision maker at the time of the stop decision, and the latter worked marginally better.
None of the methods removed the bias completely.

The bias becomes stronger with the novelty of the idea and with team emotions.
In the experiment of Schmidt and Calantone (1998, abstract), managers who saw identical poor performance data were more optimistic about, more committed to, and more likely to continue the project when it involved a really new product, and they more often let it proceed all the way to commercialization.
Huang et al. (2019) followed 569 decisions by 66 teams in an entrepreneurial simulation and reported that the relationship between group hope and continuing a failing venture was stronger than the relationship between group fear and terminating it.

Organizational incentives also make projects hard to stop.
Simester and Zhang (2010, abstract) showed theoretically that rewarding success encourages managers to forge ahead even when demand is low, whereas rewarding decisions to kill products undermines the rewards for success, so that the design of incentives contains an inherent conflict.
Manso (2011, abstract) showed that the optimal incentive scheme for motivating innovation combines tolerance for early failure with reward for long-term success.

The words of executives and practitioners point to the same mechanism.
Fujita (2025), the president of CyberAgent, writes that even he finds exit decisions difficult.
When he has been involved in the decision to start a business and can picture the faces of the people he brought in, he postpones the decision even though he knows that continuing will expand the loss.
McGrath (2009) lists statements that teams use to resist an exit and states that if most team members agree with about a third or more of them, the team is at risk of escalated commitment; she recommends developing a disengagement plan as carefully as the business plan.
The UK guidance on gateway reviews of public projects (NISTA 2026) states that it is normal for officials developing a project to be overly optimistic about its cost and timetable (optimism bias) and that an independent reviewer is needed to recognize this.
The UK appraisal guidance also states that sunk costs should not affect decisions about what to do next (HM Treasury 2026).

The number of projects also matters.
Cooper writes that when there are too many projects in the pipeline, people are spread too thinly and even important projects are under-resourced, and he recommends monitoring projects on an ongoing basis and killing bad projects in real time (InnovationManagement.se).

## What helps to stop projects?

### Separating the decision makers from the project's originators

One practitioner answer is to separate the people who decide whether to stop from the people who started the project.
At CyberAgent, a business that meets certain criteria is sent for review, and people from other departments analyze it objectively in several teams and decide whether it continues or exits (Fujita 2025).
Fujita explains that the company built this system because, left to himself, even he finds it hard to decide on an exit.
The UK Gateway Review assigns reviews to independent reviewers, such as senior officials from a different ministry (NISTA 2026).
Among Japanese public programs, NEDO's support for deep-tech startups has external experts review, at the time of fundraising, whether the project may continue and receive further support, and it may stop support midway depending on the result (NEDO 2026a).
JST's external evaluation of the Moonshot program refers decisions on discontinuing research to the governing committee (JST 2025).

Experiments also show that changing the decision maker or adding an outside view works relatively well.
The effect of a new decision maker in Boulding et al. (1997, abstract) was described in the previous section.
Behrens and Ernst (2014, abstract) conducted an experiment with 137 R&D managers who had to decide on projects with increasing losses within a stage-gate system, and reported that visual decision aids and consultant advice reduced decisions to keep funding a losing course of action, and that using both had the strongest effect.
According to the same abstract, the effect was larger before commercialization.
Eliëns et al. (2018) showed, in both a correlational study and a randomized experiment with 184 experienced practitioners, that inducing a rational thinking style before gate decisions made escalation of commitment less likely than relying on intuition.

### Setting criteria in advance and also deciding how to review them

X, CyberAgent, and Recruit all set exit criteria in advance.
Teller (2016a) of X, Google's research and development organization, writes that teams agree on "kill signals" when a project starts, and that reaching them (or, in many cases, failing to reach them within a certain timeframe) indicates that the team should walk away from the project.
CyberAgent sends businesses that meet criteria such as "two consecutive quarters of declining revenue and profit" for review (Fujita 2025).
Recruit's Ring program sets a deadline for each stage, and a business that has not met the criteria by then is withdrawn (Toyo Keizai 2022).

However, criteria set in advance are not always followed.
In the experiment of Boulding et al. (1997, abstract), self-specified rules set in advance helped little.
The Innovation Research Institute (2026) surveyed 888 people involved in new businesses at companies with at least 1,000 employees and found that 72.7% had written exit criteria, and that 28.5% of these did not follow the criteria when the decision came.
Criteria written before the business started were not followed in 38.1% of cases, about twice the 17.2% for criteria written after the start, and similar to the 40.8% of cases in which a verbal agreement alone did not work well.
The institute suggests that criteria set before the start may be used less because they reach the decision point without reflecting the facts learned through validation, and it proposes deciding in advance when, by whom, and on what basis the criteria will be reviewed.
The survey was conducted by a firm that sells new-business support, and it shows only associations.
Ishikawa, formerly of Recruit, also states that business plans should be revised every year and that the initial decision criteria should be changed as needed, because criteria that reflect what has been learned after launch should be more accurate (Sogyo Techo).

Setting criteria in advance and revising them later may seem contradictory.
To reconcile the two, what is decided in advance is not the criteria themselves but the timing of their review and the people who review them.
When the reviewers are separate from the project's originators, revisions to the criteria are less likely to become an excuse for those who want to continue.

### Releasing money by validation stage

The practitioner methods read here all recommend releasing money by validation stage rather than all at once.
McGrath (2010) states that if the results live up to the assumptions, more funds are released until the next checkpoint, and if they do not, the project is altered or killed.
Ries (2017) recommends giving teams clear, metered funding to preserve scarcity, and creating a growth board.
At X, the biggest risks become milestones, and knocking down a major risk unlocks more budget or permission to grow the team a little (Teller 2016a).
Recruit's Ring program moves through four stages (MVP, SEED, ALPHA, and BETA), checking in turn whether there is a need, whether the business can be profitable, whether it can be offered widely, and whether more should be invested; a business that meets the conditions moves to the next stage with a larger budget (Toyo Keizai 2022).
Osterwalder and Viki (2021) recommend investing small amounts in many teams, killing at least half of the projects that cannot produce evidence after a short time, and giving follow-up investment only to teams that produce evidence.
This "half" is a claim of the method, not a measured value.
Bello of McKinsey stated that new businesses are most successful when funded in a milestone-based way, with separate tranches for feasibility and go/no-go testing, business design, the MVP, and scaling (CFO Dive 2025).

Staged funding is also the investment pattern of venture capital.
Gompers (1995, abstract) showed, with data from a random sample of 794 venture-capital-backed firms, that venture capitalists periodically gather information and maintain the option to discontinue funding projects with little probability of going public.
Ewens et al. (2018) showed that as the cost of experimentation fell, an investment approach became more common in which investors provide a little funding to more startups that they are more likely to abandon.

Nevertheless, the empirical effect of staging itself is conditional.
Klingebiel and Adner (2015, abstract) reported that, in product innovation, sequencing resource allocation was associated with higher new product sales, whereas low initial commitment and reallocation each showed no direct effect on sales.
However, combinations in which low initial commitment and reallocation matched (both present or both absent) increased performance.
One reading is that sequencing is associated with sales, while starting small and moving resources do not work when only one of them is introduced; what matters is whether the two fit together.

### What should gates look at?

What gates look at in each stage has also changed in practitioner methods.
Teller (2016b) of X states that the team chooses what to work on first not by what is most fun or easiest but by the most likely reason the project will not make it.
Viki (2020) of Strategyzer recommends deciding whether to continue or stop on the basis of evidence of how close each project is to a business model that works, rather than on projections in a business case.
Osterwalder and Viki (2023) treat stronger evidence as more risk reduced and write that, in the early stages, a significant portion of projects are retired because of insufficient evidence.
Libarikian of Leap by McKinsey stated that the milestones in the first year of a new business are not financial but concern the product, customer feedback, attracting talent, and the like (CFO.com 2023).

Practice in Japanese companies differs from this.
In a survey by PwC Consulting (2025), 64% of "successful companies" (those with a project that reached investment recovery) and 63% of other companies used medium-term (within three years) contributions to sales and profit as a KPI for new businesses.
From the period of these KPIs and the frequency of organizational changes, PwC reads a tendency among Japanese companies to judge the success or failure of a new business within three years.
The survey did not ask directly whether companies stop projects after three years.

## Where do excessive stopping and formalization occur?

When the mechanisms for stopping are strengthened, the second and third types of failure appear.
Sethi and Iqbal (2008, abstract) reported, from a survey of 120 projects that used Stage-Gate, that repeated application of strictly enforced and objective evaluation criteria made projects more inflexible, that inflexibility led to learning failure, and that learning failure lowered the performance of novel new products.
The effect of inflexibility on learning failure was stronger when the firm's technological environment was turbulent.
Introducing gate conditionality, a relaxation that lets projects pass with conditions, did not mitigate the adverse effect.

Public programs face the same criticism.
The US SBIR program, which funds research and development by small businesses, requires companies that received 21 or more Phase I awards over the past five years to have at least 0.25 Phase II awards per Phase I award, and companies that fail cannot apply for Phase I for one year (SBA).
From April 2023, increased standards apply to more experienced firms.
GAO (2024) found that six companies did not meet the new standards in fiscal year 2023 and judged that few businesses are affected.
The National Academies (2026) examined SBIR at the Department of Defense and wrote that enforcing higher transition rates may limit exploratory work.
In the same report, the Phase I to Phase II conversion rate for experienced firms had dropped to about 0.4 by fiscal year 2020.
When the pass rate of a stage itself becomes a target, there is an incentive to choose only ideas that are likely to pass.

The workload of reviewers can also disadvantage novel ideas.
Criscuolo et al. (2017, abstract) showed, in the selection of R&D projects at a professional services firm, that projects with intermediate novelty were more likely to receive funds, while too much novelty was selected against.
According to the same abstract, a high panel workload reduced the preference for novelty, whereas diversity of panel expertise and a shared location between panel and applicant increased it.
Böttcher and Klingebiel (2025) showed, in a model of how to aggregate the evaluations of several decision makers, that when budgets are tight and only a few projects can be selected, letting each evaluator rank projects often performs better than averaging scores or counting approval votes.
Delegation to experts performed better only when the organization employs the relevant experts and can assign projects to them.

Practitioners also disagree.
Libarikian of Leap by McKinsey stated that CEOs are at risk of pulling the plug too early (CFO.com 2023).
Cooper and Zhao (2026) propose that gates should maximize learning velocity rather than optimize investment decisions, and that the default should be a conditional go unless there is a legitimate reason not to proceed.
By contrast, Fishman (2026), formerly of Mozilla, states that he would rather shut something down prematurely than let a team flail for twelve months without any end in sight.
Neither position is supported by an empirical comparison.
Walrave et al. (2022) followed one collaborative project between a software firm and a hardware firm and showed that mixing Agile and Stage-Gate in development may cause projects to fail (the title describes it as "keeping up appearances").
Because this is a single case, it does not provide a general evaluation of hybrid methods.
Combining the results of Sethi and Iqbal (2008, abstract) with those of Schmidt and Calantone (1998, abstract) suggests that novel ideas are hard for their originators to stop and easy for strict gates to crush.
One reading is therefore that novel ideas are exposed to both errors: not stopping when they should stop, and being crushed when they could have grown.

## What has changed since 2022?

Three changes around gates can be seen since 2022.
The first is that the leash on money has become shorter.
In a McKinsey (2025) survey of 715 respondents, the average investment before a new business broke even fell from about 125 million US dollars in the previous year to about 77 million US dollars.
Forty-seven percent said that new ventures launched in the past five years had met or exceeded expectations.
Bello of McKinsey stated that the appetite for giving businesses a long leash has gone down and that companies expect returns earlier than before (CFO Dive 2025).

The second is a proposal to move gates from reviews at fixed times to continuous monitoring.
Cooper (2025) wrote that predictive analytics will update the business case continuously and reduce the need for formal, discrete reviews.
Cooper (2026) described a design in which AI scores projects and recommends go, kill, or hold but never has a vote, and prompts an interim review when metrics drift outside predefined bands.
These are proposals by the developer of the method, and their effects have not been tested.
In Japan, Recruit uses a tool trained on several years of comments by Ring reviewers to show proposers, before they apply, the points on which they are likely to be challenged (ITmedia 2025).
CyberAgent revised its system with exit criteria for businesses in October 2022 and now decides on business integration and exits at a meeting held every half year (CyberAgent).

The third is the tightening of standards in public stage-based programs and the evaluation of their side effects.
The increased SBIR standards (from April 2023) were described in the previous section.
The EU's EIC Transition selected 40 projects from 611 submissions in its 2025 call (EIC 2026).
NEDO's support for deep-tech startups reports the number of projects that passed stage-gate review in its announcements of selected projects: 5 of the 18 selected in a 2025 round and 2 of the 15 selected in a 2026 round had passed such a review (NEDO 2025, 2026b).
Because the number of projects that underwent the review is not published, the pass rate is unknown.
The end-of-program evaluation of NEDO's earlier startup support program also noted that the program's effectiveness could be evaluated if the share of firms continuing to use the programs for each growth stage were known, which means that continuation rates by stage have not been measured (NEDO Research Evaluation Committee 2025).

## How should gates be designed for each type of failure?

The materials above can be rearranged into designs for each type of failure.

| Type of failure | Mechanism | Gate design | Sources |
|---|---|---|---|
| Projects that should stop do not stop | Originators' involvement, optimism, hope, sunk costs, incentives that reward only success | Separate the people who decide whether to stop from the project's originators, and add outside advisers | Boulding et al. 1997, abstract; Behrens and Ernst 2014, abstract; Fujita 2025; NISTA 2026 |
| Agreed criteria are not followed | Criteria set before launch do not reflect what has been learned | Decide in advance when and by whom the criteria will be reviewed, rather than only the criteria themselves | Innovation Research Institute 2026; Ishikawa (Sogyo Techo) |
| Too much money is committed | A lump-sum budget with no time to stop | Release money by validation stage, start small, and move resources according to results | McGrath 2010; Ries 2017; Toyo Keizai 2022; Klingebiel and Adner 2015, abstract |
| Too many projects, all under-resourced | New ideas are added without stopping others | Monitor projects continuously and kill bad projects in real time | Cooper (InnovationManagement.se) |
| Projects that could grow are stopped early | Repeated strict and objective criteria, pass-rate targets, reviewer workload | Judge novel ideas on evidence of learning, do not target pass rates, reduce reviewer workload, and mix reviewers' expertise | Sethi and Iqbal 2008, abstract; National Academies 2026; Criscuolo et al. 2017, abstract; Cooper and Zhao 2026 |
| The gate becomes an approval ritual | Meetings that pass nearly everything, and stop decisions not treated as learning | Discuss stop decisions openly and share what was learned | Jones (Stage-Gate International) |

In this table, separating the decision makers and releasing money by stage support the designs in the other rows.
When the decision makers are separate, revisions to the criteria are less likely to become an excuse for those who want to continue.
When money is released by stage, a single decision does not have to stop a project, and novel ideas keep a path to gather the next piece of evidence with a small amount of money.

## What this review can and cannot claim

This review can claim that the mechanisms by which projects that should stop do not stop (originators' involvement, optimism, hope, and incentives) are consistently shown in experiments and theory, that separating decision makers and adding outside advice worked relatively well in experiments, that a survey found repeated strict criteria to disadvantage novel products, and that practitioner methods are converging on staged funding and decisions based on evidence.

However, it cannot claim three things.
First, no study was found that randomly assigned gate designs in corporate new businesses and compared business outcomes.
The effects of changing the decision maker and of outside advice were measured in experiments with simulated decisions.
Second, no survey was found that counted what gates Japanese companies use and how many projects they stop.
Third, Cooper's gates built on AI and learning velocity are at the proposal stage, and there is no evidence of their effect.

## Related notes

- [What Counts as Evidence in New-Business Acceptance Research: From Words to Behavior, and Then Generative AI](https://jamjet-iqbal.com/en/notes/new-business-acceptance-research/): Practice and research on acceptance research for new businesses, which concerns what gates should treat as evidence

## Unverified items

No unverified statement remains in the main claims.
The full texts of the following works could not be obtained (publishers refused automated retrieval, and Unpaywall listed no open version), so they were used only within the scope of their abstracts: Boulding et al. (1997), Schmidt and Calantone (1998), Simester and Zhang (2010), Behrens and Ernst (2014), Sethi and Iqbal (2008), Klingebiel and Adner (2015), Criscuolo et al. (2017), Gompers (1995), and Manso (2011).
Ewens et al. (2018) was checked against the full text of the NBER working paper version (2018).
Figures cited in Cooper's articles on AI (such as the share of wrong approval decisions) were not used because their sources and samples are not shown.

## References

All sources were accessed on 2026-11-11. The ledger and verification records are in `source/review/new-business-gate-design/industry.md`.

### Public bodies

- European Innovation Council. 2026. "EIC Selects 40 New Transition Projects to Bring Research Results Closer to Market." https://eic.ec.europa.eu/news/eic-selects-40-new-transition-projects-bring-research-results-closer-market-2026-02-09_en
- HM Treasury. 2026. *The Green Book (2026)*. https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government/the-green-book-2026
- Japan Science and Technology Agency (JST). 2025. "Mūnshotto-gata kenkyū kaihatsu jigyō Reiwa 7 nendo purojekuto gaibu hyōka (chūkan hyōka) ni tsuite (mokuhyō 1)" [External (interim) evaluation of Moonshot projects in fiscal year 2025, Goal 1] (in Japanese). https://www.jst.go.jp/moonshot/program/pdf/ms1_r7_gaibu_pj.pdf
- National Academies of Sciences, Engineering, and Medicine. 2026. "Experienced SBIR/STTR Firms." Chapter 9 in *Review of the SBIR and STTR Programs at the Department of Defense*. https://www.nationalacademies.org/read/29329/chapter/11
- National Infrastructure and Service Transformation Authority (NISTA). 2026. *Guidance to Conducting Gateway Assurance Reviews: For International Use*. https://assets.publishing.service.gov.uk/media/6a7ddaf40ad4bc475be7de10/NISTA_IGR_Gateway_Assurance_Reviews_Guidance_August2026.pdf
- New Energy and Industrial Technology Development Organization (NEDO). 2025. Announcement of selected projects, DTSU 7th call and GX 4th call (in Japanese). https://www.nedo.go.jp/koubo/CA3_100483.html
- New Energy and Industrial Technology Development Organization (NEDO). 2026a. Application guidelines for the Deep-Tech Startup Support Program and the GX Deep-Tech Startup Program (in Japanese). https://www.nedo.go.jp/content/800058597.pdf
- New Energy and Industrial Technology Development Organization (NEDO). 2026b. Announcement of selected projects, DTSU 8th call and GX 5th call (in Japanese). https://www.nedo.go.jp/koubo/CA3_100497.html
- NEDO Research Evaluation Committee. 2025. End-of-program evaluation report of the R&D-oriented Startup Support Program (in Japanese). https://www.nedo.go.jp/content/800026078.pdf
- U.S. Government Accountability Office. 2024. *Small Business Research Programs: Increased Performance Standards Likely Affect Few Businesses Receiving Multiple Awards* (GAO-24-106398, Highlights). https://www.gao.gov/assets/d24106398high.pdf
- U.S. Small Business Administration. n.d. "Performance Benchmark Requirements." SBIR.gov. https://www.sbir.gov/performance-benchmarks

### Research firms and consultancies

- CFO Dive. 2025. "Firms Fund Skinnier Ventures on Shorter Leashes." https://www.cfodive.com/news/companies-fund-skinnier-ventures-shorter-leashes-report-finds/804042/
- CFO.com. 2023. "Secrets to Launching a New Business Internally." https://www.cfo.com/news/new-business-building-ventures-generative-AI-stage-gated-funding-incumbent-advantage/701543/
- Innovation Research Institute. 2026. "Shinki jigyō ni kansuru jittai chōsa" [Survey on new businesses] (press release, in Japanese). https://www.oricon.co.jp/pressrelease/3023816/
- McKinsey & Company. 2025. "The Way to Win in Corporate Venturing: Serial Building and AI." https://karriere.mckinsey.de/capabilities/business-building/our-insights/the-way-to-win-in-corporate-venturing-serial-building-and-ai
- PwC Consulting. 2025. "Shinki jigyō kaihatsu no torikumi ni kansuru jittai chōsa 2025" [Survey on new-business development initiatives 2025] (in Japanese). https://www.pwc.com/jp/ja/knowledge/thoughtleadership/2025/assets/pdf/new-business-development-survey-2025.pdf

### Practitioners

- Cooper, Robert G. n.d. "Unlocking 'Pipeline Gridlock': Effective Portfolio Management Is the Key." InnovationManagement.se. https://innovationmanagement.se/?p=64618
- Cooper, Robert G. 2025. "Stage-Gate Agentic: The Coming Revolution in the New Product Process." PDMA Knowledge Hub. https://community.pdma.org/knowledgehub/bok/product-innovation-process/stage-gate-agentic-the-coming-revolution-in-the-new-product-process
- Cooper, Robert G. 2026. "AI Powered Stage Gate: Supercharging Your Idea to Launch Process." PDMA Knowledge Hub. https://community.pdma.org/knowledgehub/bok/product-innovation-process/ai-powered-stage-gate-supercharging-your-idea-to-launch-process
- Cooper, Robert G., and Xinjin Zhao. 2026. "New Product Governance in a Time-Compressed World." PDMA Knowledge Hub. https://community.pdma.org/knowledgehub/bok/product-innovation-process/new-product-governance-in-a-time-compressed-world
- CyberAgent. n.d. "Shinki jigyō no sōshutsu" [Creating new businesses] (in Japanese). https://www.cyberagent.co.jp/way/info/detail/id=20357
- Fishman, Adam. 2026. "Adam Fishman on Why Most Innovation Labs Fail at Mozilla." Growth Talent. https://www.growthtalent.org/insights/adam-fishman-on-why-most-innovation-labs-fail-at-mozilla-and-beyond
- Fujita, Susumu. 2025. "'Shinki jigyō o tachiageru dake de wa dame'" ["Just launching new businesses is not enough"] (in Japanese). Bunshun Online. https://bunshun.jp/articles/-/83858?page=2
- ITmedia AI+. 2025. "'Zekushi' 'R25' o unda Rikurūto shinki jigyō seido: sono un'eisha wa, seisei AI o kō tsukau" [How the operator of Recruit's new-business program uses generative AI] (in Japanese). https://www.itmedia.co.jp/aiplus/articles/2503/04/news083.html
- Jones, Michelle. n.d. "Improving Go/Kill Decision Culture." Stage-Gate International. https://www.stage-gate.com/blog/improving-go-kill-decision-culture/
- McGrath, Rita Gunther. 2009. "How to Shut Down a Project Gracefully." strategy+business. https://www.strategy-business.com/article/09214?pg=all
- McGrath, Rita Gunther. 2010. In "Game Changer," by Russ Banham. CFO.com. https://www.cfo.com/news/game-changer/669106/
- Osterwalder, Alexander, and Tendayi Viki. 2021. "Why Killing Ideas Is Key in Innovation." Strategyzer. https://www.strategyzer.com/library/why-killing-ideas-is-key-in-innovation
- Osterwalder, Alexander, and Tendayi Viki. 2023. "How Should Your Innovation Board Make Decisions?" Strategyzer. https://www.strategyzer.com/library/how-should-your-innovation-board-make-decisions
- Ries, Eric. 2017. In "In Eric Ries' New Book, He Tells Companies to Turn Every Unit into a Cash-Strapped 'Startup'," by Connie Loizos. TechCrunch. https://techcrunch.com/2017/10/13/in-eric-ries-new-book-he-tells-companies-to-turn-every-unit-into-a-cash-strapped-startup/
- Sogyo Techo. 2024. "Inkyubētā Ishikawa Akira: shinki jigyō no tachiagekata kōza" [Incubator Akira Ishikawa on launching new businesses: success and exit criteria] (in Japanese). https://sogyotecho.jp/interview-ishikawa-2/
- Stage-Gate International. n.d. "Discovery-to-Launch Process." https://www.stage-gate.com/discovery-to-launch-process/
- Teller, Astro. 2016a. "The Moonshot Factory Operating Manual." X. https://x.company/blog/posts/a-peek-inside-the-moonshot-factory-operating-manual/
- Teller, Astro. 2016b. In "Astro Teller, Captain of Moonshots at X, on the Future of AI, Robots, and Coffeemakers," by Erico Guizzo. IEEE Spectrum. https://spectrum.ieee.org/automaton/robotics/artificial-intelligence/astro-teller-captain-of-moonshots-at-x
- Toyo Keizai Online (produced by Toyo Keizai Brand Studio). 2022. "Rikurūto 'tettai suru jigyō' 'seichō suru jigyō' no sa" [Recruit: the difference between businesses that exit and businesses that grow] (in Japanese). https://toyokeizai.net/articles/-/577974
- Viki, Tendayi. 2020. "To Kill or Persevere: How Do You Score Innovation Projects?" Strategyzer. https://www.strategyzer.com/library/to-kill-or-persevere-how-do-you-score-innovation-projects

### Academic

- Behrens, Judith, and Holger Ernst. 2014. "What Keeps Managers Away from a Losing Course of Action? Go/Stop Decisions in New Product Development." *Journal of Product Innovation Management* 31 (2): 361–374. https://doi.org/10.1111/jpim.12100
- Böttcher, Lucas, and Ronald Klingebiel. 2025. "Organizational Selection of Innovation." *Organization Science* 36 (1): 387–410. https://doi.org/10.1287/orsc.2023.17357
- Boulding, William, Ruskin Morgan, and Richard Staelin. 1997. "Pulling the Plug to Stop the New Product Drain." *Journal of Marketing Research* 34 (1): 164–176. https://doi.org/10.1177/002224379703400114
- Criscuolo, Paola, Linus Dahlander, Thorsten Grohsjean, and Ammon Salter. 2017. "Evaluating Novelty: The Role of Panels in the Selection of R&D Projects." *Academy of Management Journal* 60 (2): 433–460. https://doi.org/10.5465/amj.2014.0861
- Eliëns, Ramon, Katrin Eling, Sarah Gelper, and Fred Langerak. 2018. "Rational Versus Intuitive Gatekeeping: Escalation of Commitment in the Front End of NPD." *Journal of Product Innovation Management* 35 (6): 890–907. https://doi.org/10.1111/jpim.12452
- Ewens, Michael, Ramana Nanda, and Matthew Rhodes-Kropf. 2018. "Cost of Experimentation and the Evolution of Venture Capital." *Journal of Financial Economics* 128 (3): 422–442. https://doi.org/10.1016/j.jfineco.2018.03.001
- Gompers, Paul A. 1995. "Optimal Investment, Monitoring, and the Staging of Venture Capital." *The Journal of Finance* 50 (5): 1461–1489. https://doi.org/10.1111/j.1540-6261.1995.tb05185.x
- Huang, Tori Y., Vangelis Souitaris, and Sigal G. Barsade. 2019. "Which Matters More? Group Fear versus Hope in Entrepreneurial Escalation of Commitment." *Strategic Management Journal* 40 (11): 1852–1881. https://doi.org/10.1002/smj.3051
- Klingebiel, Ronald, and Ron Adner. 2015. "Real Options Logic Revisited: The Performance Effects of Alternative Resource Allocation Regimes." *Academy of Management Journal* 58 (1): 221–241. https://doi.org/10.5465/amj.2012.0703
- Manso, Gustavo. 2011. "Motivating Innovation." *The Journal of Finance* 66 (5): 1823–1860. https://doi.org/10.1111/j.1540-6261.2011.01688.x
- Schmidt, Jeffrey B., and Roger J. Calantone. 1998. "Are Really New Product Development Projects Harder to Shut Down?" *Journal of Product Innovation Management* 15 (2): 111–123. https://doi.org/10.1111/1540-5885.1520111
- Sethi, Rajesh, and Zafar Iqbal. 2008. "Stage-Gate Controls, Learning Failure, and Adverse Effect on Novel New Products." *Journal of Marketing* 72 (1): 118–134. https://doi.org/10.1509/jmkg.72.1.118
- Simester, Duncan, and Juanjuan Zhang. 2010. "Why Are Bad Products So Hard to Kill?" *Management Science* 56 (7): 1161–1179. https://doi.org/10.1287/mnsc.1100.1169
- Walrave, Bob, Sharon Dolmans, Kim E. van Oorschot, Arno L. P. Nuijten, Mark Keil, and Stefan van Hellemond. 2022. "Dysfunctional Agile–Stage-Gate Hybrid Development: Keeping Up Appearances." *International Journal of Innovation and Technology Management* 19 (3): 2240004. https://doi.org/10.1142/s0219877022400041