When and how to stress test your grant proposal
The 30/70 Rule
This article was first published on LinkedIn - the Impact Newsletter
When you are developing a grant proposal, it is surprisingly easy to lose sight of the big picture. This is especially true when multiple collaborators are involved. Each contributor frames their section around their own priorities, what they want to bring to the project, what they feel confident defending. And because these are your collaborators, there is a social friction that inhibits honest criticism.
It is also easy to slide into groupthink. Research on long-lived teams confirms what you probably already sense: high cohesion produces fast consensus, but often at the cost of critical thinking [7]. Groups tend to suppress opinions that conflict with the emerging view — what The Art of Co-Creation calls tunnel vision [6]. The longer the collaboration, the more entrenched the norms, and the harder it becomes for anyone to say the uncomfortable thing.
An external Red Team review addresses both problems at once. A Red Team is an independent group that challenges strategy, plans, or assumptions by adopting an adversarial or outsider perspective. Unlike a standard peer review, which often focuses on incremental improvements or methodological critiques, a Red Team’s goal is to identify blind spots, expose hidden assumptions, and stress-test the “big picture” before a project is finalized. By operating outside the internal hierarchy and social constraints of a team, a Red Team provide the objective friction necessary to overcome groupthink and ensure a proposal is both strategically sound and sufficiently ambitious. A good Red Team will challenge your assumptions rather than confirm them.
A Red Team review is about taking a strategic approach to proposal development. The most efficient way to do any kind of work is to spend more time on strategy than on production. For a funding proposal that translates into spending 30% of your time on content and 70% on strategy.
Why Your Network is the Wrong Place to Get Feedback on Your Grant
The instinct is to ask experts in your network. But this recreates a peer review process and peer review has a well-documented dampening effect on ambitious science. Studies have found that grant reviewers are systematically biased against novel and high-risk proposals, favouring incremental work with a predictable chance of success [3]. Ioannidis and Thombs, writing in EMBO Reports, describe how low success rates exacerbate “conservative, short-term thinking in applicants, reviewers, and funders” [2]. Braben is blunter in Scientific Freedom: “Hence the folly of using peer review to assess virtually all proposals” [4]. Expert reviewers also carry social constraints — they worry about how their feedback will be received, particularly when the applicant is a colleague or a colleague’s close collaborator.
A short, simple rule of thumb is that 30% of the input for your proposal should come from outside your network and area of expertise.
Should you be ambitious or incremental?
A story from my own experience in the early years of Horizon 2020 makes this concrete. A peer reviewer at a large university told me he had reviewed eleven proposals in a Health call where success rates were extremely low — in some calls during this period, fewer than 5% of submissions were funded [11]. Of those eleven, the proposal he was most confident would not pass was the only one that got funded. His objection? It was too ambitious. He was right once the project was running, I looked at their website and agreed. It was overambitious. And yet it got funded. Had they taken that peer review advice, they would have trimmed precisely what made the proposal compelling.
The point is that there is a balance between ambitious and incremental research. It depends upon the funding programme and the call topic. Some call for clearly ambitious proposals. If you fall victim to the peer review bias against ambitious proposals you may not stand out against other proposals.
When to stress test your funding proposal.
The reason that seemingly overambitious consortium did not listen to the feedback was probably timing. Unwinding a proposal in the final week before submission is genuinely difficult. This is one of the most underappreciated aspects of Red Team reviews: timing matters as much as content. In my experience, the optimal window is when the proposal is roughly 70% complete, enough structure to critique seriously, enough flexibility to act on what you hear. It is even better if you can start with is a concept-focused review before detailed writing begins, at 30% completion, and then again at 70%.
Research on team coaching provides indirect support for this kind of staged approach. Strategy interventions at the midpoint of a project consistently outperform those delivered too early or too late [1]. The pre-mortem concept from reinforces the same logic. A structured critique works best before the outcome is locked, not after [10].
What distinguishes a genuine Red Team from a peer review is not just independence, it is cross-disciplinary range. Domain experts are reliably blind to the assumptions of their own field. This is not a character flaw; it is a cognitive one. Fisher and colleagues showed in Cognitive Science that deep expertise leads people to systematically overestimate how much of their knowledge is shared by others, the Curse of Expertise [5]. Assumptions that feel self-evident to a specialist are often not self-evident at all; they are the product of years of field-specific immersion. A reviewer from another discipline will simply ask the question that no one inside the field thinks to ask [13]. I have watched this happen many times: a question that seems naive forces a room of specialists to articulate something they had never made explicit, and a new idea forms, either because the assumptions turn out not to be universal, or because someone from an adjacent field recognises the same underlying structure as a problem they have already solved.
Expert peer reviews also rarely provide strategic perspective: where this proposal sits in the wider research and innovation landscape, whether the vision is genuinely differentiated, whether the ambition is calibrated to what funders are actually awarding. A good Red Team covers all of this and adds value at any stage of proposal development.
AI vs. Human Review: Why Your Grant Needs Both to Win
Could you use AI instead? Yes, and an AI review can be technically useful. But for collaborative proposals it misses something essential: the generative power of human interaction. The strongest proposal concepts do not emerge from a document with tracked comments. They emerge from critique, followed by a conversation in which the consortium works through the implications together. A 2025 Springer Nature survey found that researchers overwhelmingly view AI as an assistant rather than an author, with human expertise remaining central to framing questions and generating insight [8]. A recent framework for Generative Collective Intelligence makes the structural case: the complementarity of human intuition and AI’s information-processing capacity is exactly where the value lies, neither alone gets there [9]. Having a human-in-the-loop Red Team Review is not a workaround for AI’s limitations. It is a deliberate design choice.
How to run the ideal human-in-the-loop Red Team review
Find a reviewer with cross-disciplinary knowledge, expertise in funding proposal writing, and the ability to facilitate productive dialogue.
Early in development, 30% completed, have that reviewer challenge the concept through an interactive dialogue session, before the structure hardens.
Engage 1–3 expert reviewers with no prior involvement and plan a structured review meeting at approximately 70% completion.
Run a well-formed AI-based review in parallel.
Use the AI review as well as the early stage review to develop questions for the expert reviewers.
Run a facilitated online or in person meeting where the facilitator drives creative dialogue to develop solutions to the most critical weaknesses
Work off the resulting recommendations to improve your proposal.
Strengthening your positioning as an innovation leader
Whether or not your proposal gets funded, developing a consortium project is a great opportunity to strengthen your position as an innovation leader. A effective Red Team review helps you develop the best possible proposal. If you are planning to develop a proposal for funding I would welcome the opportunity to discuss how you can implement a Red Team review. Schedule a call.
References
Hackman, J.R. (2011). Collaborative Intelligence: Using Teams to Solve Hard Problems. Berrett-Koehler.
Ioannidis, J.P.A. & Thombs, B.D. (2019). The troubles with peer review for allocating research funding. EMBO Reports, 20(12). https://www.embopress.org/doi/full/10.15252/embr.201949472
Science/AAAS — Funding agency’s reviewers were biased against scientists with novel ideas. https://www.science.org/content/article/funding-agency-s-reviewers-were-biased-against-scientists-novel-ideas
Braben, D.W. Scientific Freedom. https://readwise.io/open/846396593
Fisher, M. et al. (2016). The Curse of Expertise: When More Knowledge Leads to Miscalibrated Explanatory Insight. Cognitive Science, 40(5), 1251–1269. https://pubmed.ncbi.nlm.nih.gov/26369299/
Rill, B.R. & Hämäläinen, M.M. The Art of Co-Creation. https://readwise.io/open/184442751
ProjectManagement.com — Avoiding groupthink on long-lived teams. https://www.projectmanagement.com/blog-post/38332/avoiding-groupthink-on-long-lived-teams
Evans, I. (2025). How AI is transforming researchers’ daily work. Springer Nature. https://spkl.io/6040ASQpp
ArXiv 2505.19167 (2026). Amplifying Human Creativity and Problem Solving with AI Through Generative Collective Intelligence. https://arxiv.org/abs/2505.19167
Make Smart Choices / HBR — pre-mortem approach. https://readwise.io/open/575042164
Horizonbook.eu — What are the success rates for proposals in Horizon 2020? https://www.horizonbook.eu/news/what-are-the-success-rates-for-proposals-in-horizon-2020/
Garrette, B., Phelps, C. & Sibony, O. Cracked It! https://readwise.io/open/367410822
Wedell-Wedellsborg, T. What’s Your Problem? https://readwise.io/open/970486361