Watch the full episode on YouTube: https://youtu.be/__Zk6ULvIJM
Assumptions mapping is a team exercise that makes the risks behind a business idea explicit, then ranks them by how important they are and how much evidence supports them. David J Bland, founder of Precoil and lead author of Testing Business Ideas with Alexander Osterwalder, used it to help DuPont and its semiconductor spin-off Qnity build an experimentation culture that won the 2026 IRI Innovation Excellence Award for Outstanding Innovative Culture. In this episode of The Innovation Show, he walks host Aidan McCullen, the Dublin-based innovation and change keynote speaker, through the case study and the habits that made it stick.
Key takeaways
- DuPont ran 35 real growth opportunities through Precoil’s Extract, Map, Test method, saving more than $12 million by stopping weak projects and generating about $24 million from the ideas that passed.
- Assumptions mapping plots every hypothesis on two axes, importance and evidence. The important assumptions with the least evidence get tested first.
- A leader should never facilitate their own team’s session, because people stay silent about the risks in the boss’s pet project.
- Most failures at DuPont came from flawed commercial and customer assumptions, according to the published case study. The technology was rarely the problem.
- Culture reflects the work people do every day. DuPont made killing projects safe with an annual Dead Projects Day.
What is assumptions mapping?
Assumptions mapping turns the hidden bets in a business idea into written hypotheses that a team can prioritise and test. Testing Business Ideas (Wiley, 2019, pp. 36 to 39) sorts those hypotheses into three types. Desirability asks whether customers want the idea. Viability asks whether it can make more money than it costs, and feasibility asks whether the organisation can build and deliver it. The book credits the exercise as adapted from Jeff Gothelf and Josh Seiden’s Lean UX.
Each hypothesis then goes onto a two-by-two. The vertical axis runs from unimportant to important, where an important hypothesis is one that would sink the idea if it proved false. The horizontal axis runs from having evidence to having none. The top-right quadrant, important with no evidence, defines the team’s next experiments.
The labels took years to settle. David tells Aidan the axes began as known and unknown, moved to certain and uncertain, and landed on evidence after a whiteboard session with Osterwalder at Stanford. He also dropped dot voting early on, because every dot ended up following the most senior person in the room.
How to run an assumptions mapping session
- Pick one live idea and the right people. The book calls for a cross-functional core team plus a supporting team from functions such as legal, safety, compliance and marketing, who hold knowledge the core team lacks.
- Bring in a neutral facilitator. David is firm that the executive backing an idea should never run the session. Borrow someone from another business unit or use an outside coach.
- Write one hypothesis per sticky note. Use the format “We believe that…” and keep each one short and specific. Use a different colour for desirability, viability and feasibility.
- Place every note on the map. Position each one by importance and by the strength of the evidence you hold today. Challenge anything in the top-left quadrant that claims evidence the team cannot show.
- Design experiments for the top right. Start with cheap, fast tests and agree in advance what result would make you commit, correct course or cut the idea.
The finished map also exposes blind spots. If nearly every risk on an early idea sits under feasibility, David reads that as a warning that the team has skipped the customer and gone straight to the build.
How strong is your evidence?
An assumptions map is only as honest as the evidence behind it. Testing Business Ideas (p. 52) offers four checks for judging whether evidence is weak or strong.
| Weaker evidence | Stronger evidence |
|---|---|
| Opinions and beliefs, such as “I would buy that” | Facts and past events, such as “Last month I spent…” |
| What people say in interviews or surveys | What people actually do |
| Lab settings, where people know they are being tested | Real-world settings |
| Small commitments, such as an email sign-up | Large commitments, such as a pre-order or putting a professional reputation on the line |
Why good ideas and bad ideas look alike early on
“A good idea and a bad idea look very similar early on,” David says. He opens his keynotes with his own misjudgements to prove it. Someone once pitched him a kitty litter brand, and he wondered how it could stand out in a crowded aisle. That company later sold for a billion dollars.
His second lesson ran the other way. Teams at a large travel company in San Francisco were running experiments everywhere, until the VP of product asked David why they were testing half of them. He had no answer. From then on, he tied every experiment back to a risk the executive team already cared about, and assumptions mapping became the link between the two.
Why business plans fail new ideas
The synopsis of Testing Business Ideas warns that ideas which look irresistible in a spreadsheet often turn out to be “a hallucination.” Aidan calls this out as his favourite line in the book. The Invincible Company (Wiley, 2020) goes further. Osterwalder and his co-authors argue that demanding a detailed business plan from project teams raises the odds of executing an idea nobody has tested.
David’s fix is to understand the business model before writing any plan. Ask an executive team to describe their business model and you will often hear five or six different answers. A five-year plan built on that confusion is, in David’s words, “creative fiction in Excel.” DuPont felt this directly. The company can build almost anything, yet a manufacturing customer might discover late in the process that switching a component would cost millions on their own production line.
How DuPont built an evidence-based innovation culture
DuPont’s leadership had pushed for experimentation since 2016. After six years of mixed results, they brought David in during 2022. He started with a few teams working on real opportunities, each supported by a coach, running cycles of roughly 12 weeks. Each cycle began with assumptions mapping on the team’s own opportunity, followed by experiments on the riskiest hypotheses. By the second cohort, DuPont’s chief of staff described the flow as a “steady state,” and the company scaled it across 35 teams.
Four elements held it together, as Aidan summarises from the case study.
- A customised decision system aligned with DuPont’s strategy and governance.
- Learning plans as the core artefact for evidence-based decisions, so every team looked through the same lens.
- Coaching as the primary way to scale.
- Executive sponsorship through a two-in-a-box model pairing business and technical leaders.
Leaders also had to learn new questions. ROI questions make little sense for pre-revenue work, so DuPont trained leaders to ask about risk and evidence. Desirability, viability and feasibility language now runs through the stage gates at both DuPont and Qnity.
Dead Projects Day and the culture that makes killing ideas safe
Killing projects damages morale unless people feel safe. Alexa Dembek, DuPont’s Chief Technology and Sustainability Officer at the time, championed Dead Projects Day, a Halloween event where leaders dressed in costume and celebrated the projects the company had stopped. It sounds silly, David admits, until you see what it signals. Honest testing will not end anyone’s career. The ritual only works with a healthy volume of ideas, since a team with three ideas will never kill one.
The Invincible Company (pp. 311 to 312) makes the same point through its Culture Map. It groups the enablers and blockers of innovation culture under leadership support, organisational design and innovation practice, and lists “nobody gets fired for experimenting” among the behaviours of a healthy culture. David adds a warning about leadership whiplash, where each new CEO swings the company from Six Sigma efficiency to an innovation drive and back again. Long-serving staff learn to wait each wave out.
Bring sales and legal in early
Salespeople sit closest to customers, so new revenue often starts with them, provided their incentives reward discovery as well as closing deals. Legal matters just as much. Explain an experiment’s size and “blast radius” at the start, perhaps a test with only 100 people, and legal is far more likely to say yes. Arrive at the end and expect a no. The simplest fix is to invite both functions to the assumptions mapping session itself, which is where Testing Business Ideas already places legal and compliance.
Using AI as a sparring partner
David uses AI to challenge his thinking and avoids handing it the whole job of generating risks and tests. He argues that some intentional friction protects a team’s critical thinking, especially while organisations are still learning how to adopt AI well.
Bring these ideas into your organisation
Aidan McCullen is an innovation and change keynote speaker based in Dublin, working with leadership teams in Ireland and internationally. He is the 2025 Thinkers50 Innovation Award recipient, author of Undisruptable: A Mindset of Permanent Reinvention (Wiley) and host of more than 650 episodes of The Innovation Show. His keynotes and workshops on AI, disruption, innovation and change draw on conversations with thinkers such as David Bland, and help organisations reinvent before circumstances force them to.
If your leadership team is wrestling with the questions in this episode, such as when to kill a project or how to build a culture that experiments, find out more about booking Aidan as a keynote speaker.
Further reading
- DuPont case study on Precoil
- DuPont’s announcement of the 2026 IRI Innovation Excellence Award
- Testing Business Ideas by David J Bland and Alexander Osterwalder
- How I Tested That, David’s podcast
Subscribe to the Thursday Thought on Substack for a chance to win a copy of Testing Business Ideas, with thanks to Kyndryl.
About the Guest
David J Bland is the founder of Precoil and lead author of Testing Business Ideas, co-written with Alexander Osterwalder. Since going independent in 2011 he has helped companies including GE, Toyota, Adobe and DuPont test high-uncertainty ideas before they commit major capital. He also hosts the How I Tested That podcast. Find him at davidjbland.com and on LinkedIn.
About the Host
Aidan McCullen is the 2025 Thinkers50 Innovation Award recipient, keynote speaker on AI, disruption, innovation and change, host of The Innovation Show, and author of Undisruptable: A Mindset of Permanent Reinvention (Wiley). Learn more about Aidan.
About The Innovation Show
The Innovation Show is the Thinkers50-recognised podcast where square pegs find their place in a world of round holes. This series is brought to you by Kyndryl, who run and reimagine the technology systems that drive advantage for the world’s leading businesses. Learn more at the Kyndryl Institute.
Listen or watch now: Apple Podcasts | Spotify | YouTube
Frequently asked questions
What is assumptions mapping?
Assumptions mapping is a team exercise from Testing Business Ideas by David J Bland and Alexander Osterwalder. Teams write the desirability, viability and feasibility hypotheses behind an idea, then plot each one by importance and by evidence. The important hypotheses with the least evidence get tested first, because they could sink the idea if false.
What counts as strong evidence when testing business ideas?
Testing Business Ideas treats facts as stronger than opinions and what people do as stronger than what they say. Real-world settings beat lab settings, and large commitments such as pre-orders beat small ones such as email sign-ups. Strong evidence makes pivot, persevere or kill decisions far less risky than opinion-based bets.
What results did DuPont get from evidence-based innovation?
Across 35 real growth opportunities, DuPont saved more than $12 million by stopping weak projects early and generated about $24 million from ideas that passed their tests. DuPont and its spin-off Qnity were co-winners of the 2026 IRI Innovation Excellence Award for Outstanding Innovative Culture.
What is Dead Projects Day at DuPont?
Dead Projects Day is an annual Halloween event, championed by Alexa Dembek, where DuPont’s innovation community celebrates the projects it killed. Leaders dress in costume and explain why each project stopped. The ritual makes it safe to end weak ideas early and stops failed projects being hidden or dressed up as success.
Who is David Bland?
David J Bland is the founder of Precoil and lead author of Testing Business Ideas with Alexander Osterwalder. He has advised companies including GE, Toyota, Adobe and DuPont on testing high-uncertainty ideas, and hosts the How I Tested That podcast.
Who is Aidan McCullen?
Aidan McCullen is a Dublin-based keynote speaker on innovation, change and reinvention, and the 2025 Thinkers50 Innovation Award recipient. He hosts The Innovation Show, a Thinkers50-recognised podcast with more than 650 episodes, wrote Undisruptable (Wiley), lectures at Trinity Business School and is a former professional rugby international.