TL;DR
- I use a six-field template in ideation workshops. Field five asks the team why their own idea might fail, while they are still in the room.
- Given the same template and the same time, teams produce wildly different answers in field five. The spread is not explained by the quality of the idea.
- Two things drive it. Familiarity, which turns fluency into a blind spot. And composition, which limits what the room is capable of seeing.
- Teams that list only technical failure modes are describing a build. Teams that name competitors, pricing and sales gaps are describing a business.
- Read field five as a readout of who was in the room rather than a judgement of the team. The fix is a seating plan.
- The strongest kill criteria are about people, not product. Agree in advance what happens to careers and compensation if the thing stops.
Here is a template I use in workshops. It has six fields.
- What is the product?
- Who is it for?
- What problem does it solve?
- Illustrate how it works.
- Why might it fail?
- What resources do we need?
The fifth field is the one that matters.
Every ideation format I have used produces more ideas. That is what ideation is supposed to do. Elaboration is how ideation works. You add features, refine personas, sketch interfaces, map journeys. The question of whether the thing should exist gets deferred to a later gate, usually owned by someone who was not in the room, applied weeks after the team has already become attached to what they built. By then the question is political rather than analytical.
The six questions put the failure question inside the exercise. It gets answered by the people who just pitched the idea, while they are still standing there, as a condition of finishing.
This template is the second step in a two part method. The first is a Problem-Solution Fit Canvas that forces teams to define the problem, who feels it, what it costs, and what the current workaround is before they are allowed to propose a solution. The six-field template comes right after. It is where a team takes their validated problem and pitches a product. Field five is where it gets difficult.
Where this sits relative to Klein
Gary Klein’s premortem is close to this territory. It asks a team to assume the project has already failed and work backward to identify why. In his 2007 HBR piece he cites 1989 research by Mitchell, Russo and Pennington finding that prospective hindsight, imagining that an event has already occurred, increases the ability to correctly identify reasons for future outcomes by 30 percent.
Klein’s method is powerful. It also runs on a project that has already been selected and resourced. Its purpose is to overcome reluctance to speak up after commitment has been made.
I am not claiming to improve on Gary Klein. I am asking the same question earlier, before there is a project to protect. I want it embedded in ideation so it cannot be skipped. Fluff added after the fact creates too rosy a picture. People fall in love with the idea, and killing the idea starts to feel lot like a crime. The better move is to ask why it might fail, an awkward and uncomfortable question, from the same people who just said it was a good idea.
The spread
Given the same template and the same amount of time, different teams produce different failure modes.
The most extreme case I have seen was one team producing six distinct failure modes while their three counterparts produced between one and three. Same company. Same projects. Very similar experience levels. That last part matters, and I will come to it later.
I have watched this asymmetry repeat across workshops and client engagements over several years. Same template. Same facilitation. Same time pressure. Wildly different answers in field five.
For the first few sessions I assumed the quality of the idea explained the difference. I assumed that a stronger idea would naturally surface more risks because there was more to interrogate. That turned out to be wrong.
If anything, it’s the other way around. In the example above, the team that was new to the problem space carefully listed six failure modes because they were not yet sure they understood it. The three teams working on something they knew deeply produced between one and three, and stopped. They were confident. They had lived with the idea long enough that the risks had become invisible.
Did those three teams only have one to three failure modes in them? No. The rest surfaced later, once the work started, which is the expensive place to find them.
Confidence, then composition
The first thing the spread reveals is about familiarity. Teams that have been thinking about a product for weeks or months develop a fluency that looks like command but functions as a blind spot. They know their idea so well that the failure question feels impolite. The team that is new to the problem has no such attachment. They are cautious because they have no reason not to be, and that caution produces better answers in field five.
That is worth noticing and it’s not the whole story.
The deeper pattern is about who is in the room. The lopsided version of field five shows up - for example - when a group is heavily technical with nobody representing the business side, or when business voices are present but get talked over. I have seen this lopsided version repeat consistently enough that I stopped treating it as coincidence.
Sticking to my example, a group of engineers will list integration challenges, messy data and missing process standards. They are not being complacent. They are being complete within the limits of what that group can see. Nobody in that group was ever going to raise a competitor’s pricing strategy or a gap in sales capability, because nobody there thinks about those things for a living. It works the other way too. A room full of sales people will not appreciate the technical risk, either.
Misplaced optimism and lopsided composition can co-exist. A team that is deeply familiar with a problem and composed entirely of one discipline will produce the narrowest answer in field five.
Confidence removes the caution. Composition removes the peripheral vision.
The answer is not to tell teams to think harder. The answer is to group the room differently. Mix technical and business perspectives deliberately. Manage seniority so one or two voices cannot dominate. And be especially alert when a team tells you they already know this space well.
There is research running adjacent to this. De Freitas, Israeli, Nave, Timoshenko and Toubia argue that the constraints on innovation are cognitive and social in origin, rooted in how people generate ideas, evaluate novelty, and communicate through social systems, rather than using the tools or using the information available. Their paper is about what generative AI does to those constraints. What I am interested in is the premise underneath it. If the bottleneck is cognitive and social, then who is in the room is not a soft variable. It is the variable. I arrived there from running workshops rather than from studying the literature, which is useful, because convergence from two directions suggests the pattern is common.
I also wish these people had published their article four years earlier. It would have saved me a lot of head scratching.
The tell
Here is the practical version, and you can check it in your next session.
Teams that list only technical failure modes are describing a build. Integration risks, data quality, missing APIs, unclear specs. are all real problems. They are also problems that live entirely inside the team’s existing capability.
Teams that name competitors, their own sales gaps, pricing risk, and what customers will actually expect the product to be are describing a business. They are thinking about the thing as it will exist in the world.
Both teams are not wrong. Reading these outcomes as a verdict on the team is unfair. Nobody chooses which lens they carry into a room.
Who decides
Then there’s the most important point. This method does not make a decision for anyone.
What it does is surface trade-offs and flag opinion as opinion rather than smuggling it in as finding. The decision stays with the people who carry the consequences. In this session, it was the four teams who had to go back to their regular work once the workshop finished.
The strongest kill criteria, it turns out, are usually about people rather than product. Agreeing in advance what happens to someone’s career and compensation if an initiative is stopped is what makes stopping possible later. It is a sobering conversation. You are putting money where your mouth is. I have watched people who were bullish about their ideas wither at that challenge, and people with less conviction grow more confident the more they understood the negatives. Through that discomfort, agreement to proceed or stop arises.
Without that agreement, it is very hard to pull the trigger regardless of what field five says.