We have an idea. We believe in it. We've mulled it over, talked with the team, and maybe even built something. Yet when we bring it to life, users don't react the way we expected. And sometimes it's not because the product is bad, but because we've been building on assumptions we never tested.
When a product team gets together around the Opportunity Solution Tree (Continuous Discovery Habits 2021), there is usually an uncomfortable moment. Someone has mapped opportunities, someone has proposed solutions, and the tree starts to take shape. Everything seems reasonable. Everything seems connected.
But beneath every branch there is an invisible layer: the assumptions. The things the team takes for granted without having verified them: users experience the product the way the team believes they do, the proposed solution solves something people actually want solved, they are willing to change their behavior to use it...
Torres calls this the work of assumption tests: before building a full solution, identify the assumptions that support it and look for the fastest and cheapest way to find out whether they hold. It seems reasonable. And yet most teams don't do it, or do it poorly.
Why? Because identifying assumptions requires a certain amount of intellectual honesty that runs against two very human forces: enthusiasm for an idea and organizational pressure to deliver. When you are convinced something is going to work, stopping to ask yourself "what would have to be true for this to work?" doesn't come naturally. It feels uncomfortable.
What exactly is a hidden assumption
An assumption is any implicit statement that underlies a decision. In product management, assumptions usually fall into four dimensions under the acronym DUDE (David Bland and Alex Osterwalder, 2020):
- Desirability (does anyone want it?).
- Usability (can they use it?).
- Feasibility (can we build it?).
- Viability (does it make sense for the business?).
But not all assumptions are equally visible. Some are explicit ("we believe users prefer paying monthly over annually.") and easy to identify and easy to formulate as hypotheses. Others are so structural that the team doesn't even see them as assumptions: "users will read the instructions before using the feature." Or worse: "the problem we are solving is the real problem."
This last level, assumptions about the problem itself, is the most dangerous and the most frequently ignored. It is the difference between teams that produce output and teams that pursue outcomes: the former assume the problem is well defined and focus on the solution; the latter periodically question whether they are tackling the right problem (Marty Cagan, 2018).
Teresa Torres goes a step further with her concept of continuous discovery: the problem is never completely defined. Our understanding of the problem evolves as we learn from users. That's why assumptions about the problem are not something validated once at the beginning, but something reviewed regularly throughout the product lifecycle.
Opportunity Solution Tree as an assumption map
If we work with Torres's OST (opportunity solution tree), we have a structure that can help us organize assumptions systematically. Each node of the tree carries its own implicit assumptions.
- In the opportunity node (the need, the pain point, the user's desire) there are assumptions about whether that opportunity exists, how frequent or intense it is, and whether the segment experiencing it is relevant enough for the business. A poorly defined opportunity (or one assumed without sufficient evidence) contaminates everything that comes after.
- In the solution node assumptions multiply. Do we believe this solution addresses the opportunity? What behavior do we expect it to trigger? What context are we taking for granted? What do we know about how current users solve that problem without our solution?
- And in the experiments, which are the OST's answer to those questions, there are also methodological assumptions: the experiment measures what we want to measure, the recruited segment is representative, the signal we get is strong enough to make a decision...
Using the OST doesn't eliminate assumptions. It gives them a location: what can be located can be challenged.
How hidden assumptions surface
There is no single method for bringing to the surface what the team takes for granted. But there are some practices that, used regularly, help.
The question "what would have to be true?"
It is perhaps the most powerful and most underrated. Faced with any proposed solution, the team asks: what would have to be true about the user, the market, the technology, and the business for this to work? Each answer is an assumption. The technique comes from the field of strategy, and Torres adapts it directly to the context of product discovery.
The pre-mortem
It is another classic tool gaining renewed prominence in agile environments. Gary Klein described this exercise in Sources of Power (1998): imagine the project has failed and work backwards to identify what could have gone wrong. In the context of product discovery, it forces the team to name the fears and doubts that tend to linger in the collective subconscious. "If this fails, it's probably because we assumed users would change their current habit, and in reality they don't have enough motivation to do it."
Explicit assumption mapping as a team activity
It is usually used in a short workshop before prioritizing experiments. It consists of writing down all the assumptions the team identifies about a solution, without judging them. The rule is simple: if anyone on the team takes it for granted, it gets written down. Then comes prioritization.
Journey map
It can be useful as a tool for revealing assumptions about behavior. When the team maps the user's journey step by step, implicit assumptions emerge about what the person does before reaching our solution, during use, and afterward. Those invisible transitions are fertile ground for dangerous assumptions.
Prioritizing: not all assumptions deserve the same treatment
Once the assumptions are on the table, the next step is to prioritize them. Not because the others don't matter, but because resources (time, energy, money) are always limited.
The most widely used matrix in this context crosses two dimensions: importance (how much does the idea's success depend on this assumption being correct?) and uncertainty (how much do we already know about whether it is valid or not?).
- Assumptions that are critical to the idea's success and on which we have little evidence are the ones to test first.
- Those that are critical but on which we already have solid evidence can be treated as premises (although still revisable).
- Those of low importance, regardless of their uncertainty, can be left for later or simply accepted.
Bland and Osterwalder call this the "high-risk zone": critical and uncertain assumptions are the ones that can sink the idea. Testing them first reduces the risk of investing in building something the market won't validate.
However, there is a trap worth naming: teams tend to prioritize the assumptions they can test easily, not the ones they should test first. The availability bias leads us to overvalue what is within reach and avoid what is difficult or uncomfortable (Kahneman 2011). Testing whether the solution is usable is easier than testing whether the problem we think we've identified is really the problem. But the second assumption is usually more critical.
Testing assumptions: from experiment to evidence
Testing an assumption is not synonymous with running a survey or launching a prototype. It means finding the fastest and cheapest way to gather enough evidence to decide whether to continue, pivot, or abandon a line of work.
Torres distinguishes between two types of experiments depending on the nature of the assumption:
- Attitude experiments measure what people say (interviews, surveys, verbally evaluated prototypes).
- Behavior experiments measure what people do (A/B tests, smoke tests, fake-demand experiments, usage data analysis). These are usually more costly to run but provide more reliable signals, because the gap between what someone says they would do and what they actually do can be enormous.
This distinction is not trivial. When a team shows a high-fidelity prototype and users say "this looks really interesting," the team tends to read that as validation. But the relevant question is not whether the user finds the idea interesting, but whether they would be willing to change their current behavior to adopt it. These are very different questions.
Jeff Gothelf and Josh Seiden (2021) propose a hypothesis structure that helps keep the focus:
"We believe that [output] for [user] will result in [outcome]. We'll know we're successful when we see [measurable signal]."
This formulation forces the experiment to be connected with the evidence it will seek, avoiding the common mistake of testing without knowing exactly what signal would be enough to make a decision.
A concrete example: suppose the team believes users of a task management app need an automatic reminders feature. The hidden assumption is that users don't create reminders manually because it's too tedious, not because they aren't interested in reminders.
- An attitude experiment ("would you use automatic reminders?") will probably get a massive "yes."
- A behavior experiment (offering the feature to a subset of users and measuring how many activate it and how often they use it) will give much more valuable information about real demand.
Organizational risk: when context makes assumption work harder
The work of identifying and testing assumptions does not happen in a vacuum. It happens in organizations with deadlines, budgets, hierarchies, and cultures that don't always reward doubt and questioning.
In many contexts, acknowledging that an idea has unvalidated assumptions can be interpreted as a lack of confidence or as an obstacle to progress. "We already have enough to go on, we need to move forward." This pressure is not irrational from the point of view of those exerting it: there is irreducible uncertainty in any product decision, and at some point you have to act. The problem arises when that pressure leads to skipping critical assumptions because they are uncomfortable or because testing them would require slowing down production.
Amy Edmondson (2018) extensively documents how psychological safety directly affects teams' capacity to learn. A team that doesn't feel safe saying "I think we're assuming things we haven't verified" is unlikely to do serious work mapping assumptions. The method matters, but the environment in which it is applied matters just as much or more.
This connects to the fact that not all organizations are in a position to do quality discovery, not because they lack the method, but because they haven't created the conditions for questioning to be welcome. The maturity diagnosis, which tools like AgiLevel can help make more structured, can be the first step in understanding what is missing before the method.
Cadence: assumptions as a continuous practice, not a phase
One of the most common mistakes when teams discover the work of assumptions is treating it as a phase. This logic is understandable, but it contradicts the central principle of continuous discovery that Torres proposes.
Assumptions change. The market changes. What users need in the first month may not be the same as what they need six months later. And solutions, as they are developed, generate new assumptions that didn't exist at the beginning.
Torres's proposal to hold weekly interviews with users is not arbitrary: it is the mechanism that allows the team to continuously update its understanding of the problem space and, therefore, its assumptions. Without that cadence, knowledge becomes obsolete and assumptions fossilize.
Integrating assumption work into the team's routine (in refinement sessions, reviews, and learning-oriented retrospectives) is more effective than treating it as a special workshop held every time something new begins.
Identifying, prioritizing, and testing assumptions is not a clean process. It requires honesty about what we don't know, willingness to ask uncomfortable questions, and an organizational environment that makes this work possible without the team paying a political cost for doing it.
The good news is that, with practice, it becomes more natural. Teams that get used to questioning assumptions make better decisions and, above all, learn faster from their mistakes. Not because they make fewer, but because they make them earlier and at lower cost.
The question we'd like to leave open is this: in your team or in the teams you work with, how much of the discovery work is spent validating ideas that were already taken as good, and how much is spent genuinely questioning whether the opportunities being pursued are the right ones?
If you have experience with the Opportunity Solution Tree or other ways of making assumptions visible in your team, we'd love for you to share it.
If you want to go deeper into Torres's approach from the fundamentals, you can start with the notes on Continuous Discovery Habits available in Open Knowledge.
Bibliography
Bland, D. J. and Osterwalder, A. (2020). Testing Business Ideas: A Field Guide for Rapid Experimentation. Wiley.
Cagan, M. (2018). Inspired: How to Create Tech Products Customers Love (2nd ed.). Wiley.
Edmondson, A. C. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.
Gothelf, J. and Seiden, J. (2021). Lean UX: Creating Great Products with Agile Teams (3rd ed.). O'Reilly Media.
Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
Klein, G. (1998). Sources of Power: How People Make Decisions. MIT Press.
Martin, R. and Lafley, A. G. (2013). Playing to Win: How Strategy Really Works. Harvard Business Review Press.
Perri, M. (2019). Escaping the Build Trap: How Effective Product Management Creates Real Value. O'Reilly Media. ISBN: 978-1491973790
Seiden, J. (2019). Outcomes Over Output: Why Customer Behavior Is the Key Metric for Business Success. Sense and Respond Press.
Torres, T. (2021). Continuous Discovery Habits: Discover Products That Create Customer Value and Business Value. Product Talk LLC.
Comments (0)
Log in to your Scrum Manager area to comment