Three-Solution Assumption Testing
Compare three ideas by testing their riskiest assumptions in small, fast cycles.
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 7
- Confidence
- 97%
Three-Solution Assumption Testing keeps a team from falling in love with its first answer. For an important opportunity, generate three distinct solution ideas and break each one into the conditions that must be true for it to work. Prioritize those assumptions by risk, then design tests small enough to evaluate one assumption rather than the entire solution. The team can run several tests across all three ideas in a week and compare the emerging strengths and weaknesses. This differs from a project-sized experiment that takes weeks and returns only a verdict on one bundled concept. The goal is not to certify an idea before delivery; assumption testing is the beginning of delivery. By keeping multiple options alive and producing fast evidence about their foundations, the team can improve, combine, or discard solutions before making an expensive commitment.
Origin
Torres teaches teams to work with three ideas at once because options enable comparison, while small assumption tests make that parallel exploration practical within continuous discovery.
Core principles
- 01Every solution is a bet whether or not the team performs discovery.
- 02Options improve judgment by making comparison possible.
- 03Test the foundations of an idea before testing the whole idea.
- 04Prioritize assumptions before spending time on experiments.
- 05Small tests make parallel discovery sustainable.
How to run it
- 1
Select an important opportunity
Choose a customer opportunity where a robust solution matters, such as a core experience or potential differentiator. State the outcome the solution should support.
Pro tip Spend less discovery effort on routine flows that are working and do not represent meaningful risk.
Watch out Applying maximum discovery to every small change wastes the team's limited attention.
- 2
Create three options
Generate three materially different ways to address the same opportunity. Keep all three available long enough to compare their trade-offs.
Pro tip Increase the number or variety of options when the team is overcommitting to an apparent obvious answer.
Watch out Three cosmetic variants of one concept do not create meaningful choice.
- 3
Expose underlying assumptions
Break each solution into the specific conditions that must hold for it to succeed. Separate those assumptions from the complete idea.
Pro tip Write assumptions so a test can produce observable evidence for or against them.
Watch out Testing a bundled solution hides which foundation caused the result.
- 4
Prioritize the risks
Rank assumptions by how uncertain and consequential they are. Focus initial tests on the assumptions that could invalidate a solution most quickly.
Pro tip Use production measurement to improve the team's judgment about which assumptions are usually risky.
Watch out Testing the easiest assumption first can create confidence while the decisive risk remains untouched.
- 5
Run small tests
Design each test to evaluate one prioritized assumption with minimal time and cost. Run multiple tests across the three ideas during the same week when possible.
Pro tip Aim for a cadence where the team can complete roughly half a dozen to a dozen assumption tests in a week.
Watch out A test that needs weeks may still be a project-based experiment in disguise.
- 6
Compare and revise
Review the evidence across all three solutions. Strengthen promising options, combine useful elements, or generate a better option when none survives comparison.
Pro tip Treat disagreement as a signal that the team may not yet have the best option.
Watch out Do not declare a winner from one convenient test while ignoring contrary evidence from another assumption.
- 7
Continue into delivery
Carry the selected and tested foundations into implementation, then instrument the released solution. Use production results as the next, larger feedback loop.
Pro tip Keep the boundary between testing and delivery intentionally fluid.
Watch out Treating assumption testing as a separate approval phase interrupts the continuous learning cadence.
In the wild
An illustrative streaming team explores three solutions for customers who cannot tell whether a show is good: richer cast information, trusted viewer signals, and a more informative preview. Instead of building three complete features, the team lists the assumptions behind each, prioritizes the weakest foundations, and runs small tests across the set. Evidence may show that recognizable cast information changes evaluation while a longer preview does not.
→ The team compares real evidence across options before committing delivery capacity to one full solution.
Common mistakes
Testing only the favorite idea
Without alternatives, the team cannot compare trade-offs and is likely to defend rather than evaluate its first solution.
Testing the whole solution
A large bundled experiment takes longer and cannot reveal which underlying assumption made the idea succeed or fail.
Separating tests from delivery
Treating assumption testing as a project before implementation undermines the continuous cadence the smaller tests are meant to enable.
Is it for you?
Best for
Important product opportunities where solution quality matters and the team needs evidence to choose among plausible options.
Not ideal for
Routine, low-risk product work whose expected impact and implementation pattern are already well understood.
From the episode
Teresa Torres on how to interview customers, automating continuous discovery, the opportunity solution tree framework, making the case for user research, common interviewing mistakes, and much more