Evaluating Solutions: The 5 Types of Assumptions that Underlie Our Ideas

Teresa Torres argues that good continuous discovery depends on first surfacing the hidden assumptions behind each idea, since teams can't test beliefs they haven't named. She proposes five assumption categories (desirability, viability, feasibility, usability, and ethical) and illustrates them with a local newspaper trying to get readers to share articles. Each category maps to a different kind of risk, and naming them makes it clear what to test. She also explains how her framing relates to Marty Cagan's four risks, keeping her own vocabulary because it makes testing more concrete. For product people, the piece offers a practical checklist for stress-testing ideas before building them.

01Key takeaways

• List the assumptions behind an idea explicitly before deciding what to build or test.
• Don't rely on asking customers 'do you want this?'; run demand tests that reveal real behavior.
• Desirability includes whether customers will do the work the solution requires, not just whether they want it.
• Check viability by confirming the economics, such as whether revenue or benefits offset delivery costs.
• Assess feasibility beyond engineering, including legal, compliance, security and organizational acceptance.
• Consider ethical risks like data collection, targeting exclusions and harm to relationships before launching.

02Key sections

Why assumptions are hard to see

Assumptions are core beliefs that must hold for an idea to work, and they are often invisible to the people holding them. Categorizing the kinds of assumptions teams usually make helps surface them.

A worked example

A newspaper team sees readers wanting to share articles and generates three share-related solutions. Each one rests on many specific assumptions that can be listed out.

Desirability and viability

Desirability covers whether customers want a solution and will do what it requires, and asking people directly is unreliable so demand tests work better. Viability covers whether the economics, including costs and revenue, make the solution worth it.

Feasibility, usability and ethics

Feasibility extends beyond engineering to legal, security and organizational acceptance. Usability asks whether customers can find and use the solution, and ethics asks what harm could result from data use, targeting or social dynamics.

Relation to Marty Cagan's risks

The categories closely overlap with Cagan's value, viability, feasibility and usability risks, but Torres keeps 'desirability' because she finds it distinct from usability and more actionable.

03From the post

“Assumption testing is at the heart of what good continuous discovery teams do week over week. It’s how we evaluate which ideas will work and which won’t. But before we can test our assumptions, we have to identify them. That’s not always as easy as it sounds.”

04Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.