Hypothesis

In product management, a hypothesis is a specific, testable statement of what a team believes will happen if it makes a change, written so that evidence can show it to be right or wrong. It turns a vague assumption into a prediction by naming the change, the people affected, the expected outcome and the signal that will confirm or refute it.

How a Hypothesis Works

A product hypothesis usually starts as an assumption, a belief the team is acting on without proof. Writing it as a hypothesis forces precision. Two widely used formats show the parts:

• Lean UX (Jeff Gothelf): "We believe [this business outcome] will be achieved if [these users] attain [this benefit] with [this feature]." This links a feature to the user benefit and business result it is meant to produce.
• Strategyzer Test Card: "We believe that..." (the hypothesis), "To verify that, we will..." (the test), "And measure..." (the metric), "We are right if..." (the success criterion). This adds how the hypothesis will be tested and what result counts.

A good product hypothesis is:

• Falsifiable. Some realistic result would show it is wrong.
• Specific. It names a user group and an observable behavior, not "users will like it."
• Measurable. It points to a metric the team can actually collect.
• Committed in advance. The success threshold is set before the test, so the team cannot move the goalposts afterward.

In a statistical test such as an A/B test, there is also a formal null hypothesis: the assumption that there is no real difference between the variants. The analysis checks whether the data are unlikely enough under that assumption to reject it.

Why a Hypothesis Matters

Without a written hypothesis, almost any result can be read as success. A launch that moves no metric gets described as "laying groundwork," and learning stops. A clear hypothesis makes the team state, before spending effort, what it expects and how it will know. It also turns a long list of ideas into a list of bets that can be ranked by risk and tested in order through a product experiment.

For a walkthrough of moving from customer discovery to testable hypotheses and data-backed requirements, read about SaaS idea validation loops.

Hypothesis Example

A team building invoicing software for freelance designers believes new users stall because they must re-enter old client details. They write: "We believe that letting new users import past invoices during onboarding will help them send their first invoice sooner. To verify that, we will offer the import option to half of new sign-ups for four weeks, and measure the share who send an invoice within seven days. We are right if that share is at least five percentage points higher than for users without the import option." The threshold is agreed before the test starts.

Hypothesis vs. Assumption

An assumption is something the team takes as true, often without noticing. A hypothesis is an assumption rewritten in a form that can be tested. Teams typically surface many assumptions, use assumption mapping to find the riskiest ones, and turn only those into hypotheses worth testing.