The scientific method tests the hypothesis.
Normalization turns an assumption into a decision.
A product leader can be punished this quarter for pursuing evidence that could change the roadmap. They may never be blamed next year when the product misses the market.
That’s the asymmetry.
When evidence is weak, preserving uncertainty creates immediate, visible internal decision risk. Building on an untested assumption creates delayed, diffuse market risk. The decision-maker carries the first. The organization absorbs the second.
So teams often choose the path that feels safer internally—even when it leaves the larger risk unresolved.
The Assumption Gap describes the condition: the distance between what an organization believes and what it knows. The Normalization Law describes the response: what happens when that unresolved assumption is treated as the answer and converted into a decision, requirements, roadmap commitment, and investment. The Assumption Gap reveals what remains uncertain; the Normalization Law explains how that unresolved uncertainty becomes embedded risk as commitment grows. One shows what the organization does not yet know. The other shows how much it commits before finding out.
Lean UX applies the logic of the scientific method to product development. It makes assumptions explicit, turns them into hypotheses, tests them, learns from evidence, and then decides what level of commitment is justified—it uses evidence to reduce market risk before investment grows. Normalization breaks that sequence by converting an assumption into a decision before it has been tested.
An assumption is not yet a hypothesis. A hypothesis is an assumption made testable. Normalization is attractive precisely because it skips that step—it reduces internal decision risk without reducing external market risk. The market consequences of a bad assumption may not surface for months, and by then tend to get attributed to execution, timing, sales, pricing, adoption, or change management rather than to the assumption itself. Normalization can be locally rational even when it’s systemically dangerous.
The cause isn’t primarily ego, ignorance, or bad leadership. Normalization converts an uncomfortable question into a plan the organization knows how to accelerate—it creates direction, alignment, estimable work, executive confidence, roadmap clarity, and visible progress. The assumption does not become less uncertain as it moves through requirements and planning; the organization simply becomes more exposed to being wrong. The unresolved uncertainty isn’t removed. It’s transferred to the market.
Uncertainty is not the same as risk. Uncertainty describes what is unknown. Risk grows when the organization commits time, money, dependencies, reputation, or opportunity cost while that uncertainty remains unresolved.
Where Both Paths Begin
The team sees an opportunity
A new product, improvement, or direction may create value.
The team discusses what should work
Experience, precedent, preferences, and constraints shape the discussion.
One direction begins to feel right
It gains support before evidence has tested it.
A direction has gained support.
The evidence is not yet strong enough to justify commitment.
What happens next—test it, or treat it as the solution?
Scientific Logic
Normalization Logic
Market risk is reduced before investment grows.
Evidence increases confidence in the direction.
Investment grows before market risk is understood.
Confidence in the plan is mistaken for confidence in the market.
Questions That Interrupt Normalization
Not comparison rows—moments to stop and ask before the direction hardens into a decision.
- What do we actually know here—and what are we just assuming?
- What do we need to learn before we commit?
- Does this help us test the idea—or let us skip testing it?
- Does the evidence we have justify this much investment?
- How much risk are we comfortable with before we have enough evidence?
Not a Story About Villains
Every hypothesis begins with assumptions—that’s not the problem. An assumption is not yet a hypothesis; a hypothesis is an assumption made testable. The problem occurs when an untested assumption is treated as though it were already established, and becomes the basis for a decision that gets converted into roadmap certainty without ever being tested. The Normalization Law describes a predictable organizational shortcut under uncertainty, not a personal failing.
Leaders sometimes must decide before sufficient evidence exists—that is not the failure. The failure is when the decision gets presented as though it were already validated, and the roadmap treats an open hypothesis as settled fact.
Testing later isn’t always wrong—some propositions require a live product to evaluate. The failure mode is using “we’ll test it later” to avoid cheap, decision-relevant learning before a large commitment, or testing only after what needs to be tested has already been removed.
The problem is not accepting risk. Product decisions always involve risk. The problem is increasing exposure while disguising how much remains unknown.
One path gets you to market. The other improves your odds of market fit.