Condition tiering
Condition tiering is a way of sorting everything that might be part of an activity into three buckets (Essential, Expected, Related) by how reliably each one shows up. Essential things are present in nearly every instance; without one, the case doesn't count as the activity. Expected things are present most of the time but not always, so an absence is worth noticing without disqualifying the instance. Related things appear in some instances but not others, coloring the activity without determining whether it counts.
Definition
A sorting operation that separates the candidate characteristics of a scoped transition into three tiers — Essential, Expected, Related — by their definitional role in the transition.
The three tiers
Each tier is defined by how often a characteristic appears across instances of the transition and what its absence implies:
- Essential (>90% of instances): definitional. If absent, the instance isn't an instance of the scoped transition.
- Expected (>50%): typical. Present in most instances by default; absent cases are diagnostic but don't disqualify the instance from the category.
- Related (<50%, directional): correlated. Plausible in the transition, not required for sorting whether an instance belongs.
What each tier does
Essential conditions cleanly separate scoped transitions from each other — they're the sorting test for category membership. Expected conditions give probabilistic signal that's useful for affordance design but doesn't decide membership. Related conditions are weakly indicative, more color than structure.
Essential characteristics operationalize as completion criteria — the IF-conditions of the resulting state that must hold for the transition to count as complete. Expected and Related characteristics don't meet that threshold, but they feed auxiliary routes: the broader terrain around the critical route that users commonly traverse without it being structurally required. The frequency distinction between Expected and Related is the only difference between them; the downstream work they do is the same.
Where candidates come from
Candidates can come from anywhere — qualitative investigation, team elicitation, common-parlance decomposition of socially-defined transitions, or any combination. The operation doesn't care where the candidates were surfaced; it cares whether the tiering itself is rigorous. What makes tiering effective is the quality of the candidates and the discipline of the sort, not the method used to generate them.
Example
A team runs a sticky-note exercise on "receiving a job offer" and surfaces candidates like: crafting and sending a reply, reading the offer letter for every detail, deciding whether to push back, telling family, doing a fist pump, researching market rates, popping champagne, worrying about a bad decision.
Tiering sorts them:
- Essential: crafting and sending a reply. The transition isn't complete without it.
- Expected: reading for detail, deciding whether to push back, telling family, worrying about getting it right. Typically present, not definitional.
- Related: fist pumps, market-rate research, popping champagne. Color the transition without affecting whether it counts as a job-offer-received.
The Essential tier becomes completion criteria once translated to state-vocabulary. The Expected and Related tiers inform downstream affordance design — the team now knows what surrounding terrain to consider when designing for users navigating this transition.