Related characteristic
A Related characteristic is something that shows up in less than half the cases when users go through a given activity, but fits plausibly when it does. It's correlated with the activity rather than required by it, and its presence doesn't determine whether the activity counts. Bringing a measuring tape to buy a couch fits the pattern: some people do, most don't, and the purchase still counts either way.
Definition
A candidate characteristic of a scoped transition that's plausibly present but only weakly indicative for sorting whether an instance belongs to the transition's category — the weakest discriminating tier in condition tiering, alongside Essential and Expected.
The three tiers
Graded condition decomposition sorts candidate characteristics of a scoped transition into tiers by their definitional role:
- Essential: definitional. If absent, the instance isn't an instance of the scoped transition. Operationalizes as completion criteria.
- Expected: typical. Present in most instances by default; absence is diagnostic but doesn't exclude.
- Related: correlated. Plausibly present in some instances, but doesn't determine whether an instance belongs to the category.
The tiers do different discrimination work. Essential conditions cleanly separate scoped transitions from their neighbors. Expected conditions give probabilistic signal. Related conditions are weakly indicative — they show up sometimes, and their presence or absence tells you little about whether the transition in front of you is the one being described.
Feeding auxiliary routes
Related characteristics aren't completion criteria; they don't meet the definitional threshold. Their job is to inform auxiliary routes — user activity that's structurally in scope for the parent transition but not causally required to ratify its parent-scale claim.
This matters for affordance design. The critical route tells you the minimum set of causally-required sub-transitions given committed endpoints. Auxiliary routes describe the broader terrain that actual users traverse around that minimum. Related characteristics populate the parts of that terrain that appear plausibly but inconsistently — design choices around them shape the experience of users who do follow those paths, even when most users don't.
Distinction from Expected
Frequency is the only distinction between Expected and Related. Expected characteristics show up in most instances; Related characteristics show up sometimes. The kind of work both do is the same — both feed auxiliary routes, both inform affordance-landscape thinking, neither serves as a completion criterion.
The frequency difference matters for prioritization, not category. Expected characteristics describe terrain you can count on most users encountering; Related characteristics describe terrain that's plausible but less consistent. Both belong in the design picture; the weight assigned to each differs.
Example
For the transition "I have grilled steak I can eat now":
- Essential (completion criteria): steak present, cooked through, at a particular doneness, seasoned.
- Expected: rested before cutting, served warm.
- Related: topped with compound butter, basted during cooking, served on a cutting board for slicing.
Removing a Related characteristic — say, the user skipped the compound butter — doesn't change whether the instance counts as grilled steak. But recognizing that some users finish steaks with compound butter changes what affordances might be worth building around the grilling transition: a grilling app that includes compound butter recipes is shaping an auxiliary route that some users follow, even though most users don't.
See also
- Scope fidelity — the constraint that a transition's name accurately compresses what's inside its boundary, including which tier each characteristic occupies.
- Resulting state — where Essential-tier characteristics live as completion criteria; the part of the transition characteristics most often anchor to.
- Re-sort validation — the validation operation that grounds tiering decisions in actual response distributions rather than analyst impression.