An interface inventory of a course-registration portal finds 31 distinct grays and 9 button styles; how do you decide which variants are drift and which are intentional?
answer
- counts alone mislead
- can anyone name its purpose?
- can users tell them apart?
- failing contrast is a defect
- confirm with the owning team
basics
~20 sCluster the variants, then test each against purpose and perception: a variant with a nameable job that users can distinguish is intentional; near-identical copies are drift; any text color that fails the WCAG contrast minimum is a defect regardless of intent.
solid answer
~50 sRaw counts only show that variety exists; the job is to classify it. I would cluster the grays by value and the buttons by visual treatment, then ask of each cluster: **can someone name its purpose** — primary vs secondary action, a disabled state, a denser table — and **can users perceive the difference**? A variant with a real purpose and a visible difference is intentional; near-identical copies from different teams or copy-paste are drift. Any text color that fails the WCAG 2.2 Level AA minimum of 4.5:1 against its background (3:1 for large text) is a defect whatever the intent. I would confirm with the owning teams, since some differences follow native mobile platform conventions on purpose, and present a table of clusters, counts, classification and evidence. Choosing the consolidated values and their token structure is the next step, not part of the inventory.
go deeper
Recall that a high variant count does not mean every variant is wrong; some differences, like primary and secondary buttons, exist on purpose.
Explain the tests: cluster, then check purpose, perceptibility, accessibility and usage, and know the WCAG 2.2 AA text contrast minimum of 4.5:1, or 3:1 for large text.
Show how you classify with the owning teams, separate drift from defects and platform-specific choices, and present evidence that turns counts into decisions.
Consider the politics of classification: labelling a team's work as drift needs evidence and tact, or the inventory creates resistance to the system it is meant to justify.
## Why the count is not the finding An **interface inventory** catalogues the UI elements a product really uses. Its headline numbers — thirty-one grays, nine button styles — are striking, but a count only proves that variety exists. Some variety is exactly what a good interface needs: a primary and a secondary button *should* differ. The analytical work is separating variants that serve a purpose from ones that accumulated by accident, so that the later consolidation removes noise without removing meaning. ## Four categories of variant | Category | Test | Example in the portal | What happens next | |---|---|---|---| | **Intentional** | Nameable purpose and a difference users can see | Primary register button vs secondary add-to-shortlist button | Keep; document the purpose | | **Drift** | No nameable purpose, or a difference users cannot perceive | Four grays within a hair of each other used for the same secondary text | Flag for consolidation | | **Defect** | Fails an accessibility or usability requirement | Light-gray help text below the contrast minimum | Flag as a bug, whatever its intent | | **Context-specific** | Deliberate difference driven by a platform or medium | The mobile app's native-style toggle vs the web app's toggle | Keep and document the divergence | ## Tests to apply to each cluster 1. **Cluster first.** Group values that are close to each other and components that share a visual treatment; classifying thirty-one individual grays one by one hides the pattern. 2. **Purpose.** Ask the owning designers and engineers what each variant is for. If nobody can name a job it does that its neighbours do not, it is probably drift. 3. **Perception.** If users cannot tell two variants apart, the difference carries no meaning for them, even if someone once intended it. 4. **Accessibility.** Check text colors against their backgrounds: WCAG 2.2 success criterion 1.4.3 Contrast (Minimum), Level AA, requires at least 4.5:1 for normal text and 3:1 for large-scale text. Criterion 1.4.11 Non-text Contrast, also Level AA, requires 3:1 for the visual information that identifies UI components and their states, with exceptions such as inactive components. A variant that fails is a defect, not a style choice. 5. **Usage.** Record how often and where each variant appears. The most used variant is strong evidence of the de facto standard, but not proof — it may be the one that fails contrast. ## Traps to avoid - **Treating every distinct value as drift.** Status colors, disabled states and emphasis levels are distinct on purpose. - **Treating intent as justification.** A variant can be deliberate and still unnecessary; intentional means it needs a documented reason, not that it survives automatically. - **Flattening platforms.** Differences between web and native mobile that follow each platform's conventions are context-specific, not inconsistency to erase. - **Deciding the future in the inventory.** Classification says what exists and why; which values survive, and how they are structured as tokens, is a separate design step. ## Presenting the result The deliverable is a table per category: cluster, count, where it appears, classification, and the evidence behind it (the owner's explanation, a contrast measurement, a screenshot pair). For the portal, that might read: nine button styles resolve into three intentional emphasis levels, four drift copies of the primary style, one disabled style with failing contrast, and one mobile-specific style kept on purpose. That framing turns a shocking number into a decision-ready list.
- Should the most frequently used variant automatically become the standard?Not automatically. Frequency is strong evidence of what teams and users are used to, but the most common gray may fail contrast, or the most common button may lack a visible focus indicator. Treat usage as one input alongside purpose, accessibility and perception, and record why the chosen variant won.
- How do platform conventions affect whether a variant counts as drift?Some differences are deliberate: a native mobile app may follow its platform's control styles or navigation patterns while the web app does not. Those are intentional divergences to document, not drift to remove. The test is whether the difference serves that platform's users or simply happened without a plan.
saying these in an interview costs you the question
- Every distinct value an inventory finds is drift and must go.
- A variant someone created on purpose must always be kept.
- The most frequently used variant should always become the standard.
- Every difference between the web and mobile apps is an inconsistency.
- Body text below the contrast minimum is fine if it was chosen deliberately.