Validation plan for the Phase 17.10 redesigned admin web for extended-session usability — focus retention, density fatigue, and recovery after interruption.
Scope#
Validates the redesigned admin shell + queue + detail surfaces under sessions of:
- 30 minutes continuous triage (
focused triage); - 2 hours mixed triage + inspection (
mixed shift); - 4 hours covering an on-call window (
oncall block).
The validator is the operator (a real human reviewer); this spec defines the protocol they follow and the metrics they record.
Metrics#
For each session length, record:
- Focus loss events — count of moments when the operator could not visually re-locate their cursor / row / decision action after a brief glance away.
- Density fatigue rating — 1–5 scale, captured at the end of each session, on the question "how tired do your eyes feel reading the chrome?". 5 = severe fatigue.
- Recovery time after interruption — the validator simulates a 90s interruption (close laptop lid, return) and records the time-to-resume in seconds.
- Banner / motion fatigue rating — 1–5 scale, on "did visual motion / banner stacking distract you?".
- Density preference — at the end of every session, which density did the
operator switch to (
comfortable/compact)?
Protocol#
For each session:
- Operator picks one workspace category (Operations, Governance, Customer, etc.) and runs typical triage.
- Operator records the metrics at the half-way mark and end.
- Validator runs the same protocol three times across the three session lengths.
Exit criteria#
- Mean focus-loss events per hour ≤ 1 across all 9 session runs.
- Mean density fatigue rating ≤ 2 / 5 in
comfortable, ≤ 3 / 5 incompact(some fatigue acceptable in compact). - Mean recovery time after interruption ≤ 5 seconds (the cursor-persistence rule from V1-AWEB-100 should make this near-instant).
- Mean banner / motion fatigue rating ≤ 2 / 5 (the no-motion- on-render rule from V1-AWEB-098 should make banner fatigue rare).
- ≥ 50% of operators choose
compactfor the 4-hour session (validates that the density toggle is useful for long sessions).
Deliverables#
- Validation report appended to this doc — per-session metrics, mean across the three runs, and any exit-criterion misses with their root-cause notes.
- Issue log — per-issue Linear-style entries for any surface that fails the exit criteria; owners assigned to the panel-author for fix.
- No automated Playwright spec — extended-session usability is human-validated; no spec is added.
Out of scope#
- Mouse-only sessions (the spec assumes the V1-AWEB-101 keyboard flow has already passed).
- Multi-operator hand-off sessions (deferred).
- After-hours theming (deferred).