Self-guided
Participants complete the same three task families on their own device. Complete sequences can inform issue discovery and vocabulary gaps.
Never advances the moderated gatePublic validation ledger · Study directory-core-v1
Validation is a claim with a receipt.
Internal QA can prove that a page renders. Only representative people attempting realistic tasks can show whether the resource hub works for them.
Not yet validated0 / 10
qualified moderated baseline sessions
Moderated evidence0 / 10
Qualified baseline sessions only.
Self-guided signals0
Reported separately; never added to the gate.
Confirmation0 / 3
New people after repeated failures are fixed.
Latest evidenceNone
The ledger stays incomplete until evidence is published.
01 · Two evidence lanes
More participation without weaker claims.
Public, self-guided studies can expose search and labeling friction at scale. They cannot replace observation, so the moderated target remains separate.
Moderated
A present facilitator follows the neutral protocol, records any help as assisted, and preserves a de-identified local record.
Must pass the qualification contractThe reporting script rejects legacy events without a mode, unattested moderated sequences, incomplete or duplicate task sets, and every session with a later withdrawal marker.
02 · Who must test it
Two audiences, balanced on purpose.
A cross-domain hub can work for a familiar expert and still fail everyone approaching from a different craft or production role.
Makers
People who sew, mend, knit, crochet, weave, dye, or otherwise work with textiles for personal, teaching, or small-scale creative use.
Apparel production
People responsible for apparel product development, technical design, pattern work, sourcing, compliance, or production.
Recruit actual or likely users, including people with disabilities and people using their normal assistive technology or device. Participant characteristics and study scope belong in the aggregate report; identities do not.
03 · What they attempt
Find. Compare. Challenge.
The task wording changes by cohort, but the product abilities under test stay fixed.
Find
Locate a useful resource for a believable need without being told which category, filter, or search phrase to use.
Compare
Compare two plausible resources and explain which one better fits the stated constraints.
Challenge and correct
Explain what a listing does not prove and find the route for reporting an inaccurate or stale detail.
Recorded for every task
- Completion without help, assisted completion, or failure
- Elapsed time to the participant's chosen useful result
- One primary friction category
- Consent-based, de-identified moderator observations kept separately
04 · Claim gates
No “user validated” badge before all five gates pass.
- 01Pending
Five makers and five apparel-production participants complete qualified moderated baseline sessions.
0/5 makers · 0/5 apparel
- 02Pending
Every qualified moderated baseline session records the three exact cohort tasks and elapsed times.
0/10 complete sessions
- 03Pending
The five most repeated failures are fixed or explicitly accepted with a reason.
0/5 repeated failures addressed
- 04Pending
At least three new participants complete qualified moderated confirmation sessions after the fixes.
0/3 new participants
- 05Pending
An aggregate validation note publishes participant mix, scope, failures, changes, and remaining limits without personal data.
No evidence note published
05 · Evidence ledger
Empty means empty.
0
No participant round has been published.
Production logs, completed sessions, repeated failures, shipped fixes, and confirmation results are still pending. Internal desktop and mobile QA are excluded from this count.
Method and privacy
The protocol is public. Participant data is not.
The task runner sends only an anonymous in-memory session code, cohort, study mode, round, task identifier, outcome, friction category, timing, and any withdrawal marker. It accepts no name, contact detail, recording, business information, or free-text response. Facilitators need separate informed consent for notes or recordings.
Exported study events are retained through the active round and scheduled for deletion within 180 days after its aggregate report. A participant can withdraw in the runner or later by anonymous session code; the reporting workflow excludes that entire sequence.
The method follows primary guidance on neutral moderated tasks, representative participants, consent, data minimization, and reporting study scope from the GOV.UK Service Manual, GOV.UK consent guidance, and W3C Web Accessibility Initiative.
Participate or facilitate
Use the instrument. Keep the claim honest.
Anyone eligible can use the self-guided route to expose friction. A qualifying moderated session starts from the facilitator kit, follows the neutral protocol, and remains a separate evidence lane.