Managing usability test assets: videos, transcripts, and notes
A usability session creates more than a recording. It also produces a task script, consent record, observer notes, participant comments and a path back to the decisions made afterwards. Treating these as separate files quickly makes research difficult to reuse.
UCDmanager provides a free, collaborative place for teams to organise that evidence within User-Centered Design projects. Teams can document roles, personas, use cases, heuristic evaluations, usability tests and web accessibility conformance work, then connect findings to the activities that produced them.
The central task is to preserve context. A clip without its task, session identifier or consent status may look precise yet prove ambiguous. A transcript without observation notes may omit what happened between spoken words, while notes without timestamps can be hard to verify.
This workflow covers recording, transcription, analysis, governance and long-term retrieval. It applies to in-person sessions and remote research across Australia, where participants, observers and decision-makers may work from different cities and time zones.
Planning a usable asset library
Start with the research purpose and the evidence the team expects to use. Decide which sessions need video, which need audio or written notes, and who will record, transcribe, code and approve each asset. This prevents a pile of recordings from becoming the project’s main output.
Build a simple asset map before recruiting. Link every session to a research question, task script version, participant pseudonym and consent record. For a study involving participants in Sydney and Melbourne, the same template should also capture device, browser, connection and whether the session was remote or in person.
Naming and organising recordings
Use identifiers rather than names in filenames. A pattern such as BANK-REMOTE-S03 can combine a project code, study mode and session number without exposing identity. Keep the source recording’s extension and use ISO dates for sorting, while spelling words out in Australian English to prevent avoidable confusion.
Give each recording a catalogue entry containing its session identifier, owner, capture date, duration, storage location and processing status. In UCDmanager, record references and links within the relevant use case or test activity, and keep raw footage, transcripts and derived clips distinguishable.
Building transcripts that support analysis
Begin transcripts with participant, facilitator and observer labels, then retain time markers against the task schedule. Automated speech recognition can accelerate the first draft, but a person should check names, overlapping speech, quiet responses and Australian expressions before the transcript becomes an analysis record.
Structure the verified transcript around task IDs, marking observations and uncertainty rather than presenting every sentence as equally important. Store a redacted working copy where needed, but retain the original under tighter access. Link short quotations to timestamps and the associated evidence.
Connecting observations to task evidence
Keep observation notes separate from spoken comments. Record what the participant did, what they said, what the facilitator did and the task outcome, including deviations from the script. This distinction helps analysts revisit an interpretation without mistaking it for direct evidence.
The same evidence model can support design reviews and accessibility work. A web accessibility audit guide provides a useful structure for pairing conformance findings with observed barriers, giving teams a clearer basis for prioritisation.
Permissions, consent and asset readiness
Consent should state what will be recorded, transcribed, quoted and retained, including any recording that occurs outside the main task. Australian teams should also check organisational privacy policies and, where applicable, the Australian Privacy Principles under the Privacy Act 1988. Where participants include Aboriginal and Torres Strait Islander communities, follow the relevant community governance and data-sovereignty protocols.
Apply access by role, not by convenience. Keep consent forms, contact details and identifiable footage apart from pseudonymised evidence, and remove names, account details and notification content before sharing a clip. A participant request may require a correction, deletion or restriction, so the procedure needs an owner.
Before recording, confirm that the following information and materials are ready:
- Consent status and recording scope linked to the session.
- Unique session ID, task version, participant pseudonym and test date.
- Video reference plus draft and verified transcript locations.
- Notes covering task outcome, observed behaviour and participant comments.
Assign ownership and governance for each part of the asset lifecycle:
- Named owners for recording, transcription, note review and participant requests.
- Role-based permissions separating raw footage from approved quotations.
- Retention dates, deletion steps and a privacy review schedule.
- Shared naming rules and version history for scripts and transcripts.
Turning evidence into reusable design artefacts
Findings should remain traceable to the sessions that support them. Link recurring behaviours to the relevant persona assumptions, use cases and design requirements rather than replacing them with untagged claims. UCDmanager’s persona templates provide a practical way to keep those assumptions visible while evidence accumulates.
Give shared clips descriptive titles and short summaries, and attach the task, participant perspective and observed behaviour. Publish accessibility evidence in accessible formats, with captions for video and readable transcripts, so consumers are not forced to download large media files simply to understand the result.
Keeping the library useful over time
Review the library after each milestone, not only at project close. Check that links still resolve, transcripts match the final recordings, permissions reflect consent and obsolete duplicates are removed. Record session times with an explicit time zone: a comparison between observers in Perth and Melbourne can otherwise appear to contain contradictory observations.
A migration or tooling decision should examine the complete evidence trail, not merely the volume of stored media. Broader platform comparisons can prompt useful questions about organisation and retrieval, while research suitability still depends on participant privacy, provenance and project context. UCDmanager lets teams browse a demonstration project or create their own, then maintain a searchable record that survives changes in staff and study priorities.