Can Anyone Find the Right Passage? UX for Media Applications
Evaluating the usability of a media application: real tasks, clear feedback, accessibility and a sound UX comparison.
Search returns a relevant lecture. Yet the student cannot find her explanation: the video lasts 90 minutes, the result opens at the beginning and the relevant passage lies somewhere in the second half. The search algorithm has recorded a success. The task remains unsolved.
Human–computer interaction investigates how people understand and use digital systems. For our example media library, this perspective begins with a concrete question: what path does someone actually have to follow to find and verify a particular statement?
Usability depends on the task
Usability describes whether specified people can achieve specified goals effectively, efficiently and with satisfaction in a specified context of use. This dependence on context is central. An editorial interface used daily can display more functions at once than an occasionally used video search. User experience also considers the experience of use and the associated perceptions and responses. ISO 9241-11:2018: Usability — Definitions and concepts
For the library, “finding an explanation” would first need closer investigation. Do students search for a known technical term? Do they remember a slide? Do they want to understand a concept or quote a particular phrase? Interviews provide clues; observing an actual search can reveal additional difficulties.
This investigation would inform the design. If people often search for individual statements, a result needs text context and a timestamp. If they frequently compare entire lectures, summaries and metadata become more important. Simply expanding both at once would not yet be a justified prioritization.
Feedback is part of the feature
Nielsen’s heuristics include visibility of system status, understandable language, consistency, user control and error prevention. They support systematic interface evaluation. An ongoing operation needs feedback; a failed action needs a clear explanation and a possible next step. Nielsen: 10 Usability Heuristics for User Interface Design
In our design, a video could already be playable while transcription is still running. “Loading” would be too vague. “Video available; transcript being created” explains which feature can already be used. After a failure, retrying should be possible without uploading again.
These messages need a reliable technical state behind them. The state models in Part 1 are therefore a prerequisite for good interaction. Better wording cannot permanently compensate for an ambiguous backend state.
A study with a verifiable question
For the example library, we propose this comparison: one variant shows a list of videos, the other timestamped transcript results. The question is whether people find specific explanations more reliably and with less effort using the second variant. This is a study plan, not a reported study result.
| Observation | What it answers in the example |
|---|---|
| Correct passage found | Was the substantive goal achieved? |
| Time until the find is confirmed | How much effort was required? |
| Irrelevant results opened | Where do detours occur? |
| Explanation of the result | Was the content understood? |
| Assessment after the task | How was the workflow experienced? |
Tasks would need to have comparable difficulty. Someone answering the same question twice already knows the answer on the second attempt. Different tasks and varying the order of the variants can reduce this effect. The task wording must also avoid revealing the interaction being tested.
An average alone is insufficient for analysis. Experienced people might benefit greatly while others overlook the timestamps. In that case, separating observed usage patterns offers better guidance than a blanket assessment of the entire interface.
Observations explain the numbers
A short completion time may mean the task was easy. It may also mean someone gave up too early or mistook an incorrect result for the right one. We would therefore evaluate timing together with a substantive check of success.
Participants could also describe their reasoning while searching. This can reveal, for example, whether they mistake an excerpt for a complete answer. Such qualitative observations support design, while reliable quantitative comparisons require a suitable sample and analysis. The scope of the claim must match the study.
Accessibility from the start
WCAG 2.2 describes testable requirements under the principles of perceivability, operability, understandability and robustness. Captions, keyboard operation and visible focus are particularly relevant to media applications. Essential information must not be conveyed by color alone. W3C: Web Content Accessibility Guidelines 2.2
In the library design, a search result must therefore be selectable without a mouse. Its timestamp needs an understandable label. Automatic captions should be reviewable and correctable; their mere existence does not demonstrate sufficient quality.
Responsive layouts and color schemes also affect use. The existing articles on Container Queries (in German) and Dark Mode (in German) explore these design topics further. Visual analysis in Part 8 extends the perspective to more complex data views.