The question behind the brief
A recovery and wellbeing app came to us with a deceptively simple concern: people needed to complete a fairly substantial opening assessment before they could begin the programme. The assessment took around 15 minutes. After that, the experience asked them to respond to short daily prompts and continue with a structured programme intended to last eight weeks.
But an app like this is not used on paper.
Someone may open it after a difficult day, when concentration is limited. They may be hopeful, wary or simply tired. A neutral-looking question can feel exposing. A progress indicator can reassure or make the task ahead look exhausting. A reminder intended as support can arrive at the wrong moment.
The useful research question was therefore not, "Can people complete the onboarding?" Most people can get through a form if asked. The better questions were about trust, readiness and expectation:
- Did people understand why the app was asking for personal information?
- Did they know what would happen to their answers?
- Could they pause without losing their work?
- Did the assessment feel relevant enough to justify the effort?
- Did people understand what the eight-week commitment involved?
- Did daily prompts feel supportive, intrusive or easy to ignore?
Those distinctions matter. Completion alone can conceal uncertainty. A person may finish onboarding while already deciding not to return.
How we approached the work
For this illustrative composite, we designed a small qualitative programme around the moments where confidence could be gained or lost.
We began with a product walkthrough alongside the client team. This was not a usability test. It helped us understand the intended service, the assumptions behind the questions and the practical constraints users never saw. We asked which answers affected personalisation, what was genuinely necessary at sign-up and where users could get human help.
Next, we ran moderated sessions with people who broadly reflected the intended audience. Recruitment for a sensitive product needed care. Participants understood the topic before consenting, knew they could skip any question and had a clear route to stop the session. We did not ask anyone to disclose more than was required to evaluate the interface. This was product research, not therapy or clinical assessment.
During each session, we observed the opening journey from the first screen through the diagnostic and into the first daily activity. We paid attention to hesitation, not only errors. A long pause before a consent question was often more revealing than a wrong tap. So was someone reading a privacy explanation twice or asking whether an answer would affect what they were shown.
We then used a short diary phase to understand the daily prompts in context. Moderated testing showed whether a reminder setting was understandable. The diary work showed how that reminder felt on a busy Tuesday morning or after a difficult evening. Entries recorded when the prompt arrived, whether the person opened it, what they expected and what made responding feel worthwhile or avoidable.
Finally, we held follow-up conversations about the full programme. A short study could not prove eight-week retention. It identified early signals such as clarity about pace, a believable sense of progress and confidence that returning after a gap would not feel like failure.
What the research revealed
In this illustrative scenario, the assessment was difficult because the effort arrived before the app had properly earned the right to ask.
Participants could see that the questions might lead to personalisation, but the benefit remained vague. The introductory copy spoke about creating a tailored plan without showing what tailoring would look like. Some people therefore treated the assessment as a hurdle. They gave quick, guarded answers, which reduced the value of the information the product was collecting.
The strongest concern appeared around control. People wanted to know whether they could stop, return later and change an answer. The interface technically allowed progress to be saved, but it did not say so at the point where reassurance was needed. This is a common usability problem: the system supports the user, but the content fails to tell them.
The daily prompts created a different tension. The client team had written encouraging messages and increased their frequency when early engagement looked low. In the composite research, several people interpreted this as pressure. A missed prompt was followed by another message, which made the app feel as though it was keeping score. For a recovery product, language that resembles disappointment can carry more weight than the team intends.
We also found a mismatch in the meaning of progress. The product measured completion. Participants described progress as noticing a trigger earlier, returning after a difficult day, or writing a short response instead of skipping. The design recognised attendance while people wanted recognition of effort.
None of these findings suggests that the app should become vague or undemanding. Structure can be valuable. The issue was how that structure was introduced and how much control people felt they retained within it.
Changes we recommended
Our first recommendation was to prepare people for the diagnostic before it began. A short explanation stated what the questions were for, how long the activity usually took, which answers affected the experience, whether progress was saved and how sensitive information was handled. The aim was informed participation, not persuasive reassurance.
We broke the assessment into meaningfully labelled sections. Each explained why the topic was relevant without implying a diagnosis. Questions not needed for immediate personalisation were deferred or made optional.
The pause-and-return route should be visible before people need it. If answers can be edited, that should also be clear. Recovery is not a stable state, and an answer given during onboarding may no longer feel accurate later.
For daily prompts, we recommended fewer assumptions in the language and more control over timing. "Would you like to check in?" left room for the person to decide. "You haven't completed today's task" framed the same moment as non-compliance. Users could adjust frequency, choose a preferred window and pause reminders without feeling that they had abandoned the programme.
We also changed how the product welcomed someone back after a gap. The return screen oriented the person gently, summarised where they were and offered a manageable next action. It did not foreground a broken streak.
Finally, we broadened the language of progress to acknowledge reflection, return and self-directed pacing. Wellbeing claims still required clinical and legal review. Usability research showed how wording was understood, but did not establish therapeutic effectiveness.
Illustrative results: sample reporting only
- Sample result: onboarding completion increased from 58% to 76% after the revised introduction, section labels and visible save-and-return option were released.
- Sample result: the proportion of users reaching the first daily activity increased from 46% to 64%.
- Sample result: support contacts about lost progress fell by 31% during the first month after release.
- Sample result: notification opt-outs fell from 27% to 16% after users were given clearer timing and frequency controls.
- Sample result: return to the app after a missed day increased by 22%, measured against the previous onboarding cohort.
- Sample result: in follow-up research, 8 out of 10 participants could explain why the diagnostic questions were being asked, compared with 4 out of 10 in the earlier sample.
These sample measures still require careful interpretation. Higher completion does not prove better wellbeing outcomes. Cohort comparisons must account for acquisition source, product changes and differences in the users recruited during each period.
What this kind of project gives a client
The client receives more than a list of interface issues. They get a clearer account of where trust is being asked for, what the product gives in return and how everyday context changes the meaning of prompts. Findings are prioritised against screens and journeys, with clear limits on what research can support and where clinical, privacy or regulatory expertise is required.
The most valuable finding may be a small one. A sentence about saved progress before a sensitive assessment can change how the task feels. Good usability work tests that moment and remains careful about what the evidence does not prove.
Have a similar usability question?
Tell us which journey matters, what evidence you already have and what decision the research needs to support. We will recommend a proportionate audit or research approach.