Duration: 3 months
Overview
ieso health is a digital mental health services provider. As the lead product designer for the ieso mobile app I led a redesign of the onboarding experience which led to improved user retention and app engagement. I was working as part of a cross-functional team with engineers, AI scientists and a product manager.
The app guides users through a 6-week therapy program aimed at supporting individuals with anxiety. The app utilises generative AI to make sessions feel conversational and personal to the user. The primary objectives were to enhance user retention and engagement, and to clearly communicate the program’s structure and benefits. Key metrics for success were reducing the drop-off rate from registration to the first session and increasing the frequency of app usage.
A dual-track agile approach
To achieve our goals, I agreed with the team to adopt a dual-track Agile approach, incorporating both discovery and delivery tracks. This approach allowed us to rapidly iterate and implement improvements based on continuous feedback and validation.
Reviewing the current onboarding experience
My first step was to thoroughly understand the current onboarding experience. I mapped out the existing user journey, identifying pain points and potential opportunities for improvement. This analysis was supported by pre-existing insights from user interviews. Key pain points and opportunities identified included expectations not being well set for the program, a lack of information about how the program works and a lack of sign posting, so users didn’t know how long the getting started journey would take, and where they are in the journey.
Reviewing onboarding experiences for other apps
To help stimulate ideas I carried out a competitor review to examine the getting started journey for similar mental health apps such as Calm, Holly health and Iona. This identified some interesting ideas to explore, such as utilising social proofing, and including customer testimonials to help build trust.
Defining ideas as lean experiments
The cross-functional Agile team I was a part of usually defined app work as user stories. However, I encouraged the team to instead think of these ideas as lean experiments. This shift emphasized the importance of measuring impact and informed our selection of additional metrics.
Prioritising ideas by impact & confidence
To help evaluate and prioritise ideas I facilitated a workshop with the team and key stakeholders. Ideas were discussed and then mapped based on their anticipated reach and impact (against the desired goals), as well as our confidence in these predictions. This helped to identify ideas that we could take to delivery as there was high confidence and impact, and those that required discovery activities (such as user testing) to help explore key assumptions and unknowns.
Mapping a new onboarding experience
Having identified the most promising ideas, I began by identifying where each idea would fit within the onboarding journey. This was also important for identifying where there were dependencies, and the order in which improvements could potentially be released.
Using wireframes to collect early feedback
To allow for rapid exploration of various design directions I started with wireframes. This allowed me to rapidly iterate early designs based on feedback from the team, key stakeholders and other designers. It also avoided getting too fixated on UI design details as the focus was on the idea, not the UI design.
Prototyping onboarding ideas
Having iterated the early designs, I developed a high-fidelity interactive prototype (shown below) using Figma. This enabled users and stakeholders to experience the proposed onboarding journey from login to the first session. A high-fidelity prototype was key to showcasing design ideas, and for carrying out testing with users.
Carrying out unmoderated & moderated user testing
In order to get user feedback for the potential onboarding improvements I conducted both unmoderated and moderated user testing via Lookback. Unmoderated sessions offered quick feedback with quantitative data (such as usability rating and ease of task), whilst moderated sessions provided deeper insights through direct interaction with users. For example, the moderated testing indicated that whilst it was useful as part of onboarding to provide more information about the program, some users felt that they were being overwhelmed with too much information.
Users carrying out unmoderated user testing were also asked to rate how useful they would find some of the key concepts, such as FAQs about the program and setting up a schedule.
Analysing insights with the team
Working with the team, I analysed the feedback from user testing, mapping it against our proposed changes. This mix of quantitative and qualitative data highlighted the most impactful ideas and revealed additional user needs, such as the demand for greater flexibility in the therapy schedule.
Defining key metrics using the Google HEART framework
It’s vital to be able to measure the impact of changes, and of course whether they are having the desired effect. In order to discuss and agree how to do this I ran a workshop with the team to identify key metrics to track, such as average app usage per week. I used the Google HEART framework to help frame these conversations.
Iteratively delivering updates
Following an analysis and then prioritisation workshop with the team, the most impactful onboarding ideas were added to the app backlog, to be delivered in upcoming sprints. I worked with the engineers to document (in Jira), develop and then review user stories prior to release, and then to evaluate the performance once improvements / experiments were live.
Results
The user-centred design approach employed ensured that these changes improved both user retention and engagement. Furthermore, by adopting a Lean, experiment-driven approach, I de-risked the project and laid a solid foundation for ongoing improvements to the user onboarding experience.
Learnings
- The power of framing ideas as experiments – Treating ideas as hypotheses to test helped to foster an experimental mindset within the team and helped to identify success metrics to track.
- De-couple changes where possible – Breaking updates into smaller, independent changes enabled faster updates and reduced the risk caused by large-scale releases.
- Moderated and unmoderated user testing complement each other – Moderated sessions provided deep qualitative insights, while unmoderated testing provided quick quantitative feedback.