Start with a small ask
The initial task was to simplify onboarding. I saw an opportunity to use a few careful questions to make the rest of the app work harder.
Case study
MyProtein Personalisation
Turning a basic onboarding improvement into a more personalised app experience.
Try the working prototype (opens in a new tab)
Sign in with any email and password. The sign-in step is only there to replicate the real app experience.
- Role
- UX Designer, THG
- Duration
- About 6 weeks
- Team
- Me and a senior UX designer. I built the prototypes, and we collaborated on the UI
- Client
- THG / MyProtein
- Platform
- MyProtein app
- Focus
- Onboarding, personalisation, recommendation logic, app engagement
Overview
The MyProtein app was going through a wider redesign. Internally, there was a strong feeling that the app experience felt outdated and was not performing as well as it could. Although app conversion was stronger than web, there was a clear opportunity to improve engagement, returning users and the overall customer experience.
As part of the redesign work, onboarding was identified as one of the areas to improve. What started as a fairly simple onboarding task became a broader opportunity to personalise the app experience around user goals, activity level and clothing preferences.
My role covered UX thinking, onboarding design, behavioural UX, interactive prototyping, testing support and collaboration with app, ecommerce, trading, marketing and brand teams.
The challenge
- 01The app was seen internally as outdated and in need of improvement.
- 02The business wanted users to get into the app quickly, understand the value, browse and hopefully buy.
- 03There was no meaningful onboarding experience at the time, so the first challenge was introducing one without making the app feel slower.
- 04The bigger opportunity was to use lightweight onboarding inputs to make the app feel more relevant straight away.
- 05The experience had to balance user goals with commercial needs across nutrition and clothing.
The opportunity I saw
The useful part of onboarding was not the screens themselves. It was what the answers could do for the rest of the app.
Use answers beyond onboarding
Goals, activity level and clothing preferences could influence homepage copy, hero banners, nutrition recommendations and clothing product rails.
Keep it lightweight
The aim was not to ask lots of questions. It was to ask the minimum needed to make the first personalised state feel useful.
Make the app feel responsive
The For You page needed to feel like it had reacted to the user, rather than dropping everyone into the same generic homepage.
Why these questions mattered
Goals
Goals helped the app understand what the user was trying to achieve. This allowed copy and product rails to mirror that intent, such as weight loss nutrition or performance support.
Activity level
Activity level gave a rough signal for how much guidance someone might need. Less active users may need more explanation, while more active users are often more familiar with products.
Clothing preference
Clothing preference helped make apparel recommendations, imagery and product rails feel more relevant without asking unnecessary or sensitive questions.
Designing the personalised flow
The flow had to be short enough to respect intent, but useful enough to change what happened next.
The initial principle was to keep onboarding short, roughly three pages at most.
The flow needed to get users into the app quickly while still capturing enough preference data to make the For You page feel different.
Dynamic copy, hero banners and product rails changed based on preference combinations.
The goal was for users to feel that the app had responded to them, not that they had completed a form and landed on a standard homepage.
01Fuel Coach asks for a goal first.
02Follow-up options change with the goal. Build muscle leads to size, strength or lean muscle.
03Clothing preference shapes apparel recommendations.
04The For You page reflects the answers straight away.
Building realistic prototypes
The prototypes needed to prove the logic, not just show the intended layout.
Rather than relying only on static Figma screens, I built working interactive prototypes using AI-assisted workflows and web development languages.
The prototypes behaved more like a real app experience. Preferences selected during onboarding affected what users saw afterwards.
Nutrition products and clothing products were stored in JSON files and called dynamically based on user decisions.
This made it possible to test the logic of the personalised experience, not just the visual design.
It also made the idea much easier for stakeholders to understand because they could use it rather than imagine it.
Fuel Coach explains that the page was built from the user's answers, and how to change them.
Products are pulled from JSON data and tagged with the goal and focus they match.
Testing and iteration
Testing helped move the work away from a neat onboarding idea and toward a more useful personalised experience.
Round one challenged the short-onboarding assumption
The first round went well, but some users wanted more opportunity to explain themselves. A few felt the available answers made them feel slightly pigeonholed.
Users were willing to give more information
The interesting part was that users were happy to answer more when they understood it would improve the personalised experience.
Fuel Coach made the connection clearer
In the second round, we explored a more AI-assisted journey with Fuel Coach guiding users through questions and then explaining that the For You page had been built around their answers. This round tested better, because users believed they were getting genuine recommendations.
The prototype's limits showed in the recommendations
Because of how the prototype was built, some recommendations did not line up with the answers users had given. It showed that the recommendation logic needs to be as credible as the interface around it.
Personalised elements were noticed
Users recognised when imagery, content and product areas changed based on their answers, which helped them understand the purpose of onboarding.
Recommendations needed more explanation
Testing continued to show promise, but users wanted clearer reasons for why specific products were being recommended.
Recommendation trust
The strongest insight was that personalisation only works if users trust it. If a product is recommended, users need to understand how it connects to their goal, not just feel like the business is pushing something at them.
The question became less “can we recommend products?” and more “how do we explain why these products fit your goals?” The third round of designs is currently exploring how to give users more nuance and clearer reasons behind recommendations.
Stakeholder alignment and business complexity
- 01Senior leadership responded positively to the prototypes and the direction.
- 02The prototypes made the opportunity easier to understand, especially around clearer recommendations and more targeted app experiences.
- 03I worked closely with App Marketing, Trading, Ecommerce and Brand teams.
- 04There were ongoing conversations with the external app team who build and manage the app.
- 05Onboarding was not automatically seen as the priority by every part of the business, so alignment across teams became an important part of the work.
- 06I invited the external app team to the second round testing playback to show user response and build support around the approach.
- 07MyProtein stakeholders were aligned around the opportunity, while the wider delivery context still involved different teams with different priorities.
More from the prototype
Outcome / current status
- The concept tested well across multiple rounds.
- Users understood the connection between onboarding and the personalised For You experience.
- Senior stakeholders were excited by the direction and the commercial and product opportunities.
- The work helped shift the conversation from simple onboarding improvements toward a broader personalised app experience.
- Further iteration is ongoing, particularly around giving users more nuance and improving recommendation transparency.
- KPIs such as conversion, returning users and successful downloads sit within the wider app redesign context, but there are no final results to claim yet.
Reflection
A small UX task opened up a much bigger product opportunity. The brief started with onboarding, but the more interesting challenge was how to make the app feel useful and relevant without slowing people down.
The work is still evolving, but the testing has already shown how important it is to give users control, explain recommendations clearly and make personalisation feel genuinely helpful rather than purely commercial.
Like what you see? Let's talk.
Email me at davidedwardsjones@googlemail.com, or get in touch through the contact page.