Arnold Wei | Product Designer

Role

Product Designer

TimeLine

6 weeks (2025)

Team

1 PM

1 Product Designer (Me)

2 Frontend Devs

1 Backend Dev

1 PM

1 Product Designer (Me)

2 Frontend Devs

1 Backend Dev

Skills

Figma, User Research

Project Overview

Athes is an AI lacrosse training platform that democratizes elite coaching by helping youth athletes improve their skills through high touch partnerships with PLL and D1 lacrosse players.

I improved Athes’ onboarding user completion rate by 17% by identifying high-friction screens, improving digestibility of each screen, and moving to a soft paywall tactic.

The Problem

Users were dropping due to high friction screens.

Our current onboarding flow had numerous choices, long forms, and didn’t help users feel like the app understood their needs.

As a result, we had an onboarding complete rate of 37%, which included many family and friends (prior to making a big marketing push).

The Problem

Users were dropping due to high friction screens.

Our current onboarding flow had numerous choices, long forms, and didn’t help users feel like the app understood their needs.

As a result, we had an onboarding complete rate of 37%, which included many family and friends (prior to making a big marketing push).

The Problem

Users were dropping due to high friction screens.

Our current onboarding flow had numerous choices, long forms, and didn’t help users feel like the app understood their needs.

As a result, we had an onboarding complete rate of 37%, which included many family and friends (prior to making a big marketing push).

How might we make the account creation process feel intuitive and quick rather than a chore?

Identifying Bottlenecks

The density of our onboarding made users frustrated.

From preliminary user research, we learned users felt frustrated during onboarding due to the high amount of information. High cognitive load ended being the killer of our completion rate and our 2 main culprits were the multi-category select page, and the too-many-inputs form.

Identifying Bottlenecks

The density of our onboarding made users frustrated.

Identifying Bottlenecks

The density of our onboarding made users frustrated.

Initial Idea Sketch

Information needed to be split up.

By separating information into their own pages, we could create lower cognitive load pages that were easy to understand and fill out, leading to higher speed of completion and hopefully, a higher completion rate.

Initial Idea Sketch

Information needed to be split up.

Initial Idea Sketch

Information needed to be split up.

Ok I know what you’re thinking but hey, it makes sense to me ¯\_(ツ)_/¯

+17%

Completion rate through onboarding.

Reduced

Cognitive load on a page-by-page basis.

Redesigned screens.

Final designs and results

Bite-Sized Pages

Single-goal pages that don’t annoy users.

By first understanding what information was crucial to the onboarding mission, we were able to reorganize our information requests into a progression that felt more comfortable for the user. Each “continue” button became a psychological win that built momentum.

Mentor Matchmaking

Matchmaker for our users.

Users originally had a high-friction task of choosing their mentor. To improve this experience, we included more lacrosse questions upfront, and made mentor recommendations based on those responses.

Additionally, we implemented video to mentor cards, so users could get a better sense of their coaching style and overall vibe.

Progressive Disclosure

Building trust before asking.

Asking for a phone number is a high-stakes request for any situation. Since it’s our primary login method, we moved this step further down the onboarding flow to appear when users reach a peak moment of momentum (getting matched with a mentor).

Paywall Transition

Transitioning to a soft paywall.

Our legacy paywall had a high bounce rate, requiring users to confirm Apple Pay before unlocking a free trial.

We pivoted to a soft paywall method by cutting out the paywall and adding it to high momentum moments in the user lifecycle (during first daily log in, and after a training session). Once the 7 days are up, users must make a choice.

Reflections & Next Steps

Further validation through Longitudinal A/B testing.

While our new onboarding received positive feedback an increase in completion rate, I would keep an eye on our Mixpanel analytics to observe where drop-off points begin to emerge, especially as we continue pushing on the marketing front and draw in larger crowds of users. Are there screens that could be moved, added, or removed? These would be the questions that I would look to validate to further increase completion rate, and eventually, conversion rate.

Further validation through Longitudinal A/B testing.

Deepening dynamic personalization.

Currently, our personalization logic matches users with a mentor based on position and skills. Due to company constraints, we prioritized shipping a low-tech solution to get more users into the app, and deferred deeper personalization. With that being said, personalization is crucial for a fitness app and I would continue exploring ways to fit the experience to a user’s specific lacrosse goal, not only within onboarding but throughout the app. For example, position-specific UI and predictive skills suggestion based on a user’s conversations in our AI mentor chat feature.

Deepening dynamic personalization.

Bridging Acquisition and Activation.

Onboarding is just a piece of the overall customer activation goal. My role at Athes also includes creating marketing assets for our Instagram page (31k+ followers). In addition to improving onboarding metrics, I’d continue working with the marketing team to improve distribution, ensuring alignment in our core value proposition and having fun ideating on strong visual and copy hooks!

Bridging Acquisition and Activation.

Next Project