I led the redesign of a mental-training product that had grown one feature at a time. The work was not to make it cleaner. It was to decide what the product was, for whom, and how the team could keep building it.
The decisionOrganise the experience around each role's intent, not the feature list the product had accumulated.
[01.00]>Read the product
At first, it looked like a navigation problem.
People were getting lost, labels overlapped and equivalent actions behaved differently. The obvious response was to simplify the interface.
That reading was useful, but incomplete. When I mapped the recurring journeys, the same problem appeared underneath every screen: a single product was trying to serve three different jobs without making the boundaries visible.
It was a starting point, a mode selector, a progress view, a communication surface and a shortcut to daily tools. Each new feature added another decision before an athlete could begin.
Navigation was where the fragmentation became visible. Underneath it, one product was serving three jobs without a shared model. Every new feature added decisions for users and exceptions for Design and Engineering.
People were getting lost in navigation.
Three roles were sharing one structure without sharing one job.
[02.00]>Choose the structure
At that point, I decided the team needed something more useful than a new set of screens. I turned the diagnosis into four working principles we could use in critiques to decide what stayed, what moved and what we deliberately did not build.
Make the next useful action obvious.
Earn personalisation after the first value.
Organise around what each role is trying to do.
Reuse product decisions, not only components.
Then I changed the organising idea. The shared system stayed underneath, but the experience above it changed according to the job each person came to do.
Defines the method and follows progress over time.
Turns the method into plans, groups and daily guidance.
Prepares, practises, reflects and comes back.
I started with activation because it exposed the product logic quickly. The old onboarding front-loaded personalisation, so we shortened the entry and moved profile questions into a staged sequence after the first meaningful session.
Less tailored on day one.
↔More chance of reaching value while motivation was still high.
[03.00]>Build, test, correct
Then I got the work out of the presentation and into delivery. Foundations came first, then the inherited flows, then role architecture and activation. I ran critiques with two product designers, brought engineering constraints in early and kept final decisions tied to the four principles.
The plan moved. A coach flow exposed assumptions in the athlete experience; an implementation constraint changed how progressive disclosure worked. We updated the sequence instead of protecting the original deck.
The primary route was clearer, but exploratory users lost context and experienced the product as restrictive.
The next action stayed obvious. Secondary information and tools returned through progressive disclosure, available when the moment called for them.
We separated daily action from long-term progress, made state and hierarchy explicit, and moved repeated decisions into shared foundations. This was where the product redesign and design-system work met.
Read the design system case
I translated the same principles into a visual language for product and communication: a recognisable wordmark, restrained iconography, an energetic but controlled palette, and photography centred on athletes rather than abstract wellness.
I carried the product language into the promotional site: the same hierarchy, tone and visual cues now introduce the value, show the experience and lead athletes towards download.
[04.00]>Make it hold
One system now connects the athlete's daily practice, preparation, training plan and conversation with a coach.
I did not treat prototype validation as business impact. It gave us confidence in the direction and a concrete measurement plan for beta.
athletes chose the intended starting path without prompting
fewer decisions before reaching a first session in the redesigned flow
aligned with product and domain leads in one architecture
defined for beta: comprehension, activation, completion and return
I set the architecture and principles, divided the work into role, activation and system workstreams, and gave each designer a clear decision space.
I used critique to connect local choices back to the product model, partnered with Engineering before flows hardened, and made the call when user needs, scope and buildability pulled in different directions.
The most important leadership decision was not defending the first solution. It was making the reasoning visible enough for the team to challenge it and continue the work.
The redesign became credible when the product stopped needing us to explain how its pieces fitted together.
> MORE WORK