The Challenge
A Family Argument With Users Caught in the Middle
Erialsoft’s fitness app was losing users fast. Navigation buried everything and the interface looked its age; people got frustrated almost immediately. The company itself was run by three family members, Dan, Sarah, and Cyndi, each with a very different vision for what the app should become.
I was brought in to lead the UX redesign from the ground up. The job had two halves: find out what users needed, and get three people with three visions to act on it in an oversaturated market.
Key Insight
People Quit Within Minutes of Opening It
I ran surveys (200 sent, 41 back), six in-depth interviews, and watched people work out at local gyms. The pattern that jumped out was speed: users gave up on the app within the first few minutes, well before they reached the content that might have kept them.
The content people wanted was already in the app, sitting a few taps too deep. New users churned before they ever surfaced any of it. That reframed the brief from “what should we add” to “what can we get out of the way.”
Competitive Analysis
Competitors Ignored Beginners and Education
I led a competitive analysis across five major fitness apps. Every one of them optimized for experienced users. None invested in beginner onboarding, and none used educational content to build confidence.
That left two openings: a difficulty system that met users at their level, and an educational layer of instructional videos and form guidance that no competitor offered.
.png)
Field Research
Social Integration Was the Top Unmet Desire
Field research at gyms and persona development kept pointing at the same gap: people treat fitness as social, and every app treated it as solitary. Users wanted to share progress and challenge friends. I built four personas around different fitness motivations, from time-pressed professionals to social gym-goers, and every single one prioritized connection.
Social was the loudest ask in all of the research. That made what happened next the hardest call of the project.
.png)
The Hard Call
We Deferred the Most-Requested Feature
The research validated three needs: findability, beginner onboarding, and social. A team of eight with one release to prove the redesign could not build all three well, and social is exactly the kind of feature that punishes a half-build. A community with nobody in it is worse than no community at all.
Sarah had championed social from the first workshop, and she wasn’t wrong; her instinct matched the research. But the churn data showed people gone long before any social graph could form. When we walked the funnel together the argument settled itself: fix the leak first, then build the things that need people to stay. Social moved behind onboarding and discovery, on purpose, with every stakeholder’s name on the decision.
From there the design phase ran on three agreed words: discoverable, progressive, social, with the third held as a constraint so we wouldn’t design ourselves into a corner. Wireframing went at the two flows that were bleeding users, onboarding and exercise discovery. Registration got cut to three steps, and I argued for stripping onboarding to the bone and letting the app earn complexity later. Testing backed that call.
What Testing Broke
Everyone Picks Intermediate
The first prototype asked people to self-select a difficulty level using identity labels: Beginner, Intermediate, Advanced, Expert. In testing nearly everyone chose Intermediate, including a man who’d never set foot in a gym and a woman who competed. Nobody wants to call themselves a beginner in a form field.
The fix kept the levels but anchored each one in behavior, a plain line under each label like “I work out 0–1 times a week.” People will happily match themselves to a habit they’d never match to an identity. The second round of testing sorted people accurately.
The Solution
A Fitness App That Meets Users Where They Are
The final design led with three moves, each traceable to a specific research finding:
Findable in Seconds. Auto-complete search, flattened navigation, and visual exercise guides replaced the buried layout; people reached what they wanted in seconds. The direct answer to the churn finding.
Progressive Difficulty. Users self-select a level during onboarding, each anchored to real behavior, and the app adapts its content and language to match. This one went at the beginner gap from the competitive analysis.
Customizable Focus Areas. A “Focus” feature lets users prioritize training goals like muscle gain, cardio, or flexibility, and the app rearranges itself around the choice, which gave people a reason to stay past the novelty period.
Social, the fourth move, shipped after launch by design, once retention showed people were staying long enough to meet each other.
From Redesign to Acquisition
improvement in user satisfaction with layout and feel
improvement in navigation effectiveness
increase in user happiness with exercise options
increase in signup rate
improvement in retention over 150 days
The satisfaction and navigation numbers come from re-running the original survey instrument with post-launch users. Signup rate and 150-day retention come from app analytics, measured against the pre-redesign cohort.
The six months after launch went to what we’d deferred: the social layer and deeper focus-area features. The three stakeholders who’d disagreed on every detail stayed aligned through it, because the sequencing was their decision as much as mine. Erialsoft was later acquired by a larger competitor. I won’t claim the redesign closed that deal, but it put the product in materially better shape going into it.
Reflection
What I Learned
On stakeholder alignment: Three people with three visions never fused into one vision. What worked was moving each argument onto ground where evidence could win it: personas, funnel data, churn numbers. The social deferral proved the approach, when the feature’s strongest advocate signed off on delaying her own idea.
On research scope: With 41 survey responses from 200 sent, the quantitative sample was limited. If I were to do this again, I’d invest more in recruiting to hit statistical significance, and I’d push harder for real usage data instead of leaning almost entirely on what people told us.
On adapting to the user: The difficulty-level system was the right call, and it stopped short. It adapted the content; a further step would have adapted the whole interface to how experienced someone was. That’s where I’d take the next iteration.