Digitizing a coaching system: what the team hadn't agreed on
DESIGN STRATEGIST | CREATIVE SCIENCE FOR MAXONE & OWN IT | 2022
OWN IT was a one-to-one coaching system built around controllables that affect performance outside training hours. It worked, and it had reached the ceiling of what one-to-one can serve. MaxOne, which sold into college sports programs, adopted the system and set out to deliver it digitally at scale. They asked us how to make the app work. The interviews showed the team had not agreed what the app was for.
Approach
Two phases: research, then a pair of decision workshops. I ran the internal interviews and led synthesis, and supported the audit, the positioning work, and the model.
Internal interviews
Stakeholders across coaching, product, technology, sales, and leadership, on the OWN IT system, its value, and their athletes. The guide opened with someone's own role and responsibilities before moving to opinion, since people talk about facts more readily than beliefs.
Psychological positioning
We mapped the product against eight peer platforms on two axes: whether a product works on the individual or the group, and whether it runs on intrinsic or extrinsic motivation. It showed the client which psychological space they were competing in.
UX and UI audit
We read each screen through behavioral science: what the experience asks of a user, and what the research says about how people respond.
Two Planning Forward workshops
The first surfaced the research and put the disagreements on the table. The second built a model and set out what to test.
Eight peer platforms placed by who they work on and what motivates the user.
Where the team disagreed
The interviews surfaced incompatible beliefs about the product that had never been said out loud. In synthesis I structured them as tensions. Each paired two things stakeholders had said, stated what each position implied, and asked the room to place themselves between them.
The coach
Augment the coach
"The app's goal is to prop up and augment coaching through data and insights."
Replace the coach
"Ideally we would make coaching all digital and automated, to reach more athletes and scale."
The system
Stay true to OWN IT
"There is opportunity for the entire OWN IT experience to live in the app, we just have to translate it to digital."
Deviate when athletes need something else
"We have freedom to deviate if new user needs arise outside of what OWN IT provides."
The experience
One to one
Replicate the coaching relationship as it works in person, for each athlete.
One to many
"Take their system and do it with millions rather than hundreds."
The data
Create understanding
"Making complex data simple and relevant to the athletes so they understand the impact of the rest of their day."
Create action
"Data is the means, but we are in the business of real habit change."
Ownership
Consensus
"It is vital that we have alignment to make sure we are all consistent with the product strategy."
A single owner, for speed
"No one owns the whole UX experience. We want to take product development more seriously."
Run live in the workshop, the exercise put the disagreement in front of the people who had to resolve it.
Reading the app through behavioral science
Each observation followed one shape: what the experience does, and what to consider about how people respond.
The app assumed that giving people data changes what they do. The research does not support that. In person, a coach interpreted the data and told the athlete what it meant. The app kept the data and removed the coach.
The audit in five themes
Select a theme to see the principle behind it, how it surfaced in the product, and the screens where it showed.
The model
The deliverable was PATA: Plan, Act, Track, Analyze, entered through onboarding and running as a loop. It builds on the habit loop of trigger, action, and reward, adapted to how OWN IT works.
Each stage carries a behavioral principle, a way to apply it, and a named risk. Onboarding leans on autonomy, framing setup as a request rather than an instruction. Planning uses the endowment effect, since co-creating a plan increases commitment to it. The coach appears as a commitment device. Tracking keeps habits few and specific. Analysis uses picture superiority, giving one metric with a color and an image instead of a wall of numbers.
Each stage also names what could go wrong and how to test for it: extrinsic rewards crowding out intrinsic motivation, reactance when guidance reads as paternalistic, present bias making onboarding overpromise, information avoidance when the data disappoints.
We put the caveats on the slide. Not final, not validated, broad strokes. The framing came from George Box, that all models are wrong but some are useful. A model is worth having when it is specific enough to be proven wrong.
Watch out for friction
We put the caveats on the slide. Not final, not validated, broad strokes. The framing came from George Box, that all models are wrong, but some are useful. A model is worth having when it is specific enough to be proven wrong.
The path forward
The recommendation was to build on the model, understand the human experience before the digital one, do the things that do not scale first and automate later, and get in front of potential users as soon as possible.
The last one carried the most weight. Thirteen internal stakeholders had told us what athletes need, want, and are unaware of. We mapped those beliefs into a pyramid and labeled it: your user, as you understand it. Inside the pyramid, commonly held views. Outside, one person's. Nobody had asked an athlete.
Read this as belief, not evidence. Every item came from an internal stakeholder describing athletes. Inside the pyramid sat the views held across the team. Outside sat one person's. No athlete was interviewed.
Where it landed
The engagement ended at handoff, which is how these ran unless we were also building. I do not have visibility into what the team decided or why.
The company has since rebranded as Coya and repositioned beyond college athletics. The product today lands on a specific side of each tension the interviews surfaced. Coaching stayed central, with human-first named as a company value and a coach conversation required before anyone joins. The eight controllables carried over intact. AI appears as a companion layered onto coaching rather than a replacement for it. The marketing now leads with the argument that data alone does not create change.
I cannot claim those decisions came from this work. Time has passed and the company changed hands and direction. What it does suggest is that the tensions were the real decisions.
Reflection
When a team disagrees about what a product is for, reporting the average of their views is no use to them. Structuring the findings as tensions turned interview notes into a decision the room could make. That is the move I would carry into any alignment problem.
We were asked how to make an app work. The interviews showed the client had not agreed what it was for, and had never spoken to the people it was for. The question a client brings is worth testing before answering it.