Protected Content - Enter password to view case study
Activating High-Value Learning at Scale
How we turned a platform feature used by 5% of coachees into a commercially differentiating learning system.
CoachHub
SaaS, Personal Development, Human Resources
Product Designer
Product Manager, UX Research, Learning Experience, Engineering
2 Quarters, phased delivery
Context
CoachHub is a leading B2B2C enterprise coaching platform connecting employees with professional coaches globally. Alongside live coaching sessions, the platform offers the Academy, an on-demand library of articles, videos, exercises, and reflections designed to extend the impact of coaching into employees’ continuous development between sessions.
The Academy was central to CoachHub’s value proposition. A year after launch, it had very low engagement, making it not utilised by the most of the people it was built for.
Challenge
Clients (HR buyers users) had dashboard visibility into their program’s learning activity. For most clients, that section was thin. Sales were in a difficult position using the Academy as a value driver in renewal conversations. The Academy was in most pitch decks, but the usage data wasn’t backing it up.
The content wasn’t the issue. 95% of coachees (users who receive coaching) who completed an activity rated it good or amazing. The Academy worked exceptionally well for the small ratio who reached it. The vast majority never did. Only 5% of coachees had ever completed an activity across the platform.
The PM flagged it with clear urgency: Q4 was CoachHub’s heaviest renewal period. Meaningful improvement was needed for this quarter.
Research
Before talking to anyone, the UX researcher and I reviewed data we have from our analytics platforms. Coachees were landing on Browse, scrolling, and leaving without clicking anything. Search usage was high but coachees weren’t clicking through after searching. Intent was there. Outcome wasn’t.
We ran 8 coach and 6 coachee interviews over four weeks. The pattern that emerged was consistent: coachees didn’t know what to look for. They arrived from a coaching session needing guidance, not a library to navigate alone. AI recommendations pulled from incomplete data nobody trusted. In structured programs, coaches were spending session time covering foundational knowledge coachees could have learned independently.
The clearest signal: completion of coach-assigned content was 95%. When a trusted person said “this is relevant to you, right now” coachees engaged completely.
The Academy wasn’t failing because the content was wrong. It was failing because 95% of coachees had no mechanism connecting them to content relevant to where they actually were in their coaching journey. Relevance required human context and the platform had no way to scale that.
From Insight to Direction
The PM and I had different reads after the research readout. The PM saw activation as the priority. Reduce overwhelm, move the number before Q4. I was more focused on the relevance gap. Better navigation wouldn’t tell a coachee what was valuable for them.
We evaluated three directions. Fixing navigation ruled out quickly. Coachees weren’t failing because of the interface, they were failing because they didn’t know what to look for. A My Learning personalised space looked promising until we mapped what would feed it. The same broken signal. A better container for broken data is still broken data.
We both landed on coach-curated bundles. Coaches building thematic content libraries, assignable to multiple coachees. It balanced both positions: scale and relevance. The 95% completion rate was proof the mechanism already worked.
Before committing to it, we shared the concept with engineering for a feasibility read and with the Learning Experience team to sense-check the content angle. Both conversations came back with significant concerns.
Engineering flagged that giving coaches a creation capability meant building four or five interconnected systems on a legacy codebase, far beyond what the Q4 timeline allowed. Learning Experience worried about quality control at scale: coaches optimising for personal preference rather than pedagogical coherence could create incoherent bundles that CoachHub couldn’t audit or fix. Both concerns were legitimate. We accepted the pushback. But we weren’t willing to let the insight go.
The Pivot
I facilitated a constraints-based workshop with the PM, Learning Experience lead, engineer lead, and a behavioural scientist. One question: given what we can’t build, what’s the simplest version of human-guided relevance feasible for Q4?
The PM sharpened the commercial focus were he emphasised that Program-level engagement had more renewal impact. The behavioural scientist confirmed the opportunity: the foundational knowledge gap was consistent across every instance of the same use case. CoachHub already knew what each program needed. Learning Experience saw the structural solution, module-based, curated by them, not coaches. Engineering confirmed it was buildable.
Primary direction: structured Learning Paths aligned to CoachHub’s use cases. The individual-level problem was deprioritised. Bigger impact at program level before Q4. But during requirements work, something became clear: the module being designed for Learning Paths was independently useful. A coach could assign it after any session. A coachee could find it browsing independently. It was the curated bundle mechanism, now curated by Learning Experience rather than individual coaches.
That became Collections. The individual-level answer found inside the program-level solution.
Solution 1: Collections
Collections bundle 4 to 7 activities around a single theme, assignable by a coach after any session or discoverable by coachees browsing the Academy independently. For coaches without structured programs it’s the complete solution, and for coachees in Learning Paths (the program level solution) it’s the individual layer alongside the program structure.
Three design decisions shaped the final product. A stacked card shape to differentiate Collections in browse was blocked by the existing grid system, which required all cards to share the same dimensions, so I solved it with a folder tab graphic overlaid on the card illustration, giving coaches and coachees a clear visual signal without breaking the layout. Progress indicators showing completion state on Collection cards were deferred to a follow-up release, because the V1 hypothesis was about coach adoption rather than coachee motivation, and the legacy architecture made real-time completion state expensive to surface at scale. One layout served both Sequential and Playlist Collection types, differentiated through interaction states rather than separate components, with locks communicating order for Sequential and checkmarks communicating freedom for Playlist.
Prototype testing changed the layout. A vertical prototype sent Playlist users scrolling back to the intro repeatedly because they needed the Collection context visible while choosing which activity to start, not just at the top of the page. That insight led to the split layout, with the Collection intro fixed on the left showing the title, description, activity count, and estimated time, while the activity list scrolls on the right, keeping context always visible without needing to scroll. Catching this early saved development rework.
The final design shows activity cards changing state visually as coachees progress, unlock and complete activities. Sequential Collections locking uncompleted activities until the previous one is finished, and Playlist Collections keeping everything accessible from the start. For coaches, the assignment flow didn’t change at all, with Collections appearing in browse and search alongside individual activities and assignable in the same way, because the friction reduction came from the content unit itself rather than a new workflow.
Solution 2: Learning Paths
Learning Paths are structured learning journeys aligned to CoachHub’s coaching use cases such as Transformational Leadership and Organisational Transformation, built from Collections as modules and configured during program setup. Curated by the Learning Experience team, they pre-teach foundational knowledge before sessions rather than during them, freeing coaching time for the deeper personalised work it’s designed for. Learning Paths are only visible to coachees enrolled in a configured program, keeping the experience relevant and uncluttered for everyone else.
Building on Collections as the foundational content unit, the central design challenge was showing the full program structure clearly, where coachees were across all modules, what was complete, what was available, and what was still locked, without adding a second navigation layer to an interface that already had one. The natural solution was a module sidebar, a pattern common in platforms like Coursera or LinkedIn Learning, but the Academy already had a left sidebar navigation and adding a second one would have required refactoring a global layout component shared across every Academy view, disproportionate for an MVP. The solution was horizontal module cards at the top of the page, all modules visible in a scrollable row showing name, activity count, and status, using an existing UI pattern with no architectural changes and keeping the Academy navigation always visible. The tradeoff was less screen real estate than a sidebar would provide, and modules truncate on smaller screens, but for a pilot validating whether the program structure worked at all, that was an acceptable constraint.
I designed distinct visual states for each module case, active, inactive, in progress, locked, and complete, so coachees could see the full shape of the program and their position within it from the module row alone.
A six-month enterprise pilot validated the model, with coachees engaging consistently across the full program duration and generating exactly the kind of program-level completion data that had been missing from HR buyer dashboards.
Impact
Collections and Learning Paths together moved the Academy from a platform feature coachees rarely reached to a curation system that generated measurable engagement and commercial outcomes at both individual and program level.
Coach assignment rate increased from 25% to 28%, organic adoption signalling that reducing friction can shift coach behaviour at scale. Within the first quarter after Collections launch, overall Academy engagement lifted from 5% to approximately 5.5%, a 10% relative increase driven purely by increased coach-assigned learning, with no changes to the browse or discovery experience.
Learning Paths pilot completion hit 41% against a pre-defined 30% threshold, well above the 10 to 20% industry benchmark for optional corporate learning programs running over six months. Post-pilot coach feedback indicated reduced session time spent on foundational content, leaving more room for the personalised depth work coaching is designed for.
For the first time, HR buyers had program-level completion data to point to in renewal conversations. The Academy section of their dashboard was no longer thin. Learning Paths became a marketed commercial feature of the CoachHub Academy, directly tied to the use cases enterprise clients were already buying, giving sales a learning story aligned to company goals rather than a generic content library.
28%
25% → 28% Coach assignment rate, organic adoption with no requirement or nudge
41%
41% Learning Paths pilot completion against a 30% pre-defined threshold
5.5%
5% → ~5.5% Overall Academy engagement lift in the first quarter after launch
“The Learning Paths reinforce essential leadership behaviours like accountability, allyship, curiosity, awareness, empathy, and collaboration. These behaviours are vital for building trust within teams and driving business success, and the collections highlight these critical traits we wish to be exemplified within our teams.”
Enterprise Pilot Client
Talent Development Partner
Reflection
This project started with a big, fuzzy problem space and ended with solutions nobody expected at the beginning.
At first glance, the obvious answer was to fix activation, reduce overwhelm, improve navigation, make the Academy easier to reach. But the more effective solution was to leverage the human relevance already built into the coaching relationship and the program structure clients were buying.
Relevance in this platform operates at two levels. At the individual level, coaches have session context no recommendation engine can replicate. At the program level, the use cases clients come to solve define the foundational knowledge every coachee needs before coaching can go deeper. Designing for both levels with one coherent system rather than two isolated features is what I’m most proud of in this project.
What made this project interesting was navigating three users with intersecting but distinct needs. HR buyers needed measurable evidence to justify renewals, coaches needed tools that reduced friction and helped them guide coachees effectively, and coachees needed learning content relevant to their specific journey.
If I were to do this again, I’d validate the Collections direction with coaches earlier, not just through the 95% completion data, but through direct conversations about whether a curated bundle model would fit their workflow before committing to design.
Looking ahead, I want to explore reducing overwhelm on the coachee side by revisiting the onboarding journey and thinking through a My Learning space, and I’d want to look deeper into ways to empower coaches to shape the coachee journey more effectively. Coaches are the thread connecting relevance to engagement in this platform, and the more the system supports them, the more value reaches coachees.