Skipper Hospitality Booking Engine · Product Design
Uncovering why customers struggled in the portal — and fixing what mattered most
Research-led
Usability testing
Customer portal
Retention
Sole designer
I led an end-to-end usability study of Spot & Tango's customer portal — designing the research protocol, facilitating 7 sessions across 5 departments, and synthesizing findings into a prioritized roadmap. Three improvements shipped, including a navigation change that directly reduced cancellation rates.
7
Usability sessions
Across 5 departments
3
Improvements shipped
From research to production
↓ CX
Cancellation tickets
Post nav update
↑ Saves
Cancel save rate
Portal vs. phone
01 · Context
About the project
Spot & Tango's customer portal is the primary self-service tool for subscribers — covering rescheduling, adjusting portions, managing add-ons, and cancelling plans. Getting these flows right is directly tied to subscriber retention and lifetime value.
Despite the portal's importance, it had never been formally usability tested. CX team feedback suggested customers regularly struggled with core tasks — but without structured research, we couldn't prioritize what to fix. I proposed and led this study as part of an evergreen retention initiative.
My role
Sole Product Designer
Research questions
1
Where does the portal break down?
What are the main friction points users encounter?
2
Which tasks cause the most confusion?
Where do users drop off, and why?
3
What would reduce CX ticket volume?
Which improvements map to measurable retention outcomes?
Business case
Improving the portal was part of an evergreen retention project — a more usable portal would reduce churn by keeping customers from reaching out to cancel, and by making the experience feel worth continuing.
Phase 01
Research Setup
Designing the study
I structured the study as a moderated remote usability test using a staging portal. Participants were internal employees from Marketing, Ops, Strategy, CX, and Finance who were unfamiliar with the portal's day-to-day usage — giving us fast access to engaged testers without external recruitment lead time.
I wrote the test script, facilitated all sessions over Zoom, and synthesized findings in Notion. I also used Data Dog session recordings alongside live sessions to triangulate behavioral patterns across a wider sample.
Participants were given realistic scenarios rather than direct instructions — so we could observe natural navigation choices, not just task completion.
Participants
Mike
Marketing
Bailey
Strategy
Georgia
Marketing
Kieran
Finance
Shohini
Ops
Sara
CX
Lexi
Finance
Test tasks
1
Rescheduling flow
Reschedule your dog's next order to a future date
2
Portion adjustment
Adjust the amount of food your dog gets
3
Cancellation path
Cancel a dog's plan
4
Recipe change
Change your dog's recipe
5
Add-on process
Add a snack or supplement to your dog's plan
6
Removal process
Remove a snack or supplement from your dog's plan
Phase 02
Findings
What we learned
The portal was functional — every participant eventually completed every task. But the average error rate of 1.3 errors per task was nearly double the 0.7 industry benchmark, and qualitative feedback consistently pointed to the same themes: unclear navigation, missing feedback, and a mismatch between where users looked for things and where they actually were.
Cancellation was the starkest finding. Every single participant struggled to find it. This wasn't just a UX issue — it had a direct business cost. Customers who couldn't find cancel on the portal called CX instead, where cancellation rates were significantly higher.
Usability metrics
100%
Task completion rate
All tasks eventually completed
1.3 / task
Error rate
vs. 0.7 industry benchmark
Cancel · 3:25
Slowest task
Deep path + multi-step friction
Remove snack · 1:00
Fastest task
Simple, linear flow
Synthesis insight
The portal was designed for the happy path. Every major pain point involved a moment where users needed more context or flexibility than the current design provided.
Recommendation stack
Phase 04
What Shipped
Three changes from research to production
Not every recommendation made it into the immediate sprint — but the three that did were most directly tied to retention and CX ticket reduction.
Shipped 01
Account option added to navigation bar
Findings
Every participant struggled to find cancellation — buried behind a profile icon no one associated with plan management. Customers who couldn't find it on the portal called CX instead, where they were far more likely to follow through with cancellation.
solution
Added a visible 'Account' option to the main nav bar, surfacing plan management — including cancellation — at the top level. Removed the need to know that cancel lived under a profile icon.
Outcome
CX ticket volume for cancellation decreased post-launch. More importantly, the save rate for customers cancelling via the portal improved relative to phone — because the portal's save flow is more consistent than what CX was delivering.
Phase 05
Outcome & Reflection
Research that directly moved retention metrics
Three changes from this study shipped into production. Each was directly traceable to a specific finding — making this one of the clearest examples of research-to-impact work in my portfolio.
7
Sessions facilitated
Across 5 departments
3
Changes shipped
From 16+ recommendations
↓ CX
Cancellation tickets
Post nav update
↑ Saves
Portal cancel save rate
Remaining backlog
Clarify charge vs. delivery date in reschedule flow
Show live price/bag updates on portion size change
Make product cards in Shop feel visibly clickable
Improve promo code visibility and real-time price updates
Explore sidebar navigation for mobile — structural fix
What this project taught me about research leadership
This was the first study I designed and facilitated entirely on my own. Before this, most of my research had been in support of a specific feature. This was broader — what's wrong with the portal, and what should we fix first? That open-endedness required a different kind of discipline.
The most valuable lesson was about the gap between task completion and task experience. Every participant completed every task — which could have been reported as "the portal works." But the error rate, the time-on-task data, and the verbal frustration told a completely different story. Numbers without qualitative context are dangerous.
What I'd do differently
I'd recruit external users rather than internal employees for future rounds. Internal participants were engaged and articulate, but they brought more product intuition than a typical customer — which likely masked some navigation failures real subscribers would hit harder.
What surprised me
Making cancellation easier to find improved retention. I expected that surfacing cancel more prominently would increase churn. Instead it pushed more customers through the portal's save flow, which outperformed CX's delivery. Friction isn't always protective.
The B2B SaaS insight
In subscription products, the post-purchase experience is as important as acquisition. This portal study had more retention impact than most acquisition-focused work I'd done — because keeping one customer is cheaper than acquiring a new one. Research that connects UX to retention metrics is the kind that gets resourced.
What came next
Add-ons in booking flow (early check-in, special requests) — highest-priority v2 feature from testing
Multi-CRS integration to support additional hotel clients beyond Journey East Hampton
Analytics dashboard for hoteliers to monitor conversion and drop-off in real time
Expanded design system to support 5+ hotel brands without per-client rebuilds