Oasis Market - Buyer Journey UX Evaluation
My Role
UX Researcher, Team of 5
Deliverables
UX Evaluation Report
Tools
Useberry, Qualtrics, Statistical Analysis
Use case
Evaluate the end to end buyer journey on a hot tub marketplace platform, from homepage to seller contact, to determine whether Oasis Market was ready to open to real traffic. The evaluation combined moderated usability testing with statistical hypothesis testing to move beyond opinions and identify, with data, exactly which parts of the journey were preventing browsers from becoming buyers.
The Challenge
Oasis Market set out to be the first platform to unify the fragmented North American hot tub market into a single, trusted online marketplace. Before opening to real traffic, the team needed real buyers to interact with the live site and let the data reveal what was working and what wasn't.
Understanding the problem
The evaluation was structured around six research questions and hypotheses, grouped into three areas critical to conversion: whether the homepage communicated the platform's value proposition, whether guided filtering reduced cognitive load during product discovery, and whether the product detail page built enough trust for users to contact a seller and buy.
Methodology
16 moderated remote usability sessions were conducted, each 30 to 60 minutes, on the live Oasis website through Useberry. Participants completed a 4 step buyer journey covering homepage impressions, product search with and without filters, product detail review, and seller contact. A mixed methods approach combined think aloud protocol, SEQ, SUS, NPS, CSAT, and statistical hypothesis testing through t tests.
Key Findings
The Value Proposition Was Invisible, Not Absent
87.5% of users, 14 out of 16, only partially understood what Oasis offered after visiting the homepage. The platform's strongest differentiator, a quiz to help buyers find the right hot tub, was found by only 1 participant. One user asked directly: "I'm just questioning, is it a marketplace or is it focused on hot tubs?"
Filters Underperformed Search, Statistically
Guided filtering showed no statistically significant improvement over plain keyword search, p equals 0.32. Task success dropped 18.75% when filters were used instead of search, and user uncertainty doubled. The root cause was concrete: the single most requested filter criterion, seat capacity, did not exist on the platform.
Trust, Not Usability, Was Blocking Conversion
This was the study's critical severity finding. 12 out of 16 users raised trust concerns without being prompted. Seller credibility scored 4.56 out of 7, below acceptable threshold, and messaging a seller produced no statistically significant increase in purchase confidence, p equals 0.86. Likelihood to Recommend scored an NPS of 0, while Likelihood to Purchase scored an NPS of negative 13, showing that users liked the idea of Oasis far more than they trusted it with their money.
The Overall Usability Score Confirmed the Pattern
The platform scored 66 out of 100 on the System Usability Scale, just under the industry benchmark of 68. The product was usable, but not yet reliable enough to convert hesitant buyers.
Recommendations
Findings were mapped on a severity and frequency matrix and translated into a phased roadmap.
Immediate action, 0 to 30 days: Fix the zero result filter bug and add seat capacity as a primary, visible filter.
Short term, 1 to 3 months: Redesign the homepage to lead with a clear value proposition, elevate the quiz to the primary call to action, and enrich seller profiles with transaction history, response time, and authorization status.
Strategic, 3 to 6 months: Build a platform wide trust framework covering guarantees, seller responsibilities, and dispute handling, then re test against this study's baseline.
Takeaways
The project demonstrated that usability and trust are not the same thing, and that a platform can be technically sound while still failing to convert.
Usability is not enough. Users did not get lost navigating Oasis, they got unconvinced by what they found.
Trust must be designed, not assumed. A verified badge and a single review were not enough reassurance for a high value purchase.
Data reframes the problem. Statistical testing revealed that features assumed to help, like filters and messaging, were not moving the metrics that mattered.
Prioritization drives impact. Mapping findings by severity and frequency turned a long list of issues into a clear, phased action plan.