The Setup
StyleHaven, a fashion e-commerce brand, relaunched their checkout with a new, streamlined UI. The result: cart abandonment climbed 11 points, from 68% to 79%.
The conversion team blamed UX. UX blamed pricing. Marketing blamed the traffic sources. The CEO wanted answers by Friday.
They had 6 usability test recordings — think-aloud sessions where users narrated their checkout experience — and funnel data from 25 sessions. This is what the integrated analysis found.
The Data
What we had:
- 6 usability session transcripts (Karen, Derek, Samantha, Megan, Chris, and one more)
checkout_funnel.csv: 25 sessions tracking cart value, country, time on each checkout stage, completion status
| Metric | Value |
|---|---|
| Overall cart abandonment | 79% |
| Avg time on payment page (abandoned sessions) | 114 seconds |
| Avg time on payment page (completed sessions) | 52 seconds |
| Completion rate — US users | ~35% |
| Completion rate — international users | ~5% |
The Investigation
Qualitative Coding
Five clear themes emerged from the sessions:
- Price Shock — unexpected shipping costs, currency confusion at checkout
- Payment Friction — missing Apple Pay, PayPal, preferred payment methods
- Trust Issues — return policy concerns, security hesitation
- Delivery Time Issues — slow shipping estimates
- Smooth Experience — sessions with no friction (the comparison baseline)
Karen's session was the opening statement:
"Wait, $14.99 for shipping? That wasn't mentioned anywhere."
"I feel tricked. Like they hid the shipping cost on purpose."
She was buying a dress. She'd already entered her address. The shipping cost appeared at the final payment screen — for the first time in the entire checkout journey.
Derek's session revealed a completely different failure mode:
"I don't see Apple Pay... There's no PayPal?"
"I'll just do this later when I have my wallet."
Derek was ready to buy. He just couldn't pay the way he wanted to. He said he'd come back later. He didn't.
Chris's session exposed the international problem:
"It's saying $79 USD... what's that in CAD?"
"Feeling lowkey scammed right now."
The site hadn't detected his Canadian location. He assumed $79 was in Canadian dollars. When he realised it was USD — roughly $108 CAD — he left.
Samantha raised a fourth concern, lower priority but worth noting:
"Customer pays return shipping... That's not great."
Return policy anxiety surfaced as a hesitation point, though not at the same severity as the top three.
The Joint Display: Where It Came Together
The Split View: Words and Numbers, Side by Side
FableSense AI's Split View joint display put every coded quote on the left and the dataset statistics on the right — simultaneously. No switching tabs. No exporting to spreadsheets.
Filtering to "Unexpected shipping cost" showed:
- Karen: "$14.99 for shipping? That wasn't mentioned anywhere"
- Chris: "It's saying $79 USD... what's that in CAD?"
- Two additional matching segments
The right panel simultaneously showed:
- Payment stage abandonment: 91% for sessions coded with this theme
- Average time on payment page: 114 seconds
- Average cart value at abandonment: $67.40
These users weren't confused by the interface. They were calculating. Shipping plus tax plus exchange rate. And deciding it wasn't worth it.
Switching the filter to "Missing payment methods" surfaced the starkest finding of the entire analysis:
- Derek: "I don't see Apple Pay... There's no PayPal?"
- Derek: "I'll just do this later when I have my wallet."
Completion rate for sessions coded "Missing payment methods": 0%.
Not low. Zero.
The Integration Matrix: Confirming the Pattern
The Integration Matrix confirmed these were systematic failures, not isolated incidents:
| Theme | payment_stage_abandonment | completed_purchase |
|---|---|---|
| Unexpected shipping cost | Strong positive | Strong negative |
| Missing payment methods | Strong positive | r ≈ -1.0 |
| Currency confusion | Strong positive | Strong negative |
| Smooth experience | None | Strong positive |
The Recommendation
Three specific, prioritised fixes — each backed by the integrated data:
1. Show shipping estimates on the product page, not at checkout Karen said "I feel tricked." The fix isn't cheaper shipping. It's showing the cost before users invest time in the checkout flow. Transparency reduces abandonment more than discounts. This is a product page change, not a checkout redesign.
2. Add Apple Pay and PayPal immediately Zero completions from users who couldn't pay their preferred way. This is not a nice-to-have. It's a hard blocker for a significant segment of users. The engineering lift is well-established; the revenue justification is now clear.
3. Auto-detect location and display local currency Every international session in this dataset ended in abandonment. The fix is one geolocation lookup and a currency conversion display. The revenue impact is disproportionate to the engineering effort.
Lower priority: Simplify the return policy language. Samantha's concern was real, but she was still considering the purchase — unlike the users blocked by the top three issues.
The Outcome
Before this analysis, the team had competing theories and no shared evidence. After it, they had three ranked recommendations, direct user quotes supporting each one, and funnel data quantifying the cost of each issue.
The conversion team and UX team stopped arguing. They were looking at the same data.
The qualitative said: here is exactly what users are thinking at the moment of abandonment. The quantitative said: here is how often this happens and what it costs.
That combination is what makes the recommendation actionable rather than just interesting.