The Client
An Australian fintech — a B2B SaaS platform for digital expense and payments management. Subscription revenue model with user acquisition via a multi-step registration funnel. High-intent audience: business owners and finance managers.
The Challenge
Qualified traffic was dropping off across an 11-step mobile funnel (vs 6 steps on desktop) before completion.
Mobile funnel (11 steps)
- Journey 83% longer on mobile than desktop (11 steps vs 6).
- Average Frustration Score: 80/100 mobile vs 50/100 desktop.
- Mobile: all 5 initial sessions scored Critical (66–100 frustration range).
- Estimated completion rate: 60% mobile vs 80% desktop (−20 percentage points).
- Primary drop-off: Step 11 — Phone Verification (40% abandonment).
- Step 7 (Redemption Options): 20% drop-off — multi-select confusion.
- Step 9 (Calculation Screen): 20% drop-off — unclear value proposition.
- Step 10 (Account Form): 20% drop-off — hesitation / privacy concerns.
Desktop funnel (6 steps)
- Step 5 (Account Form): 40% drop-off — referral-code error triggering on all users.
- Step 6 (Phone Verification): 60% drop-off initially (later stabilized to 20%).
Technical issues discovered
- CLS Score: 0.95 — 10× worse than Google’s 0.1 threshold.
- Form not responsive at 374px viewport width.
- Plus/arrow button not clickable at 374px.
- Accessibility-tool conflict at Step 9.
The Strategy & Methodology
Qualitative session analysis using ContentSquare — 80 to 100 sessions analyzed every day. Methodology: frustration scoring → funnel-step mapping → drop-off root-cause identification → prioritized recommendations (P0/P1/P2) → daily consolidated stakeholder reports.
Tool Stack: ContentSquare (session replays, heatmaps, funnel analysis, frustration scoring), Hotjar, GA4.
Execution
- Daily analysis across the engagement window. 33+ daily Excel analysis files + 33+ matching Word stakeholder reports delivered.
- 4 professional 300 DPI charts created and embedded.
- Both platforms later standardized to a 10-step form (A/B variant consolidation).
| Friction pattern | % impact | Priority |
|---|---|---|
| Early drop-offs Steps 1–4 | 25–44% | P0 Critical |
| Urgency gap at Steps 7–8 | 11–20% | P0 Critical |
| Form not responsive 374px | 3–6% | P1 High |
| Background-effect distraction | 3–11% | P2 Medium |
The Results
| Metric | Before | After | Change |
|---|---|---|---|
| Conversion Rate | Baseline | Baseline × 1.80 | +80% |
| Time to Conversion | Baseline | 30–40% faster | −30 to −40% |
| CLS Score | 0.95 | Flagged for immediate fix | Diagnosed |
| Form Structure | A/B variants | Unified 10-step flow | Standardized |
Why It Worked
- Daily rhythm beats monthly reports — 80–100 sessions analyzed every day exposed patterns that monthly analysis would have buried for weeks.
- Frustration scoring made invisible friction visible — 80/100 on mobile is not a “slightly worse” experience, it’s a fundamentally broken one.
- Phone verification was the silent killer — at Step 11, after 10 completed steps, users abandoned. Moving it earlier removed the end-of-journey barrier.
- ContentSquare precision — session-level frustration data mapped the exact moments users gave up, removing guesswork from recommendations.
Key Takeaways
- Daily session analysis is CRO infrastructure, not a service — it compounds over time.
- Frustration scores tell you HOW users are failing; drop-off data tells you WHERE.
- Mobile journey parity is non-negotiable in B2B SaaS — 83% longer journeys kill conversion.
- Phone verification is a trust gate — place it late and you lose; move it early and brief users.
- CLS 0.95 is a silent conversion killer — invisible to users, lethal to completion rates.