PH Experience
2.0
Redesigning Cowbell's policyholder portal from a passive document store into an active lifecycle management platform — grounded in session data, usability research, and three measurable KPIs.
Sr. Staff Designer (Director-Level IC) · Cowbell Cyber · 2025–2026
0.9%
Nav. efficiency gain
vs. baseline audit
0%
Fewer dead-ends
Session recording analysis
0/30
Usability tasks
R1→R3 improvement
0/8
KPI criteria met
Across 3 dimensions
01 — Problem
The platform was invisible to the people it served
Policyholders had a portal — but session recordings showed the majority never returned after their first login. The platform held critical data: policy documents, renewal timelines, coverage limits. But it surfaced none of it proactively.
3,394
Sessions with zero meaningful interaction
+70%
Bounce rate on policy detail pages
2×
More time on navigation than on content
+7,816
Support contacts for self-serviceable tasks
02 — Evidence
From session recordings to actionable evidence
Before proposing solutions, I built an evidence base from FullStory session recordings, click heatmaps, and support ticket taxonomy. The data revealed five compounding failure patterns.
Outdated Renewal Awareness
Policyholders missed renewal windows because the portal displayed dates without context, priority, or notification.
Dead-End Page Flows
Users landing on policy detail pages had no clear next action — 70%+ exited rather than engaging with coverage data.
Worst-Click Destination Bias
Click maps showed 3× more traffic to support links than to core portal features — users were routing around the product.
Confidence Gap
No visual representation of coverage adequacy or risk posture. Users couldn't answer "am I protected?" from the portal alone.
Strategy Vacuum
No design precedent, no component library, and no analytics infrastructure existed before this project began.
Compliance Exposure
Policy document access was buried 4+ clicks deep. For audit or claims scenarios, this created meaningful risk exposure.
0%
of active policyholders never used the dashboard — the portal's primary surface. A design product isn't a design product if it forces users to work around it.
03 — Process
Research-first, outcome-anchored
Phase 01
Session Analysis & Baseline Audit
3,394 FullStory sessions reviewed. Click heatmaps, dead-end flows, and support ticket taxonomy mapped into a failure inventory.
FullStory auditHeatmap analysis
Phase 02
Competitive Landscape & Flow Analysis
Benchmarked 10 fintech and insurtech portals. Identified navigation architecture patterns that reduce time-to-information for non-expert users.
10 platformsIA benchmarking
Phase 03
Navigation Architecture & Dashboard Concepts
Three dashboard concepts tested with 12 policyholders. The "Lifecycle" model — surfacing time-sensitive actions — won on task completion and confidence scores.
3 concepts12 participants
Phase 04
Stakeholder Review & Engineering Feasibility
Aligned on scope with PM, Engineering, and Legal. Defined data contracts for real-time policy status, renewal countdown, and compliance indicators.
Data contractsScope alignment
Phase 05
KPI Definition & Success Criteria
Three KPI dimensions finalized — Reliability, Usability, Revenue — each with baseline measurement and tracking method defined before production build began.
3 KPI dimensionsMeasurement plan
04 — Design
Before & After: The Dashboard Redesign
Two screenshots — same product, 14 months apart. The original portal surfaced no proactive context and offered no lifecycle visibility. The redesigned Mission Control platform puts the policyholder's risk posture, next steps, and policy status front and centre on first login.
Previous Design · v6.1
Static Document Store
Dashboard buried below marketing banner — zero policy context on load
Cowbell Factors chart requires expert knowledge to interpret
No renewal timeline, no compliance status, no next-step prompts
Support and eRiskHub surface equally to core policy actions
PH Experience 2.0
Active Lifecycle Platform
Personalised welcome + active policy summary visible on first screen
Policy period progress bar — renewal awareness without clicking
Quick Actions surface the 5 highest-value tasks based on user state
Intelligence Modules: risk score, recommended actions, and documents unified
Passive document store
→
Active lifecycle dashboard
Generic CRS promotional banner
→
Personalised policy card with live data
Static risk chart (expert-only)
→
Cowbell Factor score + recommended actions
No prioritised next steps
→
5 Quick Actions ranked by urgency
Previous Design
Static policy document store
No renewal or expiry visibility
Coverage data buried 4+ clicks deep
Zero proactive notifications
No risk posture or compliance status
Users exit to broker for basic answers
Updated Design
Active lifecycle management platform
Renewal countdown with 60/30/7-day triggers
Coverage summary visible on first screen
Proactive in-portal action prompts
Live compliance status with remediation paths
Self-serve answers to 80% of common questions
05 — KPIs
Three KPIs grounded in research
Reliability
Session return rate
+0%
Target: +25% · Exceeded
Policyholders returning within 30 days of first login — the primary proxy for perceived platform value.
Usability
Task completion rate
0/30
Tasks completed unaided · R3 testing
From 9/30 at baseline to 17/30 post-launch across three rounds of testing.
Revenue
Renewal capture rate
0/8
KPI criteria met · 87.5% achievement
In-portal renewal actions initiated without broker intervention — tying design directly to revenue retention.
06 — Impact
What the project moved forward
0.9%
Navigation efficiency gain
Time-to-task vs. baseline
0%
Fewer dead-end sessions
Post-launch vs. pre-launch
7/8
KPI criteria met at launch
Across all 3 dimensions
0%
Dashboard adoption (was 0%)
Active users engaging dashboard
−0%
Support tickets for self-serve tasks
Ops workload reduction
0
Critical post-launch defects
All device types tested
07 — Learnings
What this project taught me
Define KPIs before wireframes
Setting measurable success criteria upfront changed every conversation — from "does this look good?" to "does this move the metric?"
Session data is a design brief
3,394 FullStory sessions gave us more honest requirements than any stakeholder interview. Users showed us where they were confused — we didn't have to ask.
Proactive surfaces beat reactive ones
The biggest usability gains came from surfacing time-sensitive information before users went looking for it. Anticipation beats navigation.
No component library is a constraint, not an excuse
Building the design system in parallel added 3 weeks — but ensured consistency across a team that had none, and reduced design debt before it accumulated.