How I transformed a paper-based system plagued by falsified data into a trusted GPS-verified field reporting system for Philip Morris Algeria’s tobacco sales operation.
📋 Project Overview
Client: Philip Morris International (Algeria Division)
Industry: Tobacco / FMCG (Fast-Moving Consumer Goods)
Platform: iPad App (Sales Field Force) + Web Dashboard (Top Management)
Timeline: 8 months (3-month MVP + 5-month iteration and dashboard build)
Role: Solo UX/UI Designer
Team: Agency Developer (CEO), PMI Stakeholders
Users: 300+ field sales agents + Algeria’s national sales management team
The Mission: Replace a paper-based data collection system where agents could falsify entries from home with a verifiable, GPS-locked digital solution that actually worked in real-world field conditions.
🧾 What is iReport?
iReport was a custom iPad application designed for Philip Morris Algeria’s field sales force, who visit thousands of retail locations each month to collect data on tobacco product availability, shelf presence (planograms), pricing, and competitor activity.
Before iReport, this process relied on paper-based surveys — an outdated method that made it nearly impossible to verify whether visits were actually taking place, or if the data collected was real.
🎯 Project Impact at a Glance
🎯 Product Goals
To build a real-time, GPS-verified reporting tool that:
Prevents data falsification Helps agents report quickly and reliably in real conditions Improves sales visibility and field accountability Ensures legal compliance with tobacco marketing/display regulations Enables fast, accurate decision-making for management This wasn't just a UX upgrade. It was a mission-critical digital transformation for a global company operating in a heavily regulated industry.
🚨 The Core Problem
Sometimes they would fill the forms at home just to submit something. There was no way for managers to verify field presence..
The Paper System's Fatal Flaws:
Data Integrity Crisis: No way to verify if agents were actually visiting sales points Cognitive Overload: Complex paper forms with 20+ fields overwhelmed users Operational Blindness: Management had zero real-time visibility into field operations Process Inefficiency: Manual transcription created delays and additional errors The Real-World Reality:
Through field visits, I discovered agents were working under challenging conditions:
Bright sunlight making screens hard to read Often using one hand while driving or standing Covering multiple territories daily under time pressure Switching between different types of sales points with varying requirements This wasn't just about digitizing forms - it was about rebuilding trust in the entire data collection process.
🧠 My Role
I was the sole UX/UI Designer working directly with agency CEO/developer and PMI stakeholders under aggressive timeline constraints and responsible for:
Field research and user immersion (ride-alongs with 6 agents) UX strategy and IA structuring Full UI design for iPad and Web Dashboard Collaborating with stakeholders and the agency developer Prototyping, user testing, iteration QA and visual refinement through delivery 🔍 Discovery Through Direct Immersion
Real-World Observation (6 Field Agents)
Instead of assumptions or lab testing, I joined field reps on live store visits. What I discovered redefined the entire design direction:
Witnessed the frustration of agents juggling paper forms while trying to take photos Agents were standing, moving, or even driving while data collection process They were overwhelmed by long, repetitive paper forms Therefore,
Managers had zero confidence in the data they received I interviewed PMI management about their data quality concerns I created quick personas focused on field constraints rather than demographics Key Insight: The problem wasn't just poor tools - it was a complete breakdown of trust between field agents and management.
💡 Strategic Design Solutions
1. GPS-Locked Access = Built-in Accountability
Problem: Agents could submit forms from anywhere—often from home.
Solution: Forms unlock only within 50 meters of a sales point via real-time GPS.
The Psychology: This felt significant enough to change behavior but not punitive enough to create rebellion. Users understood they were being held accountable, but the system was helping them do their job properly rather than catching them cheating.
Clear message: “Move to the sales point entry to begin.” Built-in resistance that felt fair, not punishing Supported by Weber’s Law: noticeable, but not disruptive Result: ↓ 50% data falsification within the first 90 days.
2. Bite-Sized, Guided Task Flow
Problem: The original paper form was overwhelming - agents had to remember details across multiple sections while standing in a store.
Solution: Redesigned the information architecture into logical, bite-sized steps:
Sales point identification (auto-populated from GPS) Product availability check Final notes and submission Result: Form completion dropped from ~20 minutes to ~5 minutes → ↑ 40% faster task completion
3. Designed for Real Use: One Hand, Outdoors, On-the-Go
The Reality Check: Watching an agent try to hold an iPad, take notes, and photograph a shelf simultaneously made me completely rethink the interface.
Design:
Large thumb-accessible targets Glare-friendly, high-contrast visuals Simplified navigation with clear visual hierarchy Instant camera access with forgiving targeting Guided by Fitts’s Law, optimized for efficiency under motion
4. 🧠 Smart Defaults and Memory Aids
Problem: Users repeated the same data across stores or forgot mid-task.
Design:
Auto-filled store info (based on GPS) Pre-filled fields from previous visits to the same location Progress indicators showing exactly where they were in the process to reduce abandonment (Zeigarnik Effect) Visual reminders for missing steps Result: ↑ 30% engagement within 1 month
🔁 Iteration Based on Real Feedback
1# The Photo Upload Problem
Discovery: During pilot testing, many agents were skipping the mandatory planogram photos.
Root Cause: They didn't understand that photos were required, not optional.
Solution: Blocked form submission + added inline guidance.
Result: 100% planogram image compliance.
2# The Error Message Disaster
Discovery: Generic error messages were causing support calls and user frustration.
Root Cause: "Invalid entry" tells users nothing about how to fix the problem.
Solution: Contextual, actionable error messages: "Move to the sales point entry to begin." or "Please select a product category before proceeding."
Impact: 23% reduction in support requests.
3# The Abandonment Issue
Discovery: Users were starting reports but not finishing them.
Root Cause: No sense of progress or completion in a multi-step process.
Solution: Clear progress indicators and a completion summary showing what they'd accomplished.
Result: 30% increase in user engagement within 3 months, with 90% field force adoption within 6 weeks.
📊 Web Dashboard Design (Top Management)
With zero time for management user research, I had to make educated design decisions based on stakeholder interviews and industry best practices.
Approach: Created a clean, data-focused interface that prioritized real-time insights over complex analytics. The key was making field data immediately actionable for decision-makers who were used to waiting weeks for reports.
Timeline: Delivered iteratively over 5 months alongside app improvements.
Result: ↑ 70% faster decision-making from management, thanks to real-time verified data.
📈 Business & Operational Outcomes
🎓 What This Project Taught Me
📍 Context Beats Assumptions
One field visit revealed more than any wireframe ever could.
Watching users in motion — not just thinking about them — changed everything.
🔐 Trust Is a Design Problem
The GPS lock wasn't just a technical feature - it was a behavioral design intervention that rebuilt trust between field agents and management.
⚡️ Constraints Drive Focus
Short timelines forced me to skip wireframes, and go straight to what worked: direct field testing, fast iteration, and business-backed design decisions.
🔁 UX Is an Ecosystem
Better field agent experience directly led to better management insights. Happy agents provided higher quality data, which made the dashboard more valuable.
🔮 What I’d Do Differently
Push for Management Research: Even 2-3 stakeholder interviews would have improved the dashboard design significantly. Build in Feedback Loops: The app should have included built-in feedback mechanisms for continuous improvement. Create Design System: With more projects coming, a component library would have accelerated future development. Measure Baselines: Better pre-launch metrics would have made the impact story more compelling. 💡 Why This Case Study Matters
This wasn’t just a UI overhaul. It was a behavioral redesign of an entire operational system.
It shows my ability to:
Lead UX strategy under real-world constraints Design for high-friction environments (motion, sunlight, skepticism) Drive measurable business change through thoughtful UX Navigate multi-stakeholder enterprise teams with speed and clarity “iReport proved that UX design isn’t just about improving screens — it’s about making truth easier than lies.”
The defining moment: Realizing that rebuilding trust between agents and management was more important than just making a prettier form. That insight shaped every design decision that followed.