It Started With A Simple Frustration
In Sri Lankan cities, that small weekly uncertainty compounds into a national problem. Residents wait with bags at the gate or miss the truck entirely; bags left out get torn open by dogs and crows. The 2017 Meethotamulla dump collapse made waste management a national issue — yet households still have no digital way to know when the truck is coming, what to separate, or how to report a missed pickup.
The System, Not The Interface
Mapping the current state exposed a service running blind: fixed municipal rounds nobody can track, schedules that vary by ward, near-zero source segregation, and an invisible informal recycling economy of bottle, paper, and metal buyers with no digital layer connecting them to supply.
Observe, Interview, Validate, Synthesize
Stakeholder Ecosystem
Waste collection touches four very different groups — residents, drivers, council supervisors, and informal collectors. I needed to see who depends on whom before designing anything.
The council sits at the centre of every flow but has the least visibility — supervisors run 12 trucks with a phone and paper logs.
It set the interview plan: one track per stakeholder group, with the council dashboard promoted to a first-class product.
- Mapped actor relationships from notes
- Suggested overlooked stakeholders
- Structured the ecosystem diagram
- Field observation & system mapping
- Prioritised the council as key node
- Drew and verified the final map
Proto Personas → Final Personas
Meet The People Behind The Product
Three research-grounded personas anchored every scoping decision — a collection driver, a household recycler, and an independent recyclables buyer.
Nimal Perera
If I knew every schedule change in real time, I could finish my route much more efficiently.
About
Nimal starts work before sunrise and follows a daily collection route across residential neighborhoods. He coordinates with his crew to ensure waste is collected efficiently while keeping to a tight schedule.
Demographics & Context
- Medium technical proficiency
- Uses an Android smartphone daily
- Works outdoors for 8–10 hours
- Responsible for completing assigned routes
Pain Points
- Route changes communicated through phone calls
- Difficult to prove whether a collection was completed
- Traffic delays disrupt daily schedules
- Manual reporting at the end of the day
- Citizens stop the truck to ask questions
Goals
- Complete collection routes on time
- Avoid missed streets or households
- Receive route updates instantly
- Report issues quickly without paperwork
Needs
- Turn-by-turn optimized routes
- One-tap collection status updates
- Ability to report blocked roads or missed pickups
- Offline support in low-network areas
Technology Use
Motivations
- Provide a clean environment for the community
- Take pride in completing routes on time
- Work safely and return home to family
Nadeesha Fernando
I never know if the truck comes at 7 or 10 — some weeks it just doesn’t come at all.
About
Nadeesha manages the household waste for her family before leaving for work. She wants to segregate properly and never miss a collection, but unpredictable schedules make it a weekly guessing game.
Demographics & Context
- High technical proficiency
- Smartphone-first, active on messaging apps
- Time-poor mornings before work
- Primary waste decision-maker at home
Pain Points
- No way to know when the truck will arrive
- Segregation rules differ by ward and change silently
- Missed-pickup complaints go nowhere
- Recyclables pile up waiting for a buyer
- Bags left out get torn open by animals
Goals
- Reliable alerts before each collection
- Clear, ward-specific segregation guidance
- Report a missed pickup in one tap
- Sell recyclables without waiting at home
Needs
- Live truck tracking with street-level ETA
- Push / SMS reminders the night before
- Illustrated trilingual segregation guide
- A simple recyclables marketplace
Technology Use
Motivations
- Set a good recycling example for her child
- Keep a tidy, responsible household
- Reclaim wasted time in busy mornings
Suresh Kumar
I ride 30 km a day hoping someone has bottles — half the time, I find nothing.
About
Suresh buys paper, glass, and metal from households on his tuk-tuk and sells to recyclers. His income depends entirely on finding supply — but today he rides blind, with no way to know who has recyclables to sell.
Demographics & Context
- Basic technical proficiency
- Uses an entry-level Android phone
- On the road 6–8 hours daily
- Sole earner for his family
Pain Points
- Rides long distances with no guaranteed supply
- No way to find who has recyclables nearby
- Prices vary and are hard to agree on
- Fuel costs eat into thin margins
- Competes with other buyers for the same streets
Goals
- Get pickup requests routed to him
- Reduce empty kilometres between stops
- Agree fair, transparent per-kg rates
- Build repeat customers in his area
Needs
- Nearby sell requests with weight and photos
- A clear in-app rate card per material
- Simple accept / navigate flow
- Low-data app that works on older phones
Technology Use
Motivations
- Keep valuable materials out of landfill
- Earn a stable, predictable income
- Provide for his three children
Assumption-based proto personas made team beliefs explicit and testable before field research.
Interviews broke two assumptions: unpredictability (not laziness) blocks segregation, and recyclables already have cash value people fail to capture.
Final personas — Nadeesha the resident, Mr. Perera the supervisor, Suresh the recyclables buyer — anchored every scoping debate that followed.
- Drafted proto personas
- Clustered behavioral patterns
- Improved persona documentation
- Conducted interviews
- Validated assumptions
- Refined final personas
Validation Templates & Results
A structured script and scoring template kept 20+ validation sessions comparable across three languages.
Live truck tracking was the single feature people said they would download for; complaints going nowhere was the sharpest pain.
Tracking + ETA became the hero loop of the MVP, and one-tap reporting was pulled forward from a later phase.
- Generated script & scoring templates
- Summarised 20+ session transcripts
- Flagged trilingual wording issues
- Ran the validation sessions
- Scored and compared results
- Made the scoping decisions
From Pain Points To Design Opportunities
- Sellable B2G SaaS for councils
- Cut truck fuel cost via routing
- Commission on recyclables marketplace
- Know when the truck is coming
- Clear, ward-specific segregation rules
- Turn recyclables into cash easily
- Unpredictable collection times
- Complaints go nowhere
- Councils run on paper
- Live tracking with street-level ETA
- One-tap missed-collection reporting
- Supervisor dashboard + optimized routes
Predictable beats powerful. One reliable ETA earns more trust than ten smart features.
Design for the gate, not the desk. Trilingual, glanceable, one-handed — usable while holding a garbage bag.
Every actor must win. Residents get certainty, drivers get routes, councils get visibility, collectors get demand.
Close every loop. Every report gets a status; every pickup gets a confirmation.
- Scored opportunities by impact/effort
- Cross-linked pains to goals
- Drafted principle wording
- Framed the opportunity matrix
- Chose the four design principles
- Aligned goals with stakeholders
Architecture Before Aesthetics
Each journey below expands into the moments that shaped the flows:
- Proposed IA alternatives
- Stress-tested flows for edge cases
- Drafted journey-map stages
- Defined the architecture
- Designed and tested the flows
- Prioritised the MVP feature set
Sketches, Wireframes, Iterations
Iteration 01 — The Home Screen
The first wireframe led with the schedule calendar; testers ignored it and hunted for the map.
Promote the live map with ETA to the top of the home screen; collapse the schedule into a single next-pickup line.
Time-to-answer for "is it coming today?" dropped from ~20s to under 5s in the next round.
Iteration 02 — Reporting A Missed Pickup
A four-field complaint form felt like council bureaucracy reborn in an app.
One tap + auto-location + optional photo; category inferred from the missed route.
Report completion in testing rose to 100%, and every report now returns a tracked status.
- Expanded the idea space in sketching
- Generated wireframe copy variants
- Suggested layout alternatives
- Sketched and selected directions
- Built and tested the wireframes
- Made every iteration decision
Three Surfaces, One Service
Citizen App
Residents never know when the truck comes or what to separate.
Live truck tracking with street-level ETA, ward-specific schedules with push/SMS/voice alerts, an illustrated trilingual segregation guide, and one-tap reporting.
The weekly uncertainty is gone — the bag goes out exactly once, at the right time.
Driver App
Drivers run memorized routes with no way to log progress or exceptions.
Today's overview — assigned, collected, pending — a live collection map, and per-stop status designed for gloved, one-tap use.
Less backtracking, fewer disputes — the day closes itself.
Admin Dashboard
Supervisors track 12 trucks with a phone and paper.
Fleet GPS, completion by ward, complaint queue with SLAs, daily tonnage, and route optimization that pays for the subscription in fuel savings.
The council sees the service live for the first time — and answers citizens with data.
- Generated edge-case UI states
- Drafted microcopy in three languages
- Produced asset variations
- Designed the final UI & interactions
- Annotated and specced every screen
- Directed the visual language
The full design, in motion:
Built To Scale Across Three Surfaces
Type scales for glanceability — big numerals for ETAs and counts, generous hit targets (min 44px) for gloved and one-handed use, and a component library shared across citizen, driver, and admin surfaces: status chips, route cards, map overlays, and trilingual text styles with Sinhala and Tamil line-height tuning. Color contrast meets WCAG AA on every status pairing.
What This Project Taught Me
Lessons learned: service design beats screen design — the hardest problems lived between stakeholders, not inside the app; validating assumptions early killed two features that felt obvious and saved months; and designing trilingually from day one shaped the layout system far more than any visual choice.
Future roadmap: AI route optimization, IoT smart bins, predictive scheduling, carbon analytics, and recycling rewards — each one building on the data loop the MVP creates.