Delivery Partner — Rider App
A delivery app designed around the rider, not the dispatch system.
Overview
Four screens — Pickup, Delivery, Earnings, Profile — plus a fifth state. Go offline and Pickup becomes Start duty, a demand map that says where to position and what an hour there is worth.
Built as a working Next.js prototype with real map tiles and live routing, so every state had to survive implementation.
Problem
Built for the Dispatch System, Not the Rider
A rider is a piece-rate worker running a small business on a phone. No salary, no floor — income tracks hours and positioning, and the rider absorbs the cost of every idle minute while the platform absorbs none of it. That asymmetry is the design brief. Judged against it, the standard app fails in five places.
Earnings show a total, not a decision. “₹2,860 today” never answers the question that changes the next hour: how close am I to the bonus?
Bonus rules are opaque. Riders reverse-engineer them from experience — and the ones who guess wrong work extra hours for a payout that was never reachable.
Offline is a blank screen. It is also when a rider makes their highest-leverage call of the day: where to position.
Handoff notes are free text. “Don’t ring the bell” gets read at pickup and forgotten at the door.
Payout status sits in another tab. Money earned and money received are two different anxieties.
Approach
One Question Per Screen, Answered in Two Seconds
Pickup
Is this order worth taking?
Delivery
Where exactly am I going, and what does this customer want at the door?
Earnings
What have I made, and what is still in reach?
Start duty
Where should I be, and what is an hour there worth?
Game Mechanics, Pointed at the Rider
The app is unapologetically gamified — a trophy-badged ₹4,000 pool, locked tiers, progress bars, and a mechanic the interface literally calls a quest. Those are also the mechanics platforms use to keep tired riders on the road: name no pool, show only the bar, nudge hardest at the point of fatigue.
The inversion is in what gets disclosed. The full pool is stated up front. Each tier carries its exact delta — 4 more orders unlocks ₹1,000, 3.5 more hours unlocks ₹3,000 — beside a lock and the hours left on the clock. Today's ₹2,860 splits into ₹2,240 base and ₹620 incentives, so the rider sees what the extra effort actually bought. Same goal gradient, opposite direction: enough information to walk away.
Offline Is a Screen, Not an Absence
Start duty gives the rider the dispatcher's view of the city: a hex-binned demand map, ranked zones priced in ₹/hr, a peak-hour histogram, and a plain reason for each — densest kitchen cluster in the city and short drops. That information asymmetry is the one that costs riders most.
Earned and Received, Together
Payout folds into Earnings, not a separate wallet tab — ₹2,860 to HDFC ···1842, pending, expected by 11:00 PM, updated 4 minutes ago. Enough to stop checking.
Screens
The Daily Loop
Pickup leads with the two numbers that decide acceptance — 1.2 km, up to ₹110 — and breaks the payout into ₹100 + ₹10 on-time so nothing surprises the rider at settlement. Delivery carries the landmark that actually finds the building. Start duty replaces the empty offline state with a positioning map.
Night Shift
Most delivery volume happens after sunset, on a handlebar mount at full brightness. The dark theme re-tunes rather than inverts: bonus progress shifts to amber to hold contrast, the heat map saturates so density still reads, and maps get a real dark basemap. Status colours keep their meaning in both themes.
One System, Every State
One component vocabulary across all screens — labelled card, stat split, progress bar with an unlock delta, map with a floating summary strip. The bottom nav is the deliberate exception: four tabs online, three offline, because tabs that can’t do anything shouldn’t hold thumb space.
Scope
Craft
Decisions & Tradeoffs
Money is the largest type on any screen it appears on. For a piece-rate worker, the number is the content — everything else is metadata.
Locked bonuses state the delta, not just progress. “Orders 16/20” is a status report; “4 more orders → ₹1,000” is an instruction. The goal gradient pulls either way — what changes is whether the rider can see the finish line clearly enough to decide it isn’t worth chasing.
The demand map is priced, not just coloured. Koramangala at ₹165–210/hr, 204 orders across 46 riders, 4 minutes away — riders choose on expected earnings, not a colour gradient.
Offline is a real screen. The app helps riders earn more per hour instead of only pushing them to stay online longer.
Constraints sit next to the control they constrain. “Prefers no call” renders beside the call button, and preferences sit directly above the delivery CTA.
Real tiles, real routes. OpenStreetMap and live OSRM routing surface the layout problems a drawn map hides — label collisions, route overlap, zoom on a long drop.
No QR or pay-on-delivery code. A substantial flow with its own failure modes — cut cleanly rather than represented badly.
Start a Project
Want work like this
built for you?
Mobile product design, AI product strategy, or a complete overhaul — let's talk.