Shipment
The World's Most Advanced Logistics Operating System — Multi-Country, Multi-Currency, AI-Powered, Built to Dominate Global Delivery Infrastructure
A comprehensive enterprise architecture document covering market positioning, multi-country platform design, AI-powered core engines, real-time QR chain-of-custody tracking, multi-warehouse orchestration, cross-border operations, and the technology blueprint for building the most technologically advanced shipment and logistics operating system on the planet — designed from day one to compete with and surpass DHL, FedEx, Amazon Logistics, and Noon at global scale.
Table of Contents
Platform & Market
- Executive Summary
- Market Gap & Opportunity
- Problems with Existing Systems
- Product Vision & Principles
- Full Stakeholder Map
- Detailed User Journeys
- Product Modules & Capabilities
Architecture & Engines
- Recommended Architecture
- Tech Stack
- Event-Driven Architecture
- Canonical Data Model
- Dispatch Engine
- Vehicle Recommendation Engine
- Pricing / Rate Engine
- SLA Prediction Engine
- Address Intelligence
- Route Optimization
Operations & Integrations
- Driver Fairness & Earnings
- Receiver Experience
- Failed Delivery Handling
- Reverse Logistics
- Multi-Branch / Multi-Country
- Marketplace Model (Wadi)
- WooCommerce Integration
- Shopify Integration
- Public API / Webhooks / SDKs
Strategy & Execution
Platform & Market
Market analysis, product vision, stakeholder mapping, user journeys, QR chain-of-custody system, multi-warehouse management, and the complete enterprise module landscape.
Executive Summary
Shipment is the operating system for physical delivery — the layer between "someone bought something" and "they have it in their hands." Every order, from every channel, through every fleet, in every country — orchestrated through one intelligent platform.
The Problem
The logistics industry — particularly in MENA and emerging markets — runs on fragmented, rigid, and opaque systems. Merchants use one courier for same-day, another for next-day, a third for intercity. Each has its own dashboard, its own tracking page, its own rate card, its own API (if any). Marketplace sellers juggle fulfillment rules across Noon, Amazon, Wadi, and their own stores. Drivers are treated as disposable resources. Receivers have zero control. Dispatchers work in spreadsheets. Finance teams reconcile COD in Excel.
No single system exists that acts as the operating system for the entire shipment lifecycle — from the moment a buyer clicks "Place Order" to the moment cash is reconciled and the merchant is paid.
What Makes This Different
| Traditional Courier | Shipment |
|---|---|
| Fixed rate cards, manual quotes | Dynamic pricing engine with real-time cost modeling |
| Manual dispatch via phone/WhatsApp | AI-powered dispatch with driver suitability scoring |
| "Your package is on the way" | Real-time map tracking, receiver controls, time-window promises |
| One-size-fits-all vehicles | Automatic vehicle selection based on cubic volume, fragility, weight |
| COD = black box until settlement | Real-time COD tracking with per-delivery reconciliation |
| APIs as afterthought | API-first architecture; every feature is an API |
| Returns are a nightmare | Reverse logistics as a first-class workflow |
| Drivers are anonymous | Driver profiles, ratings, specializations, earnings transparency |
Scale Design Point — Enterprise From Day One
This is not a startup experiment. Shipment is built as an enterprise-grade, multi-country, multi-warehouse logistics operating system from day one. Every feature — multi-currency settlement, warehouse management, cross-border customs, QR chain-of-custody, AI dispatch, computer vision POD verification — is core, not "Phase 2." We are building the system that makes DHL, Aramex, and Noon's logistics look outdated.
Revenue Model (Hybrid)
| Revenue Stream | Description |
|---|---|
| Per-shipment fee | 10 AED flat fee per delivery or return — simple, competitive, market-leading |
| COD handling fee | 1-3% of COD amount collected |
| Subscription tiers | Monthly plans for merchants (Free / Growth / Enterprise) |
| API usage | High-volume API consumers pay per-call beyond free tier |
| Marketplace fulfillment | Wadi merchants pay fulfillment fees (pick, pack, ship) |
| Float income | Interest on COD float between collection and settlement |
| White-label licensing | Courier companies license the platform under their brand |
Market Gap & Opportunity
The MENA Logistics Landscape
The MENA eCommerce market is projected to exceed $50B by 2027. Yet logistics infrastructure remains the #1 bottleneck. Address chaos, COD dominance (50-70% of orders), fragmented last-mile, driver churn, and no unified platform create massive opportunity.
Competitive Landscape
| Player | Strengths | Fatal Weaknesses |
|---|---|---|
| Aramex | Brand, network, cross-border | Legacy tech, terrible tracking UX, slow API, rigid pricing |
| SMSA | Saudi coverage, gov contracts | No marketplace integration, no real-time tracking |
| Noon logistics | Integrated with marketplace | Captive — only serves Noon, not a platform |
| Amazon logistics | Scale, tech excellence | Captive — only serves Amazon |
| Shippo / ShipStation | Aggregation, multi-carrier | Label printing only — no fleet, no dispatch, no driver app |
| Onfleet / Bringg | Last-mile dispatch | No COD, no marketplace model, no rate engine, US/EU focused |
Nobody has built the Stripe of logistics — a platform that handles the entire shipment lifecycle end-to-end, works for ANY merchant, ANY marketplace, ANY fleet, is API-first, handles MENA realities (COD, addresses, multi-language), and scales from 10 to 1M+ shipments/day.
Problems with Existing Logistics Systems
Merchant Pain Points
No Single Pane of Glass
Merchants managing 3-5 couriers have 3-5 dashboards, 3-5 logins, 3-5 rate cards, 3-5 tracking pages. No unified view.
Opaque Pricing
"We'll quote you" is the standard. No real-time pricing. Surcharges appear on invoices weeks later.
COD Black Hole
When does the money arrive? Which orders were collected? Which driver has the cash? Most systems cannot answer in real-time.
No Delivery Intelligence
No courier gives a probabilistic answer to "When will my shipment arrive?" Just "1-3 business days."
Driver Pain Points
Opaque Earnings
Drivers don't know how much they'll earn before accepting. Commission structures are complex and hidden.
No Route Optimization
Drivers with 20 packages get a list, not an optimized route. They use personal judgment and Waze separately.
Receiver Pain Points
No Control
Cannot reschedule, redirect, choose a time window, or communicate with the driver until they're already at the door.
Tracking is Useless
"In transit" for 3 days. No map, no ETA, no updates. The tracking page is a static list of statuses.
Systemic Industry Problems
Most courier systems were built as monoliths in 2005-2015. They can't handle real-time events, webhooks, or concurrent operations at scale. No multi-tenancy. No intelligence layer. Mobile-last design. No event streaming.
Product Vision & Principles
"Shipment is the operating system for physical delivery — the layer between 'someone bought something' and 'they have it in their hands.'"
10 Design Principles
Platform First, Product Second
Every feature must be accessible via API. Our own dashboards are just consumers of our own API.
Intelligence by Default
The system must always know more than the human operator. Suggest, predict, warn, auto-resolve.
Multi-Everything by Design
Multi-tenant, multi-country, multi-currency, multi-language, multi-branch, multi-fleet — from day one.
Driver = Partner
Drivers are the physical execution layer. Every decision must make their job easier and more rewarding.
Receiver = First-Class User
The receiver is not a passive endpoint. They should have control, visibility, and options.
Real-Time or Nothing
Everything — tracking, dispatch, status, alerts, COD — must be real-time. 2-second propagation.
COD as Core Competency
In MENA, COD is the default. The system must treat cash flow with the same rigor as shipment flow.
Fail Gracefully, Recover Automatically
Automated recovery, retry logic, and graceful degradation for every failure mode.
Build for the Last Mile of the Last Mile
The hardest deliveries are where competitors fail. We design for these first.
Data is the Moat
Every shipment generates data. Over time, dispatch gets smarter, pricing sharper, SLA predictions more accurate. This compounds.
Full Stakeholder Map
Every actor in the logistics ecosystem has a defined role, interface, and set of needs within the Shipment platform.
| Stakeholder | Role | Primary Interface | Key Needs |
|---|---|---|---|
| Merchant / Shipper | Creates shipments, tracks orders | Web Dashboard, API | Reliability, pricing transparency, fast COD settlement |
| Marketplace (Wadi) | Bulk creation, multi-vendor fulfillment | API, Admin Console | Split shipments, seller-level tracking, SLA enforcement |
| Driver | Picks up, delivers, collects COD | Mobile App | Clear instructions, fair earnings, optimized routes |
| Fleet Partner | Provides drivers and vehicles | Fleet Portal | Driver management, earnings, SLA compliance |
| Dispatcher | Monitors ops, handles exceptions | Dispatch Console | Real-time visibility, alerts, bulk operations |
| Branch Operator | Hub operations, sorting | Branch Console | Scan workflows, inventory, transfers |
| Receiver / Customer | Receives deliveries | Tracking Page, SMS | Control, visibility, easy returns |
| Finance Team | COD reconciliation, invoicing | Finance Console | Real-time cash tracking, automated reconciliation |
| Support Team | Handles escalations, disputes | Support Console | Full timeline, communication tools, resolution workflows |
Permission Model
Permissions are additive — roles can be combined. Every role is scoped to a tenant and optionally to a branch. API keys inherit the permissions of the role they're created under. All permission checks happen at the API gateway level.
Merchant Onboarding Flow
Email, business name, trade license upload
Trade license, TRN, bank details (automated + manual review)
Based on volume tier, category, service levels
API keys generated, webhook configured, plugin installed
Guided first shipment with dummy data
Production API access, full dashboard
For enterprise merchants (1000+ shipments/month): dedicated account manager, custom rate negotiation, technical integration support, data migration from existing courier (Aramex/SMSA/DHL historical data import via CSV), branded tracking page setup, and custom webhook configuration.
Data Migration Strategy
Merchants switching from Aramex, SMSA, DHL, or other couriers can import historical data:
Detailed User Journeys
Merchant: First Shipment
Name, phone, address/coordinates
Weight, dimensions, item type
Same-day, next-day, scheduled
Geocode, verify, flag risk
Based on volume, weight, fragility
Rate card + surcharges + COD fee
"2-4 PM, 87% confidence"
Unique tracking number
PDF + ZPL thermal
Auto-dispatch via DSS
Driver: Typical Delivery Day
Start Shift
Report location, receive optimized delivery queue, accept delivery batch.
Pickup
Navigate to merchant, scan barcode, verify package count, confirm pickup.
Deliver
Navigate to receiver, hand over package, collect COD if applicable, capture POD.
End Shift
View daily earnings summary, deposit COD at branch, review performance.
Receiver: Tracking Experience
SMS on creation → Proactive tracking link → Live map with ETA → "Driver 5 min away" alert → Delivery confirmation with POD → Rating widget → One-tap return if needed
Finance: Daily COD Reconciliation
Real-time ledger updated
Operator scans & confirms
Collected vs deposited
Bank transfer confirmed
Product Modules & Capabilities
The platform is organized into four module families: Core Platform, Operations, Financial, and Intelligence.
Shipment Management — All at Launch
| Capability | Status |
|---|---|
| Single + bulk shipment creation (API, dashboard, CSV, QR-initiated) | Core |
| AWB + QR code generation (every package gets unique QR) | Core |
| Full lifecycle status tracking with QR scan at every handoff | Core |
| Multi-piece + split shipments | Core |
| Cross-border shipments with customs documentation | Core |
| Multi-warehouse routing & inventory orchestration | Core |
| Scheduled, recurring, and template-based shipments | Core |
QR Chain-of-Custody System
Like Noon but more advanced. Every package gets a unique QR code at creation. At every handoff point — seller packs → driver picks up → hub inbound → sorting → transfer → hub outbound → delivery driver → receiver — the QR is scanned. This creates an immutable chain-of-custody with GPS coordinates, timestamp, actor ID, and photo evidence at each scan. Lost package? Check the chain. Damaged? We know exactly which handoff caused it.
Packs & labels package
GPS + timestamp + photo
Automated sorting trigger
Intercity manifest verify
GPS + confirms possession
Confirms receipt (optional)
Dispatch Engine — AI-Powered
| Capability | Status |
|---|---|
| AI auto-dispatch with Driver Suitability Score (DSS) | Core |
| Manual dispatch override for dispatchers | Core |
| Real-time driver tracking (WebSocket + GPS) | Core |
| Batch dispatch with VRP solver (Google OR-Tools) | Core |
| Zone-based dispatch with geo-fencing | Core |
| Predictive pre-positioning (ML demand forecasting) | Core |
| Cross-fleet dispatch (partner fleet integration) | Core |
Financial Module — Multi-Currency, Multi-Country
| Capability | Status |
|---|---|
| COD collection tracking with real-time driver cash balance | Core |
| Multi-currency settlements (AED, SAR, KWD, BHD, OMR, QAR, USD, EUR) | Core |
| Automated merchant invoicing with per-shipment line items | Core |
| Dynamic rate cards with A/B testing | Core |
| Real-time reconciliation engine with exception queue | Core |
| Cross-border duty/tax estimation & FX rate locking | Core |
| Driver payroll automation with multi-country tax compliance | Core |
Warehouse Management System (WMS)
| Capability | Status |
|---|---|
| Multi-warehouse inventory tracking across countries | Core |
| QR-based pick, pack, and label workflows | Core |
| Bin/shelf location management with barcode mapping | Core |
| Wave picking & batch picking optimization | Core |
| Returns processing & quality inspection workflows | Core |
| Cross-warehouse transfer management | Core |
| Real-time inventory dashboard with aging alerts | Core |
Architecture & Engines
System architecture, tech stack decisions, event-driven design, data models, and the core algorithmic engines that power intelligent dispatch, pricing, SLA prediction, and route optimization.
Recommended Architecture
Cloud-native, event-driven microservices architecture deployed on DigitalOcean Kubernetes (DOKS). Each service owns its data and communicates via events and APIs.
React + Next.js
React Native
REST + GraphQL
DigitalOcean Spaces CDN
Kong on DOKS
JWT + API Keys
CRUD, status, labels
DSS, assignment
Rate cards, quotes
Real-time, ETA
Primary datastore
Cache, real-time, pub/sub
Full-text search
Files, POD photos
Launch config: Multi-region DOKS (UAE primary, Saudi DR), PostgreSQL with read replicas + PostGIS, Redis Cluster (6 nodes), Apache Kafka (3-broker), multi-AZ deployment. No single points of failure. Auto-scaling from 100 to 100,000+ pods. Every service runs minimum 3 replicas. Database failover is automatic. This is not "scale later" — it's "scale now."
Recommended Tech Stack
| Layer | Technology | Rationale |
|---|---|---|
| Backend | Node.js (NestJS) | Fast dev, TypeScript, enterprise patterns, excellent for event-driven |
| Primary DB | PostgreSQL 16 | JSONB, PostGIS for geo, RLS for multi-tenancy |
| Cache | Redis 7 (Cluster) | Sub-ms reads, pub/sub for tracking, sorted sets for dispatch queue |
| Event Streaming | Redis Streams (included in Redis) | Built into Redis we already pay for. Handles inter-service events at our volume. Upgrade to Kafka when hitting 50K+ deliveries/day. |
| AI/ML Runtime | TensorFlow Lite (on-device) + Claude API (LLM) + LightGBM (server) | Edge AI on driver phones, LLM chatbot for customer support, ML models for ETA/demand/fraud |
| Computer Vision | TensorFlow Lite (MobileNet) | On-device POD verification, damage detection, QR scanning. No cloud round-trip needed. |
| Mobile | React Native | Single codebase iOS+Android, large ecosystem, code sharing with web |
| Web Frontend | React + Next.js | SSR for tracking pages, rich ecosystem, TypeScript |
| Maps & Logistics | HERE Maps (single provider) | Geocoding, routing, tour planning, geofencing, fleet tracking, traffic ETA. Best logistics features. 10x cheaper than Google. ~800 AED/month. |
| Map Display | OpenStreetMap + Leaflet (free) | Tracking page and dispatch console. Zero per-load charges. Driver uses Google Maps/Waze free app for navigation. |
| WhatsApp API | 360dialog (official Meta BSP) | Zero markup on Meta rates. $5/month platform fee. WhatsApp buttons + keyword bot + Claude Haiku fallback for complex queries. |
| Payments | HyperPay + Stripe | HyperPay for MENA (MADA, Apple Pay), Stripe for international |
| Cloud | DigitalOcean | Simpler, cost-effective; DOKS, managed PostgreSQL, managed Redis, Spaces |
| Monitoring | DigitalOcean built-in + custom dashboard | DO monitoring for infrastructure. Custom admin dashboard for business metrics. Zero extra cost. |
| CI/CD | GitHub Actions | Integrated with code, good ecosystem |
HERE Maps as single provider for geocoding, routing, tour planning, geofencing, fleet tracking, and traffic ETA. Cost: ~800 AED/month at launch (Google would be $7,500+). Includes logistics-specific features (vehicle routing, multi-stop optimization) that directly increase deliveries per driver. Driver navigation via free Google Maps/Waze app. Tracking page maps via free OpenStreetMap + Leaflet.
Event-Driven Architecture
All state changes produce domain events. Services react to events, enabling loose coupling, audit trail, and easy addition of new consumers.
Core Event Catalog
| Event | Producer | Key Consumers |
|---|---|---|
shipment.created | Shipment Service | Dispatch, Pricing, Notification, Analytics |
shipment.assigned | Dispatch Service | Tracking, Notification, Driver App |
shipment.picked_up | Driver App | Tracking, Notification, Webhook |
shipment.delivered | Driver App | COD, Notification, Analytics, Webhook |
shipment.delivery_failed | Driver App | Dispatch, Notification, Webhook |
driver.location_updated | Driver App | Tracking, Dispatch, Geofencing |
cod.collected | COD Service | Finance, Driver Wallet, Reconciliation |
settlement.completed | Settlement Service | Notification, Merchant Dashboard |
Synchronous (HTTP): Pricing calculation (merchant needs price immediately), driver availability query. Asynchronous (Event): Dispatch assignment, webhook delivery, COD recording, POD upload, analytics.
Canonical Data Model
Core entities with multi-tenancy built into every record via tenant_id and PostgreSQL Row-Level Security.
Core Entities
Shipment Status State Machine
Dispatch Engine Design
Traditional dispatch: "Find the nearest available driver." Shipment dispatch: "Find the driver who maximizes delivery success probability while minimizing cost and ensuring fairness."
Driver Suitability Score (DSS)
+ (w2 × VehicleMatchScore)
+ (w3 × CapacityScore)
+ (w4 × PerformanceScore)
+ (w5 × FairnessScore)
+ (w6 × SpecializationScore)
+ (w7 × ZoneAffinityScore)
+ (w8 × CODCapacityScore)
+ (w9 × TimeEfficiencyScore)
− (Penalties)
Default Weights
| Weight | Factor | Value | Adjustable By |
|---|---|---|---|
| w1 | Proximity | 0.25 | Tenant settings |
| w2 | Vehicle Match | 0.20 | System (hard constraint) |
| w3 | Capacity | 0.10 | Tenant settings |
| w4 | Performance | 0.15 | Tenant settings |
| w5 | Fairness | 0.10 | Tenant settings |
| w6 | Specialization | 0.08 | System |
| w7 | Zone Affinity | 0.07 | Tenant settings |
| w8 | COD Capacity | Hard constraint | System |
| w9 | Time Efficiency | 0.05 | Tenant settings |
Dispatch Edge Cases
No Driver in Zone
Expand radius progressively: 5km → 10km → 15km → broadcast to all nearby drivers.
All Drivers at COD Capacity
Alert dispatcher; option to send non-COD driver + redirect payment to online.
Vehicle Breakdown Mid-Route
Auto-reassign remaining shipments to nearest available driver.
100 Shipments Simultaneously
Batch dispatch: optimize collectively using VRP solver, not individually.
Vehicle Recommendation Engine
Automatic vehicle selection based on multi-dimensional analysis: cubic volume, fragility, temperature needs, quantity, item type, route type, and business rules.
| Vehicle Type | Code | Max Weight | Max Volume | Use Cases | Hourly Cost |
|---|---|---|---|---|---|
| Motorcycle | MOTO | 10 kg | 30,000 cm³ | Documents, small parcels, food, instant | 25 AED |
| Compact Car | CAR_S | 30 kg | 200,000 cm³ | Medium parcels, multi-stop | 35 AED |
| Cargo Van (Small) | VAN_S | 500 kg | 3M cm³ | Furniture, appliances, bulk | 70 AED |
| Refrigerated Van | VAN_R | 500 kg | 3M cm³ | Food, pharma, temp-sensitive | 120 AED |
| Box Truck | TRUCK_B | 3,000 kg | 20M cm³ | Hub transfers, bulk fulfillment | 150 AED |
Pricing / Rate / Surcharge Engine
Dynamic, transparent, and configurable per tenant. Supports static rate cards, dynamic pricing, and hybrid models.
(BaseRate + DistanceCharge + WeightCharge + CODFee + Surcharges + ZoneAdj)
× TimeMultiplier × DynamicPricingMultiplier × (1 - MerchantDiscount)
)) × (1 + TaxRate)
Surcharge Types
| Surcharge | Condition | Typical Amount |
|---|---|---|
| Fuel surcharge | Always | 5-15% of base |
| Fragile handling | is_fragile | 5-10 AED |
| Temperature-controlled | is_temperature_sensitive | 15-30 AED |
| COD fee | payment = COD | 2.5% of COD amount |
| Instant delivery | service = INSTANT | 50-100% premium |
| Remote area | zone.difficulty > 4 | 10-25 AED |
SLA Prediction Engine
Not "1-3 business days." We produce specific, probabilistic delivery windows: "Delivery between 2:00 PM - 4:00 PM today, 87% confidence." This powers the delivery promise shown to receivers and SLA compliance tracking.
Service Levels
SLA Prediction Model
+ TransitTime + HubProcessingTime + LastMileDeliveryTime + BufferTime
ConfidenceInterval = f(historical_variance, address_quality, time_of_day, weather, Ramadan)
SLA Risk Score (Real-Time)
| Score Range | Status | Action |
|---|---|---|
| 0.0 - 0.2 | GREEN | On track. No action. |
| 0.2 - 0.5 | YELLOW | At risk. Alert dispatcher dashboard. |
| 0.5 - 0.8 | ORANGE | Likely miss. Auto-suggest reassignment. Notify merchant. |
| 0.8 - 1.0 | RED | Will miss. Auto-attempt reassignment. Update receiver ETA. SLA credit triggered. |
When confidence is >90%, we offer: "Delivered by [time] or shipping is free." No courier in MENA offers this. Cost: ~3-5% of revenue if we hit 95%+ on-time rate. Value: massive trust signal that drives merchant acquisition.
Address Intelligence & Location Verification
Get the driver to the exact delivery point with < 50m accuracy, on the first attempt, every time. This is where most competitors fail in MENA.
Address Resolution Pipeline
Text, Makani number, or pin drop
HERE Maps API
0-1 based on completeness + history
PostGIS polygon lookup
Address Enrichment Layers
Layer 1: Input-Time
Autocomplete suggestions (HERE Autosuggest), Makani number resolution (free UAE API), map pin drop, Plus Codes acceptance.
Layer 2: Intelligence
Address clustering (group deliveries within 50m). Driver-confirmed locations (GPS at delivery = ground truth). Building database built over time.
Layer 3: Confidence Scoring
geocoding_confidence × 0.3 + completeness × 0.2 + historical_success × 0.3 + pin_accuracy × 0.2. Score < 0.7 = flag for verification before dispatch.
Layer 4: Self-Healing Database
After every successful delivery, actual GPS vs stored coordinates compared. Drift > 100m = auto-update. After 100K deliveries = 50K+ verified addresses. Impossible to replicate.
Makani is UAE's official addressing system. Every building has a unique 10-digit number. We accept Makani as input, resolve to coordinates via free Makani API, cross-reference with commercial geocoding. For Dubai, this is free and more accurate than any commercial geocoder.
Route Optimization & Batching
Four optimization layers: single route, multi-stop TSP, batch VRP, and hub-spoke network routing. Instant delivery gets dedicated, direct assignment with pre-positioned drivers.
Single route: A → B
Multi-stop TSP (10-100 stops)
Batch VRP (50+ shipments)
Hub-spoke network
Operations & Integrations
Driver strategy, receiver experience, failed delivery handling, reverse logistics, multi-branch operations, marketplace integration, and channel connectors for WooCommerce, Shopify, and the public API.
Driver Fairness & Earnings Strategy
Own-fleet model: drivers are employees or long-term contractors, not anonymous gig workers. Compensation = Base Salary + Per-Delivery Commission + Performance Bonus + Tips (100% pass-through).
Driver Onboarding Flow
Online form + document upload
Criminal record, license verification
App training, QR scanning, COD handling, POD capture
Phone setup, app install, biometric registration
Monitored performance, enhanced oversight
Full access, bonus-eligible
Fairness Mechanisms
Equitable Distribution
No driver gets <60% or >140% of average daily deliveries. Premium deliveries distributed across all drivers. Fairness index tracked daily (target: StdDev/Mean < 0.3).
Zone Rotation
Drivers rotate through zones weekly. No permanent assignment to "easy" or "hard" zones. Specialized zones (cold-chain) assigned to qualified drivers only.
Earnings Transparency
Real-time earnings in app. Per-delivery breakdown. Weekly projection. Monthly bonus progress: "You're at 94% success — 1% more unlocks 500 AED bonus."
Anti-Gaming Protections
Cherry-picking detection (3+ rejections = temp queue removal). Fake delivery prevention (GPS + QR + photo CV). COD verification via receiver WhatsApp confirmation.
IoT & Hardware Integration
GPS trackers on all vehicles — real-time location even if driver phone is dead. Temperature sensors in refrigerated vans — continuous cold-chain monitoring with alerts if temp exceeds range. Smart scales at warehouses — auto-weigh packages, verify declared weight, flag discrepancies. Dash cams — incident recording, driver safety. Smart lockers — QR-authenticated pickup, 24/7 availability, no staff needed.
Receiver Experience Strategy
The receiver is a first-class user with agency, control, and delight. A branded, mobile-first tracking page (no app install required) serves as their primary interface.
First Contentful Paint: < 1s. Map Load: < 1.5s. Real-time Update Latency: < 2s. Lighthouse Score: > 95. No app install required. Works on any browser, any device.
Tracking Page Features
Live Map with ETA Countdown
Real-time driver location on interactive map. Updates every 5 seconds when out for delivery. ETA countdown narrows as driver approaches. Pulsing circle animation when driver is < 5 min away.
Reschedule & Redirect
Receiver can reschedule delivery to another day/time, redirect to a different address (same zone), or authorize safe drop — all from the tracking page, up to 30 min before driver arrives.
Payment Flexibility
Switch from COD to online payment directly on tracking page (card, Apple Pay, STC Pay). "Pay now and skip the cash at the door." Incentive: merchant-funded 5% discount for online switch.
Driver Communication
Masked phone call (privacy-preserving — neither sees real number). In-page chat with pre-built quick messages. Receiver can share live GPS location with driver for hard-to-find addresses.
Rating & Feedback
Post-delivery: 1-5 star rating on professionalism, package handling, timeliness. Optional text feedback. Tip option (5/10/20 AED, 100% to driver). Shareable delivery summary card.
One-Tap Return
Within merchant return window: one tap to initiate return. Select reason, upload photo evidence, schedule pickup. Full return tracking on same page.
Receiver Communication Flow
| Trigger | Channel | Message |
|---|---|---|
| Shipment created | "Your order from [Merchant] is being prepared. Track: [link]" | |
| Out for delivery | "Your package is out for delivery! ETA: [time]. Track: [link]" | |
| 5 min away | "[Driver] is 5 minutes away with your package" | |
| Delivered | "Delivered! View proof: [link]. Not happy? Start a return: [link]" | |
| Delivery failed | "We missed you! Reschedule: [link]. Reply with preferred time." | |
| Return picked up | "Return picked up. Refund in [X] days." |
Contactless & Smart Delivery Options
Failed Delivery Handling & Retry Logic
Target: >95% first-attempt success rate. Failed deliveries are the #1 cost driver in last-mile logistics. We attack this with 3-layer prevention and AI-powered retry scheduling.
Every failed attempt costs 15-25 AED (driver time, fuel, opportunity cost) with zero revenue. At 50K shipments/day, reducing failure from 20% to 5% saves ~$3.5M/year.
3-Layer Prevention System
Confidence > 0.7 or flag
HLR lookup — is number active?
Past failures at this address?
First-time + high-value = verify
"Will you be home tomorrow?" [Yes/Reschedule]
"Pay online now, skip cash at door"
Reschedule outdoor deliveries in sandstorms
"Package out for delivery today"
"Not home? [Reschedule/Safe Drop]"
"Driver arriving now"
Driver calls if no response
Smart Retry Logic
| Failure Reason | % of Failures | Auto-Retry Strategy |
|---|---|---|
| Not available | 40% | Check receiver's historical patterns. Morning fail → retry afternoon. Ask receiver via WhatsApp for preferred time. |
| Wrong address | 20% | Send pin-drop link via WhatsApp. After update → immediate re-queue. |
| Refused (changed mind) | 15% | Confirm with merchant. Initiate return. |
| Can't pay COD | 10% | Send online payment link via WhatsApp. Auto-convert to prepaid if paid. Retry within 4 hours. |
| Damaged package | 5% | Photo documentation, damage claim, replacement shipment. |
| Can't find location | 5% | Request live location via WhatsApp. After received → immediate re-dispatch. |
3 failed attempts → automatic return to merchant. Receiver explicitly cancels → immediate return. No additional charge for first retry. 2nd attempt: 50% of delivery charge. 3rd: 75%. Return: 50-100% of original charge. This pricing incentivizes merchants to provide accurate data.
Reverse Logistics / Return Architecture
Returns are a first-class workflow, not an afterthought. Six return types, QR-tracked through the entire reverse chain, with combined delivery + return routes that reduce cost by 50%.
Return Types
| Return Type | Initiated By | Trigger | SLA |
|---|---|---|---|
| Customer Return | Receiver | Changed mind, wrong item, damaged | Pickup within 24h |
| RTO (Failed Delivery) | System | Max delivery attempts exceeded | Immediate, auto-triggered |
| Exchange | Receiver + Merchant | Swap item (return old + deliver new) | Single trip: pickup old + deliver new simultaneously |
| Partial Return | Receiver | Multi-item order, return some | Pickup within 24h |
| Merchant Recall | Merchant | Product recall, shipping error | Urgent pickup |
| Rejected at Door | Driver/Receiver | Visible damage, wrong item | Immediate return, no extra trip |
Return Lifecycle (QR-Tracked)
Via tracking page or WhatsApp
Merchant auto-approve or manual
Combined with delivery routes
Driver has return package
Inspection & QC check
Auto-credit to receiver
RTO Reduction Strategy
Multi-Country / Multi-Warehouse Operations
Full organizational hierarchy: Platform → Region → Country → City → Branch/Warehouse. Not "future-ready" — operational from launch.
| Facility Type | Purpose | Staffing |
|---|---|---|
| Main Hub | Central sorting, intercity dispatch, returns processing, QR scan gates | 20-50 staff |
| City Hub | Intra-city sorting, driver dispatch, COD collection, QR inbound/outbound | 10-20 staff |
| Micro-Hub | Instant/minutes delivery staging, pre-positioned inventory | 3-5 staff |
| Fulfillment Warehouse | Full WMS: pick, pack, label, QR assign. Multi-level racking with bin locations | 20-100 staff |
| Dark Store | Quick-commerce fulfillment (Wadi Express), inventory pre-positioning | 10-30 staff |
| Cross-Dock Facility | No storage — packages arrive, get sorted, depart same day. Intercity transfers | 5-15 staff |
| Pickup Point / Locker | Customer self-collection, QR-authenticated pickup, 24/7 smart lockers | 0-2 staff |
Shipments crossing country borders automatically generate customs documentation (commercial invoice, packing list, HS codes), estimate duties/taxes for the receiver, and route through designated customs-clearance hubs. Supports both DDP (Delivered Duty Paid) and DDU (Delivered Duty Unpaid) models. Multi-currency settlement handles FX conversion with locked rates at shipment creation.
Marketplace Model for Wadi
Wadi is a tenant on the Shipment platform — no special code paths. Same APIs, same dispatch, same tracking. This is critical: what works for Wadi works for any marketplace.
Marketplace-Specific Flows
Multi-Vendor Order Splitting
Customer buys 3 items from 3 sellers → 3 separate shipments, linked by marketplace_order_id. System detects proximity and batches pickups. At destination hub, consolidates to one delivery driver if possible.
Fulfillment by Wadi (FBW)
Seller inventory pre-stored in Wadi warehouses. Order placed → QR pick list generated → pick, pack, label at warehouse → single shipment to customer. Fastest SLA, lowest cost.
Seller Pickup Model
Non-FBW sellers: order placed → seller confirms "ready" within 4 hours → pickup driver dispatched → QR scan at seller → deliver to customer. Late confirmation = seller penalized.
Wadi Express (Instant/Minutes)
Items pre-positioned in micro-hubs/dark stores. Order → pick in <5 min → nearest motorcycle dispatched → delivery in 15-30 min. Pre-positioned drivers at dark store locations.
Wadi Settlement Flow
Online or COD to driver
Minus shipping fee
Minus marketplace commission
Driver → Platform → Wadi → Seller
WordPress / WooCommerce Integration
Full WordPress plugin published on WordPress.org. Zero-code setup: install plugin, enter API key, done.
Plugin Capabilities
| Feature | Description |
|---|---|
| Live Shipping Rates | At checkout, fetches real-time rates from Shipment API based on cart weight, dimensions, and delivery address. Shows multiple service levels with prices and ETAs. |
| Auto Shipment Creation | Order paid → shipment auto-created via API. AWB + QR code generated. Label downloadable from WooCommerce order page. |
| Status Sync | Webhook-driven: shipment status changes auto-update WooCommerce order status. Order notes added at each milestone. |
| COD Support | Maps WooCommerce COD payment method to Shipment COD. COD fee shown at checkout. Configurable max amount and first-time restrictions. |
| Tracking Widget | Embeddable tracking widget for any WordPress page. Customer sees tracking in "My Orders" with link to full tracking page. |
| Bulk Operations | Select multiple orders → create shipments in bulk. Export labels as batch PDF. |
| Return Management | Customer-initiated returns from "My Account" page. Creates return shipment automatically. |
| Arabic + English | Full RTL support. Translation-ready with .po/.mo files. |
Shopify Integration
Shopify App published on Shopify App Store. Uses Carrier Service API for live rates + Fulfillment Service API for auto-shipment creation. Embedded dashboard via App Bridge + Polaris.
Shopify App Features
| Feature | Description |
|---|---|
| Live Rates at Checkout | Shopify calls our Carrier Service endpoint. We return rates for Instant, Same Day, Next Day, Economy with ETAs. |
| Auto-Fulfillment | Order paid → fulfillment request sent to our API → shipment created → tracking number returned to Shopify. |
| Embedded Dashboard | Full merchant dashboard embedded inside Shopify admin (App Bridge). View shipments, analytics, COD balance without leaving Shopify. |
| Bidirectional Webhooks | Shopify → Shipment: orders/create, orders/cancelled, refunds/create. Shipment → Shopify: picked_up, delivered, failed, returned. |
| COD via Payment Extension | Adds "Cash on Delivery" as payment method at Shopify checkout. COD collection and reconciliation handled by Shipment platform. |
| Label Printing | Print shipping labels with QR codes directly from Shopify order page. Supports thermal printers (ZPL) and standard PDF. |
Public API / Webhook / SDK Strategy
REST-first, versioned, idempotent, with comprehensive documentation. Every feature in the platform is an API. Our own dashboards consume the same APIs third parties use.
API Design Principles
Core API Resources
| Resource | Key Endpoints |
|---|---|
| Shipments | POST /shipments, GET /shipments/:id, POST /shipments/bulk, GET /shipments/:id/tracking, GET /shipments/:id/label |
| Quotes | POST /quotes (price + SLA + vehicle recommendation in <500ms) |
| Tracking | GET /tracking/:awb (public, no auth, rate-limited) |
| Addresses | POST /addresses/validate, POST /addresses/autocomplete |
| Webhooks | POST /webhooks, GET /webhooks/:id/deliveries |
| Returns | POST /shipments/:id/return |
Webhook System
HMAC-signed payloads (SHA-256). Exponential retry: immediate → 5m → 30m → 2h → 8h → 24h (6 attempts). Delivery logs via API. Auto-disable after 72h of continuous failure with email alert. Events: shipment.created, shipment.picked_up, shipment.delivered, shipment.delivery_failed, shipment.returned, cod.collected, settlement.completed.
SDKs & Sandbox
Official SDKs
JavaScript/TypeScript (@shipment/sdk), PHP (shipment/shipment-php), Python (shipment-sdk). Auto-generated from OpenAPI spec. Type-safe, with built-in retry logic.
Sandbox Environment
Full API parity with production. Test API keys (sk_test_...). Simulated dispatch, tracking, and webhooks. Simulation controls: POST /sandbox/simulate/:id/deliver. Data reset every 30 days. Zero charges.
Rate Limits
| Tier | Rate Limit | Burst |
|---|---|---|
| Free | 100 req/min | 20 req/sec |
| Growth | 1,000 req/min | 100 req/sec |
| Enterprise | 10,000 req/min | 500 req/sec |
| Custom | Negotiated | Negotiated |
Strategy & Execution
Security architecture, AI/ML at every layer, fraud prevention, operational analytics, cutting-edge 2026 technology stack, enterprise launch scope, risk analysis, innovative moat ideas, and the 28-day implementation sprint plan.
Security / Compliance / Audit
Enterprise-grade security architecture. Every layer protected. Every action audited. Every regulation complied with.
Security Architecture
DigitalOcean Cloud Firewall
All traffic encrypted
Per API key, per tenant
1h access, 7d refresh, rotatable
Tenant + branch isolation
Schema enforcement on every request
All PII, financial data
Driver ↔ receiver via platform relay
Every action logged, 7-year retention
PII Handling
| Data Type | Storage | Display | Retention |
|---|---|---|---|
| Customer phone | AES-256 encrypted | Masked: +9715****4567 | 2 years |
| Driver national ID | Encrypted + tokenized | Never shown in UI | Employment + 1 year |
| Bank accounts (IBAN) | Encrypted + tokenized | Last 4 digits only | Active + 1 year |
| GPS locations | Plaintext (operational) | Authorized roles only | 90 days, then aggregated |
| POD photos | Encrypted at rest (S3) | Authorized roles only | 1 year |
| QR scan chain | Blockchain-hashed | Full chain viewable | 7 years (immutable) |
Compliance
Disaster Recovery & Business Continuity
RPO: < 15 minutes (WAL archiving + streaming replication). RTO: < 15 minutes (automated failover). Multi-AZ: all services run across 2+ availability zones. Multi-region DR: Saudi Arabia as warm standby. Daily backups: encrypted, tested monthly with restore drills. Chaos engineering: quarterly failure injection tests to validate resilience.
Fraud Prevention & Dispute Management
10 fraud vectors identified and defended against. Real-time rule engine + ML anomaly detection from launch.
Fraud Vectors & Prevention
| Fraud Type | Actor | Prevention |
|---|---|---|
| Fake delivery | Driver | GPS must be within 200m of delivery point. Photo POD verified by on-device CV model. QR scan required. |
| COD skimming | Driver | Receiver confirms COD amount via WhatsApp. Discrepancies auto-flagged. Daily random cash audits. |
| GPS spoofing | Driver | Device fingerprinting, emulator detection, accelerometer cross-validation with GPS movement. |
| Return fraud | Receiver | QR scan at return pickup. Hub QC inspection with photos. Weight comparison vs original shipment. |
| Address fraud | Receiver | First-time + high-value COD = phone OTP verification before dispatch. |
| Account takeover | External | API key rotation, IP whitelisting, anomalous call pattern detection. |
| Phantom driver | Internal | Biometric selfie at shift start. Device binding (one device per driver account). |
| Insurance fraud | Merchant | Declared value vs product category cross-check. Photo evidence at pickup. |
Dispute Lifecycle
By merchant, receiver, or system
Pull POD, GPS trail, QR chain, COD record
ML model suggests resolution
Auto-resolve or escalate to human
SLA breach: auto-credit 50% shipping fee. COD mismatch < 5 AED: auto-adjust. "Not delivered" but POD exists: send POD to receiver for review. Damaged + insured: auto-initiate claim, credit declared value. These auto-resolutions handle 60%+ of disputes without human intervention.
AI & Automation — Built Into Every Layer
AI is not a feature — it's the nervous system of the platform. Every decision point has an AI model behind it, from dispatch to pricing to fraud detection to customer communication.
AI Dispatch (Reinforcement Learning)
RL agent learns optimal driver-to-shipment assignment by observing outcomes. Discovers patterns humans can't see: "assign this to the farther driver because their route passes through the pickup area." Trained on simulation, deployed with A/B testing against rule-based DSS.
ML ETA Prediction
Custom LightGBM model trained on delivery data: Google Maps ETA as baseline + parking time + building access + floor + COD payment time + weather + Ramadan adjustments. 30-40% more accurate than raw Google ETA.
LLM-Powered Customer Support
Claude API as the brain behind WhatsApp chatbot. Understands natural language: "where is my package?" "can you deliver tomorrow instead?" "the driver was rude." Handles 70%+ of support queries without humans. Arabic + English.
Computer Vision
On-device TensorFlow Lite models: POD photo quality verification (rejects blurry/dark), package damage detection (flags before receiver opens), QR code scanning (instant decode), driver selfie verification (anti-fraud).
Demand Forecasting (Prophet + LSTM)
Predicts order volume per zone per hour, 24-72h ahead. Factors: historical patterns, weather, holidays, Ramadan, payday cycles, marketing campaigns, trending products on Wadi. Drives driver scheduling and pre-positioning.
Fraud Detection (Anomaly ML)
Real-time anomaly detection on delivery patterns, COD amounts, GPS trails, driver behavior. Flags fake deliveries, COD skimming, GPS spoofing, and account abuse. Auto-escalates to investigation queue.
Operational Analytics / KPI Model
30+ KPIs across operations, financial, customer experience, and driver metrics. Real-time dashboards in custom admin console. Operations Command Center for dispatchers. Merchant analytics dashboard with export.
Unit Economics & Profit Model
This section answers the critical question: how does Shipment make money, how much, and when does it become profitable? All numbers are based on UAE market rates, own-fleet model with 20 drivers at launch, and realistic order volumes.
Revenue Streams — How We Earn Per Delivery
| Revenue Stream | How It Works | Rate (AED) | % of Orders | Weighted / Delivery (AED) |
|---|---|---|---|---|
| Shipping Fee (Flat) | 10 AED fixed — same for delivery or return. Simple, competitive. | 10.00 | 100% | 10.00 |
| COD Handling Fee | 2.5% of COD amount. Avg COD = 300 AED. | 7.50 | 60% | 4.50 |
| Instant/Express Premium | 30-min delivery premium on top of flat fee. | 20.00 avg | 10% | 2.00 |
| Return Fee | Same 10 AED flat for return pickup. | 10.00 | 8% | 0.80 |
| Wadi Fulfillment (FBW) | Pick (3) + Pack (2) + Label (0.5) on top of shipping. | 5.50 | 15% | 0.83 |
| Merchant Subscriptions | Growth (500) / Enterprise (1,500 AED/month) plans. | ~1.50 | 100% | 1.50 |
| Insurance | 1.5% of declared value. Optional. | 7.50 | 5% | 0.38 |
| COD Float Income | Cash held 3-7 days before settlement. | Passive | — | |
| Blended Avg Revenue Per Delivery | 20.01 AED | |||
10 AED is aggressively competitive — most UAE couriers charge 15-25 AED. This is our merchant acquisition weapon. The flat fee is razor-thin but blended revenue reaches 20 AED through: COD fees (60% of orders), instant premiums (3x markup), FBW warehouse fees, merchant subscriptions, and scale efficiency.
Cost Structure — What Each Delivery Costs Us
All Fixed Monthly Costs (20 Drivers + Operations)
| Cost Item | Monthly (AED) | Notes |
|---|---|---|
| Driver Costs (20 Drivers) | ||
| Driver base salary (20 × 2,500) | 50,000 | Fixed monthly salary per driver |
| Fuel (20 drivers × 3,000 km/month @ 0.70 AED/km) | 42,000 | Avg 3,000 km/driver/month |
| Vehicle maintenance (20 × 350) | 7,000 | Tires, oil changes, service |
| Vehicle insurance (20 × 1,200/year ÷ 12) | 2,000 | Commercial third-party, cheapest UAE option (RSA ~1,200 AED/yr) |
| Operations Costs | ||
| Operations manager (1 per 20 vehicles) | 7,000 | Manages dispatch, drivers, daily ops |
| Technology (DigitalOcean + HERE Maps + 360dialog + Claude Haiku + Sentry) | 3,000 | Cloud infra 1,250 + HERE Maps 800 + 360dialog WhatsApp 400 + Claude AI 75 + Sentry 95 + misc 280. Redis Streams for events. GitHub Free. DO built-in monitoring. |
| Miscellaneous (labels, supplies) | 2,000 | Thermal labels, QR stickers |
| TOTAL FIXED MONTHLY COST | 113,000 | Drivers: 101,000 + Ops: 12,000 |
Vehicle purchase: 3,500 AED/vehicle × 20 = 70,000 AED (owned forever).
Uniforms + safety equipment: ~500 AED/driver × 20 = 10,000 AED.
Vehicle branding/wrapping: ~1,500 AED/vehicle × 20 = 30,000 AED.
Total one-time startup: ~110,000 AED.
Driver visa, personal insurance, and phone/device are the driver's own responsibility.
Commercial third-party (required by UAE law). Cheapest: RSA Insurance (~1,000-1,200 AED/year), Orient Insurance (~1,100-1,400), Oman Insurance (~1,200-1,500). We budget 1,200 AED/year per vehicle = 100 AED/month. Third-party is sufficient for delivery operations.
Variable Cost Per Delivery
| Variable Cost | Per Delivery (AED) | Notes |
|---|---|---|
| Driver per-delivery commission | 2.00 | Fixed 2 AED max per order, offset by 2,500 AED base salary + bonuses |
| Fuel per delivery (avg 6 km optimized route) | 2.50 | Route optimization reduces avg trip distance |
| WhatsApp notifications (4 msgs/delivery) | 0.08 | 360dialog + Meta conversation pricing |
| Maps API cost per delivery (HERE) | 0.10 | HERE Maps with aggressive caching |
| Label + QR code printing | 0.20 | Thermal printer, bulk label stock |
| COD cash handling risk reserve (0.3%) | 0.90 | QR chain reduces risk significantly |
| Failed delivery cost (prorated @ 5% rate) | 0.75 | AI prevention reduces failure rate to 5% |
| Total Variable Cost Per Delivery | 6.53 |
Unit Economics Per Delivery
Blended revenue per delivery: 20.01 AED. Variable cost per delivery: 6.53 AED. Contribution margin: 13.48 AED (67.4%). The 10 AED flat fee is the merchant acquisition weapon — COD fees, instant premiums, FBW, and subscriptions push blended revenue to 20 AED while costs stay lean.
Breakeven Analysis (20 Drivers)
= 113,000 AED / 13.48 AED per delivery
= 8,383 deliveries/month
= 279 deliveries/day (30-day month)
= ~14 deliveries per driver per day
Trivially easy. Industry benchmarks in UAE: Noon drivers do 25-35 deliveries/day. Aramex drivers do 20-30. Even the worst-performing drivers do 15+. We break even at just 14 — less than half of what average drivers do. Everything above 14 is pure profit. At 30/driver (industry average), we're making 129K AED/month.
Profit Scenarios (Monthly, 20 Drivers)
| Scenario | Del/Driver/Day | Del/Month | Revenue (AED) | Total Cost (AED) | Net Profit (AED) | Margin |
|---|---|---|---|---|---|---|
| Below Breakeven | 10 | 6,000 | 120,060 | 152,180 | −32,120 | −27% |
| At Breakeven | 14 | 8,383 | 167,880 | 167,880 | 0 | 0% |
| Good (20/driver) | 20 | 12,000 | 240,120 | 191,360 | 48,760 | 20.3% |
| Strong (25/driver) | 25 | 15,000 | 300,150 | 210,950 | 89,200 | 29.7% |
| Great (30/driver) | 30 | 18,000 | 360,180 | 230,540 | 129,640 | 36.0% |
| Excellent (35/driver) | 35 | 21,000 | 420,210 | 250,130 | 170,080 | 40.5% |
| Outstanding (40/driver) | 40 | 24,000 | 480,240 | 269,720 | 210,520 | 43.8% |
At 10 AED flat, profitability requires high volume per driver. This is exactly the Noon/Amazon model — low per-unit margin, massive volume. The 10 AED price is a market penetration weapon that undercuts every competitor, locks in merchants, and builds the density needed for profitable operations. Once density is established, margins improve dramatically because fixed costs are spread across more deliveries.
Where the REAL Money Is (Beyond Flat Fee)
The 10 AED fee gets merchants in the door. The real profit comes from:
1. Instant delivery premium: 100 instant orders/day × 20 AED premium = 60,000 AED/month
2. FBW warehouse fees: 300 FBW orders/day × 5.50 AED = 49,500 AED/month
3. Enterprise subscriptions: 30 merchants × 1,000 AED avg = 30,000 AED/month
4. COD float income: ~15,000 AED/year at current volume, 150,000 AED/year at scale
Combined: These add ~140,000 AED/month on top of base delivery profits.
How Top UAE Logistics Companies Maximize Profit
Density Over Distance
Noon, Aramex, and Careem focus on dense urban zones (Dubai Marina, JLT, Downtown, Business Bay) where 5-8 deliveries happen within 2 km. More drops per hour = higher revenue per driver hour. Our zone-based dispatch + route optimization targets this.
COD Float Income
COD cash sits in our account 3-7 days before merchant settlement. At 500 deliveries/day with avg 300 AED COD (60% COD rate): float = ~630,000 AED constantly held. At 5% annual return = 31,500 AED/year free income. At scale (5,000/day) = 315,000 AED/year.
Return Revenue
Returns are charged at 50-75% of original shipping. With ~8% return rate: 500 deliveries/day × 8% = 40 returns/day × 18 AED avg = 21,600 AED/month extra revenue on work you're already partially doing (combined routes).
Wadi Fulfillment Fees
For FBW (Fulfillment by Wadi) orders: charge pick (3 AED) + pack (2 AED) + label (0.5 AED) = 5.50 AED per unit ON TOP of shipping fee. At 200 FBW orders/day = 33,000 AED/month additional revenue from warehouse operations.
Merchant Subscription Revenue
Enterprise merchants pay monthly subscription (500-2,000 AED/month) for premium features: priority dispatch, custom tracking pages, dedicated account manager, advanced analytics. 20 enterprise merchants = 20,000-40,000 AED/month recurring.
Instant/Minutes Premium
Instant delivery charges 3-5x base rate. A 15 AED standard delivery becomes 45-75 AED for instant. Even at 10% of volume being instant: 50 instant deliveries/day × 40 AED premium = 60,000 AED/month pure premium.
Scaling Economics: 20 → 50 → 100 Drivers (@ 40 deliveries/driver/day)
| Metric | 20 Drivers | 50 Drivers | 100 Drivers |
|---|---|---|---|
| Deliveries/day | 800 | 2,000 | 4,000 |
| Deliveries/month | 24,000 | 60,000 | 120,000 |
| Monthly revenue (@ 20.01 AED blended) | 480,240 | 1,200,600 | 2,401,200 |
| Fixed costs (drivers + ops) | 113,000 | 278,500 | 547,000 |
| Variable costs (@ 6.53/delivery) | 156,720 | 391,800 | 783,600 |
| Total costs | 269,720 | 670,300 | 1,330,600 |
| Net profit/month | 210,520 | 530,300 | 1,070,600 |
| Net margin | 43.8% | 44.2% | 44.6% |
| Annual net profit | 2.53M AED | 6.36M AED | 12.85M AED |
Fixed ops costs don't double when drivers double. One dispatcher handles 20-40 drivers. One hub handles 50 drivers. Tech costs are nearly flat. At 100 drivers, ops overhead per delivery drops 40%. This is why Noon and Amazon invest in scale — the unit economics get dramatically better. Our target: 20%+ net margin at 50+ drivers.
Time to Profitability
Ramp-up. 150-250 deliveries/day. Building zone density.
Hit 279/day breakeven (14/driver). Trivially easy target.
600-900/day (30-45/driver). Strong margins. Adding drivers.
Scale mode. 50+ drivers. 1,500+/day. 20%+ margins. Expand to Abu Dhabi.
"In logistics, you lose money on the first delivery to a zone and make money on every subsequent one. The game is density — the more deliveries per square kilometer per hour, the more profitable each delivery becomes. This is why we focus on zone domination before geographic expansion." — Shipment Financial Strategy, March 2026
Launch Scope — Full Enterprise System
No MVP. No feature gates. The launch system includes every core capability: multi-country operations, multi-currency settlement, multi-warehouse management, QR chain-of-custody, AI-powered dispatch, real-time tracking, computer vision POD verification, and full API platform. This is the system that makes competitors obsolete.
Auto-dispatch accuracy: >90%. QR scan chain completion: 100%. Tracking page: <1s load. Driver app crash rate: <0.5%. COD reconciliation: >99%. API uptime: >99.9%. Multi-country: UAE + 2 GCC countries operational. Cross-border: customs docs auto-generated.
2026 Cutting-Edge Technology Stack
AI/ML at Core
Not bolted on later. LLM-powered customer support chatbot (Claude API), ML demand forecasting, computer vision for POD quality + damage detection, reinforcement learning for dispatch optimization. All from day one.
Computer Vision POD
Driver captures delivery photo → on-device ML model verifies: package visible, not blurry, correct location, no damage. Rejects bad photos instantly. Detects package damage automatically and flags before receiver opens.
Edge Computing on Driver App
ML models run ON the phone (TensorFlow Lite): QR scanning, POD verification, address matching, offline dispatch. No internet needed for core operations. Sync when connected.
Blockchain Audit Trail
Every QR scan event is hashed into an immutable chain. Tamper-proof chain-of-custody. Exportable for insurance claims, legal disputes, and customs verification.
Real-Time Event Streaming
Redis Streams for inter-service events at launch (included in Redis — zero extra cost). Upgrade to Apache Kafka when volume exceeds 50K deliveries/day. Every event (shipment, driver, financial) streams in real-time to all consumers.
Multi-Region Deployment
DigitalOcean regions: UAE (primary), Saudi, Europe. Data residency compliance per country. Active-active for global availability. Sub-100ms API responses in every served region.
Estimated Enterprise Infrastructure Cost
| Item | Monthly Cost |
|---|---|
| DigitalOcean (DOKS multi-region + managed DBs + Redis Cluster + Spaces + LBs) | ~$1,500-2,500 |
| HERE Maps (geocoding, routing, tour planning, geofencing, fleet tracking) | ~$220 |
| 360dialog (WhatsApp API) + Meta conversation fees | ~$110 |
| Claude API Haiku (AI chatbot fallback) | ~$20 |
| Sentry (error tracking) + GitHub Free | ~$26 |
| Redis Streams (event streaming — included in Redis) | $0 |
| Domain + misc | ~$55 |
| Total Technology Infrastructure | ~$820/month (~3,000 AED) |
DHL spends $2B+/year on technology. Noon spends $50M+/year on logistics tech. Our $3-6K/month buys us a system that is architecturally superior to both — because we're building on 2026 cloud-native infrastructure while they maintain 2010-era monoliths. This cost scales linearly with volume, not exponentially.
Risks, Tradeoffs & Anti-Patterns
Critical Risks
| Risk | Probability | Impact | Mitigation |
|---|---|---|---|
| Driver acquisition & retention | HIGH | CRITICAL | Competitive pay, transparent earnings, good app UX, streak bonuses, partner treatment |
| COD cash handling fraud/loss | HIGH | HIGH | Daily audits, cash limits, dual-confirmation for high-value, QR scan at every handoff, insurance |
| Map API costs at scale | MEDIUM | HIGH | HERE Maps as single provider (10x cheaper than Google). Aggressive caching. Logistics-grade features included. |
| Address quality in UAE | HIGH | MEDIUM | Makani integration, pin drop, driver-confirmed GPS locations, crowdsourced address database |
| System reliability at scale | MEDIUM | CRITICAL | Multi-AZ, automated failover, circuit breakers, chaos engineering, <15min RTO |
| Regulatory compliance (multi-country) | MEDIUM | HIGH | Legal counsel per country. Data residency per jurisdiction. Transport permits. Central Bank COD license. |
| Competition from incumbents | HIGH | MEDIUM | Technology superiority (AI, QR chain, real-time). Better UX. Lower cost. API-first openness. |
| 28-day timeline execution risk | MEDIUM | HIGH | All architecture decisions pre-made. Zero planning weeks. 15+ engineers in parallel. Daily standups. No scope creep. |
Architecture Tradeoffs
Microservices from Day 1
Chosen over modular monolith. Tradeoff: harder to build initially but eliminates painful monolith-to-microservices migration later. With Kafka from day 1, services are properly decoupled.
PostgreSQL over NoSQL
Chosen over MongoDB. Why: ACID for financial data, PostGIS for geo, RLS for multi-tenancy, JSONB for flexibility. NoSQL would need separate solutions for transactions and geo.
DigitalOcean over AWS
Chosen over AWS. Why: Simpler, 30-50% cheaper, sufficient for our needs. If we need specific AWS services (SageMaker), we can go hybrid later.
HERE Maps over Google Maps
Chosen over Google Maps. Why: 10x cheaper at scale. HERE has logistics-specific features (tour planning, geofencing, fleet tracking, vehicle routing). Quality gap in UAE is small.
Anti-Patterns to Avoid
WhatsApp-based dispatch: build the dispatch console first. Excel reconciliation: COD service from day 1. Shared databases between services: each service owns its data. Driver favoritism: algorithmic dispatch with fairness scoring. Ignoring failed deliveries: automated retry with root cause analysis. Custom everything: buy commodities, build differentiators.
"Crazy" Innovative Ideas That Create a Moat
Money-Back Delivery Guarantee
"We promise delivery by 4 PM. If we miss it, shipping is free." No courier in MENA offers this.
Receiver Reputation System
Receivers build reputation based on availability, accuracy, and behavior. High-rep gets priority slots.
Crowdsourced Address Database
Every delivery GPS-confirms the exact door location. Over 100K deliveries = 50K+ verified addresses. Impossible to replicate.
Driver Streak Bonuses
"5 on-time in a row = +20 AED. 10 without failure = +50 AED." Gamification without punishment.
Predictive Inventory Positioning
Pre-ship popular items to nearest micro-hub before orders arrive. 15-minute delivery becomes possible.
Open Delivery Network
Other courier companies plug into our infrastructure. Their drivers, our tech. Network effect creates moat.
Final System Blueprint
React Native
React
Next.js SSR
React
+ PostGIS
Cache + Pub/Sub
Event Bus
S3 Storage
| Property | Specification |
|---|---|
| Architecture | Microservices from day one (NestJS + Kafka + DOKS) |
| Availability | 99.95% at launch → 99.99% at scale |
| Latency | P95 < 200ms API, < 3s tracking updates |
| Throughput | 500/day → 1M+/day |
Implementation Readiness Pack
Enterprise Engineering Workstreams — 28-Day Sprint
Platform Foundation + Shipment Core (Days 1-5)
DOKS infra, CI/CD, NestJS monorepo, DB schema, auth, API gateway. Shipment CRUD, AWB + labels, address validation with HERE Maps.
Engines: Pricing + Vehicle + Dispatch (Days 3-8)
Rate card engine, pricing formula, vehicle recommendation, DSS scoring, auto-dispatch, driver location tracking via WebSocket + Redis. All 3 engines built in parallel by separate devs.
Driver App — Full Flow (Days 4-14)
React Native: OTP auth, shift start/end, delivery queue, Google Maps/Waze navigation, pickup scan, delivery POD (photo + signature), COD collection, earnings view. Offline queue sync.
Merchant Dashboard + Admin Console (Days 6-16)
Merchant: shipment creation (single + CSV bulk), tracking with map, COD balance, label download, API key management. Admin: tenant, user, branch, zone, rate card CRUD.
Tracking Page + WhatsApp Notifications (Days 10-18)
Next.js SSR tracking page (live map, ETA, status timeline, Arabic/English). 360dialog WhatsApp integration for 4 touchpoints + buttons + Claude Haiku chatbot fallback
Dispatch Console + Branch Console (Days 12-20)
Dispatch: real-time map with drivers + shipments, alert panel, manual assign/reassign. Branch: inbound/outbound scan, sorting view, COD deposit recording.
Financial Module + Webhooks (Days 14-22)
COD ledger (per-delivery tracking), driver cash balance, deposit reconciliation, merchant invoicing, settlement tracking. Webhook service with retry logic.
Wadi Integration + E2E Testing (Days 18-25)
Wadi API client, shipment creation from orders, webhook status sync, COD flow. Full end-to-end testing of complete lifecycle.
Load Testing + Security + Bug Fixes (Days 23-27)
Load test at 2x expected volume, security audit, penetration testing, critical bug fixes, performance tuning.
Soft Launch + Go Live (Day 28)
Soft launch with 50-100 shipments/day, driver training, merchant onboarding, monitoring war room. Full launch ramp-up.
Recommended Build Sequence — 4-Week Sprint
4-week timeline requires maximum parallelization: all 10 workstreams run concurrently with dedicated owners. Daily standups, no blockers tolerated. Team of 11+ engineers working simultaneously across backend, mobile, frontend, and DevOps.
| Day | Deliverables | Team |
|---|---|---|
| Day 1-2 | DOKS cluster live, CI/CD pipeline, NestJS monorepo, PostgreSQL schema, Redis, auth service | DevOps + 1 Backend |
| Day 3-5 | Shipment CRUD API, AWB generation, label PDF, address validation (HERE Maps), quote API | 2 Backend |
| Day 3-8 | Pricing engine, vehicle recommendation engine, dispatch engine (DSS + auto-assign) | 2 Backend (parallel) |
| Day 4-14 | Driver app: auth, shift, queue, navigation, pickup/delivery flows, POD, COD, offline mode | 2 Mobile |
| Day 6-16 | Merchant dashboard + admin console (shipment creation, tracking, rate cards, user mgmt) | 2 Frontend |
| Day 10-18 | Tracking page (Next.js SSR, live map, ETA) + WhatsApp (360dialog) + buttons + Claude Haiku chatbot | 1 Frontend + 1 Backend |
| Day 12-20 | Dispatch console (real-time map, alerts) + branch console (scan, sort, COD deposit) | 1 Frontend |
| Day 14-22 | COD ledger, reconciliation, invoicing, settlement, webhook service with retry | 1 Backend |
| Day 18-25 | Wadi integration (API client, webhooks, COD flow) + full E2E testing | 1 Backend + QA |
| Day 23-27 | Load testing (2x volume), security audit, critical bug fixes, performance tuning | All hands |
| Day 28 | Soft launch — 50-100 shipments/day, driver training, merchant onboarding | Full team |
This is not an MVP build — it's a full enterprise system in 28 days. Multi-country, multi-currency, multi-warehouse, QR chain-of-custody, AI dispatch, computer vision, LLM chatbot, real-time Kafka streaming — all at launch. With a 15+ person team running all workstreams in parallel, every capability ships on Day 28. The technology stack is decided. The architecture is designed. Day 1 = code starts.
"This document is a living artifact. Update it as decisions are made, assumptions are validated or invalidated, and the product evolves." — Shipment Architecture & Product Strategy, March 2026