IT204 E Commerce

E CommerceUnit 918 min read

E-Commerce Applications & Case Studies: Models, Platforms & Real-World Impact

Unit 9 of E Commerce explores how e-commerce theories apply in practice through real-world business models (B2B, B2C, C2C), platform architectures (marketplaces, social commerce), and case studies (eSewa, Daraz, NEPSE). It covers technical infrastructure, customer engagement strategies, and emerging trends like AI-driv

TAKEAWAYS:

  • E-commerce applications are built on technical frameworks (hardware, software, networks) that enable seamless transactions, but their success depends on business model alignment (e.g., Daraz’s marketplace vs. eSewa’s payment gateway).
  • Case studies (like NEPSE’s digital trading platform or Pathao’s ride-hailing) reveal how environmental factors (regulation, infrastructure, user behavior) shape e-commerce strategies.
  • Social media and software agents (chatbots, recommendation algorithms) are critical for customer engagement, but their effectiveness varies by platform (e.g., WhatsApp Business vs. Instagram Shopping).
  • B2B vs. B2G vs. C2C models require different infrastructure (EDI for B2B, secure portals for B2G, peer-to-peer networks for C2C), each with unique security and scalability challenges.
  • Emerging technologies (blockchain for supply chain, AI for dynamic pricing) are transforming e-commerce, but their adoption depends on cost, regulation, and user trust.
  • Exam focus: Compare models, analyze case studies for infrastructure/strategy trade-offs, and explain how real-world constraints (e.g., Nepal’s internet speeds, NTC’s bandwidth limits) impact e-commerce design.

1. E-Commerce Applications: The Core Models and Their Technical Backbone

E-commerce applications are not just websites or apps—they are systems that integrate hardware, software, networks, and business logic to enable transactions. The three primary models—Business-to-Business (B2B), Business-to-Consumer (B2C), and Consumer-to-Consumer (C2C)—each rely on distinct technical infrastructures but share common components.

1.1 The Technical Infrastructure Framework

Every e-commerce application sits atop a layered technical foundation. Below is the OSI-inspired e-commerce stack, adapted for business needs:

Smartphones (Android/iOS)PCs/LaptopsIoT Devices (e.g., Smart TVs)User DevicesInternet (NTC/Ncell)4G/5G (Nepal Telecom)Satellite (e.g., Starlink for rural areas)Network LayerCloud (AWS, Azure, Google Cloud)On-Premise (NTC Data Centers, local servers)Server LayerWeb Servers (Apache, Nginx)App Servers (Node.js, Django, Laravel)Application LayerSQL (MySQL, PostgreSQL)NoSQL (MongoDB, Firebase)Database LayerSSL/TLS EncryptionFirewalls (e.g., Cloudflare)DDoS ProtectionSecurity LayerE-Commerce Technical Stack (Nepal Context)
Layered architecture of e-commerce platforms in Nepal, adapted from OSI model for business needs.

Key Components Explained:

  • Network Layer: In Nepal, NTC’s fiber-optic backbone and Ncell’s 4G/5G networks are critical for low-latency transactions. For example, Pathao’s ride-hailing app relies on real-time GPS updates, which fail if network latency exceeds 200ms.
  • Server Layer: Daraz’s cloud infrastructure (hosted on Alibaba Cloud) must handle 10,000+ concurrent users during sales events. Nepalese startups often use AWS or local data centers (e.g., Nepal Telecom’s hosting services) to reduce latency.
  • Database Layer: eSewa’s transaction database uses PostgreSQL for ACID compliance (ensuring no double-spending). NEPSE’s trading platform uses real-time Oracle databases to match buy/sell orders in milliseconds.
  • Security Layer: SSL/TLS certificates (like those issued by GlobalSign) encrypt data between users and servers. Khalti’s 3D Secure (3DS) protocol adds an extra authentication layer for high-value transactions.

1.2 Worked Example: How Daraz’s Marketplace Handles a Peak Load

Scenario: During the Dashain sale, Daraz expects 50% more orders than usual. How does its infrastructure scale?

2077 BS (2020 AD)Daraz NepalLaunched (Alibaba Grou2078 BS (2021 AD)Peak Load Test:Dashain/Tihar Season (2079 BS (2022 AD)Auto-Scaling CloudServers (AWS) Deployed2080 BS (2023 AD)AI-Based InventoryPrediction (Reduced St
Key milestones in Daraz Nepal’s infrastructure scaling for peak demand.
Component Normal Day Dashain Sale (Peak Load) Nepal-Specific Challenge
Network Ncell 4G (avg. 50ms latency) NTC fiber + CDN (Cloudflare) Rural areas may drop to 3G (300ms+)
Servers 500 AWS EC2 instances Auto-scaling to 5,000 instances Power outages in Kathmandu
Database MySQL (read replicas) MongoDB sharding + Redis caching High transaction fees on local banks
Payment Gateway eSewa/Khalti (90% success) Failover to IME Pay NPR 100+ transaction limit delays
CDN Cloudflare (global) Edge caching in Singapore/India Slow DNS resolution in Nepal

Why This Matters for Exams:

  • Question: "How does Daraz ensure availability during peak traffic?" Answer: Horizontal scaling (more servers), caching (Redis), and geo-distributed CDNs reduce latency. In Nepal, local failovers (e.g., switching to IME Pay) handle payment gateway limits.

2. Real-World E-Commerce Models: B2B, B2C, C2C, and B2G

Each model has unique technical requirements and business strategies. Below is a comparison table:

Model Definition Key Infrastructure Nepal Example Security/Compliance Needs
B2B Business sells to another business EDI, API integrations, ERP systems Nepal Oil Corporation (NOC) suppliers GDPR-like data sharing agreements
B2C Business sells directly to consumers Mobile apps, payment gateways, CRM systems Daraz, Hamrobazaar PCI-DSS for payments, COPPA for kids’ data
C2C Consumer sells to another consumer Peer-to-peer networks, escrow systems OLX Nepal, Facebook Marketplace Fraud detection, buyer/seller ratings
B2G Business sells to government Secure portals, e-procurement systems e-Governance Nepal (eGN) vendors IT Act 2006 compliance, digital signatures

2.1 Case Study: NEPSE’s Digital Trading Platform (B2B + B2C Hybrid)

Nepal Stock Exchange (NEPSE) uses a hybrid B2B/B2C model where:

  • Institutional investors (B2B): Use API-based trading with real-time data feeds (provided by Bloomberg Terminal).
  • Retail investors (B2C): Use the NEPSE mobile app with biometric authentication (fingerprint + OTP).

Technical Challenges in Nepal:

  1. Low Internet Penetration: Only 50% of Nepal’s population has smartphones (CIA World Factbook).
    • Solution: NEPSE offers USSD-based trading (dial *123#) for feature phones.
  2. Power Outages: Kathmandu’s load-shedding causes trading halts.
    • Solution: Backup generators at NEPSE’s data center in Kathmandu.
  3. Fraud Prevention: Ponzi schemes (e.g., One Network scam) require AI-driven anomaly detection.

Worked Example: How NEPSE Prevents Fraud

  • Step 1: User logs in via biometric + OTP.
  • Step 2: AI checks for unusual patterns (e.g., 100 trades in 1 minute).
  • Step 3: If flagged, manual review by NEPSE’s compliance team.
  • Step 4: Blockchain audit trail (since 2021) ensures transparency.

3. Social Media and Software Agents: The Invisible Engines of E-Commerce

3.1 Social Commerce: How Facebook and Instagram Drive Sales

Social media platforms (Facebook, Instagram, TikTok) are now e-commerce enablers, not just marketing tools. In Nepal:

  • Facebook Marketplace: 30% of C2C transactions (OLX Nepal competitor).
  • Instagram Shopping: Hamrobazaar uses Instagram’s checkout feature to sell handmade goods.
  • WhatsApp Business: Local kirana shops use catalog links to take orders.

Technical Workflow:

sequenceDiagram
    participant User
    participant Instagram
    participant Merchant
    participant PaymentGateway
    participant InventorySystem

    User->>Instagram: Views product in "Shop" tab
    Instagram->>Merchant: Fetches product via API
    User->>Instagram: Clicks "Buy Now"
    Instagram->>PaymentGateway: Redirects to Khalti/eSewa
    PaymentGateway-->>User: Returns to Instagram after payment
    Instagram->>InventorySystem: Updates stock (via ERP)
    Merchant->>User: Ships order (Pathao/Dhlokhi)

Why This Works in Nepal:

  • Low smartphone penetration: Feature phone users can access Marketplace via Facebook Lite.
  • Trust issues: Verified seller badges reduce fraud (e.g., Hamrobazaar’s "Gold Seller" status).

3.2 Software Agents: Chatbots and Recommendation Systems

Software agents automate customer interactions and personalize experiences.

Text: ‘Check balance’Voice: ‘Pay electricity bill’User InputIntent RecognitionEntity Extraction (e.g., ‘1000 rupees’)NLP ProcessingDatabase Query (User Balance)Payment Gateway TriggerActionText/Voice OutputConfirmation SMSResponseeSewa Chatbot Workflow
How eSewa’s chatbot processes user requests using NLP and integrates with backend systems.
Agent Type Example in Nepal How It Works Exam Tip
Chatbots eSewa’s WhatsApp bot NLP (Natural Language Processing) + FAQ database Ask: "How does eSewa’s bot handle ‘refund’ queries?"
Recommendation AI Daraz’s "Frequently Bought Together" Collaborative filtering (like Amazon) Compare with traditional cross-selling
Price Optimization Nepal’s fuel price apps (e.g., PetrolPump) Scrapes data from NOC, HPCL websites Discuss real-time vs. static pricing

Worked Example: eSewa’s Chatbot for Refunds

  1. User: "I didn’t receive my order from Hamrobazaar."
  2. Bot: "I’m sorry! Please share your order ID and payment screenshot."
  3. Bot checks:
    • Database: Order status = "Shipped" → "Check with Pathao."
    • Payment Gateway: Refund initiated if dispute confirmed.
  4. User: "I want a refund."
  5. Bot: "Refund processed to Khalti account #12345678. ETA: 3-5 days."

4. Environmental Factors Impacting E-Commerce in Nepal

E-commerce in Nepal is shaped by unique challenges not found in Western markets:

Factor Impact on E-Commerce Example Mitigation Strategy
Internet Speed High latency → abandoned carts NTC’s 100Mbps fiber vs. rural 2G Edge caching (Cloudflare in India)
Payment Infrastructure Limited card usage → cash-on-delivery dominance 90% of Daraz orders use COD eSewa/Khalti push notifications
Logistics Poor road networks → delayed deliveries Pathao’s "No Delivery Zone" in remote areas Dhlokhi’s drone trials (2023)
Regulation IT Act 2006 → strict data privacy rules NEPSE’s biometric authentication GDPR-like compliance training
Power Outages Data center downtime → trading halts NEPSE’s backup generators Solar-powered data centers (future)

5.1 Blockchain for Supply Chain Transparency

Problem: In Nepal, counterfeit goods (e.g., fake Ncell chargers) cost businesses $50M/year. Solution: Blockchain-based tracking (used by Daraz’s "Authentic Seller" program).

How It Works:

  1. Product tagged with QR code (scannable via Daraz app).
  2. Each step logged on blockchain (manufacturer → warehouse → delivery).
  3. Consumer verifies authenticity in real-time.

5.2 AI for Dynamic Pricing

Example: Nepal’s hotel industry uses AI pricing tools (like RateGain) to adjust room prices based on:

  • Demand (e.g., Dashain festival → +300% prices).
  • Competitor rates (e.g., Hotel Everest vs. Hotel Himalaya).
  • Weather (e.g., clear skies in Pokhara → higher bookings).

Worked Example: Pokhara Hotel Pricing

Factor Low Demand (Monsoon) High Demand (Dashain)
Base Price NPR 3,000/night NPR 12,000/night
AI Adjustment -20% (NPR 2,400) +150% (NPR 18,000)
Competitor Price Hotel Annapurna: NPR 2,800 Hotel Himalaya: NPR 15,000
Final Price NPR 2,500 NPR 17,000

## In the Real World

  1. eSewa’s Payment Gateway

    • Idea Used: Multi-factor authentication (MFA) + real-time fraud detection.
    • How It Works: When you pay via eSewa, the system checks:
      • Device fingerprint (IP, browser, OS).
      • Behavioral biometrics (typing speed, mouse movements).
      • Transaction velocity (e.g., 5 payments in 1 minute → blocked).
    • Nepal Impact: Reduced fraud by 40% since 2020 (eSewa Annual Report).
  2. Daraz’s "Lightning Deal" Algorithm

    • Idea Used: Collaborative filtering (like Netflix recommendations).
    • How It Works:
      • Tracks user browsing history (e.g., you viewed iPhone 13).
      • Shows limited-time discounts to similar buyers (e.g., "Only 3 left at NPR 50,000!").
    • Nepal Example: During Tihar sales, Daraz’s AI pushed diya lamps to users who bought incense sticks earlier.
  3. Pathao’s Ride-Hailing Infrastructure

    • Idea Used: Geohashing + real-time routing.
    • How It Works:
      • Step 1: User requests ride in Kathmandu.
      • Step 2: Pathao’s algorithm checks:
        • Nearest driver (within 500m).
        • Traffic data (from Google Maps API).
        • Driver availability (e.g., not in a school zone).
      • Step 3: Dynamic pricing (e.g., surge pricing during Dashain).
    • Nepal Challenge: Poor GPS accuracy in hilly areas → Pathao uses multiple satellite fixes.

## Exam Tip

1. Compare Models with Real Examples

  • Question: "Differentiate B2B and B2C e-commerce with Nepal examples."
  • Answer:
    Aspect B2B (NOC Suppliers) B2C (Daraz)
    Transaction Size Bulk orders (NPR 500,000+) Small orders (NPR 1,000-)
    Payment Method Bank transfers, EDI eSewa, Khalti, COD
    Security Digital signatures, API keys PCI-DSS, 2FA
    Infrastructure ERP systems (SAP) Mobile-responsive websites

2. Explain Infrastructure with Diagrams

  • Question: "Explain the technical infrastructure of eSewa."
  • Answer: Draw a 3-layer diagram:
    1. User Layer: Mobile app (Android/iOS) + USSD.
    2. Processing Layer: Khalti integration + Nepal Rastra Bank’s NPCI.
    3. Backend: AWS servers + PostgreSQL database + Fraud detection AI.

3. Analyze Case Studies for Trade-offs

  • Question: "How does Pathao’s dynamic pricing affect drivers?"
  • Answer:
    • Pros: Higher earnings during surge periods (e.g., airport rides).
    • Cons: Lower demand when prices spike (e.g., NPR 800/km vs. usual NPR 300).
    • Nepal-Specific: Drivers in rural areas earn less due to lower surge pricing.

4. Link Theory to Nepal’s Constraints

  • Question: "Why is C2C e-commerce more popular in Nepal than B2B?"
  • Answer:
    • Low trust in businesses → peer-to-peer transactions (e.g., Facebook Marketplace) feel safer.
    • Limited EDI adoption → SMEs can’t afford B2B platforms.
    • Cash preference → C2C allows COD, while B2B requires bank transfers.

5. Discuss Emerging Trends with Local Relevance

  • Question: "How can blockchain improve e-commerce in Nepal?"
  • Answer:
    • Problem: Fake products (e.g., counterfeit Ncell batteries).
    • Solution: Blockchain tracking (like Daraz’s "Authentic Seller").
    • Challenge: Low smartphone penetration → QR codes may not work for rural sellers.
    • Future: USSD-based blockchain verification (e.g., dial *123# to check product authenticity).

Final Visual Summary:

mindmap
  root((E-Commerce Applications in Nepal))
    B2B
      Example: NOC Suppliers
      Tech: EDI, ERP
    B2C
      Example: Daraz
      Tech: Mobile Apps, AI Pricing
    C2C
      Example: OLX Nepal
      Tech: Escrow, Ratings
    B2G
      Example: e-Governance Nepal
      Tech: Digital Signatures
    Social Commerce
      Example: Instagram Shopping
      Tech: API Integrations
    Software Agents
      Example: eSewa Chatbot
      Tech: NLP, Fraud Detection
    Emerging Trends
      Blockchain: Daraz Authentic Seller
      AI: Dynamic Pricing in Hotels

Based on the TU BBM syllabus for E Commerce (IT204), unit 9.

Discussion

Loading…