Building & Managing IS: Systems Development, Agile, SDLC, Feasibility, Maintenance

TAKEAWAYS:

• The Systems Development Life Cycle (SDLC) is a structured process (planning → analysis → design → implementation → maintenance) to build reliable IS, but Agile offers iterative flexibility for fast-changing projects.
• Feasibility studies (technical, operational, economic, schedule) determine if a project is viable before investment—critical for banks like Nabil Bank evaluating loan-processing systems.
• Change management (resistance, training, communication) ensures user adoption; Pathao’s ride-hailing app failed in Nepal partly due to poor driver training.
• Maintenance (corrective, adaptive, perfective, preventive) keeps systems running—NTC’s fiber-optic network upgrades require constant adaptive maintenance.
• Outsourcing vs. insourcing trade-offs (cost, expertise, control) matter: Daraz outsources IT infrastructure to AWS but insources customer-service bots.
• Ethical/social trade-offs (privacy, job displacement) arise in automation—eSewa’s AI chatbots replace human agents, raising unemployment concerns.

1. Systems Development Life Cycle (SDLC): The Waterfall Model

The SDLC is a sequential, phase-based approach to building information systems. Each phase must be completed before moving to the next, making it rigid but predictable. The classic waterfall model has six phases:

flowchart TD
    A["Planning"] --> B["Analysis"]
    B --> C["Design"]
    C --> D["Implementation"]
    D --> E["Testing"]
    E --> F["Maintenance"]

Phases Explained with a Real Example: Nabil Bank’s Loan Processing System

Assume Nabil Bank wants to automate its loan approval process. Here’s how SDLC applies:

Advantages and Disadvantages of SDLC

2. Agile Methodology: Iterative and Flexible Development

While SDLC is predictable, Agile prioritizes speed, adaptability, and customer feedback. Used by Pathao, Daraz, and Google, Agile breaks projects into small, iterative cycles (sprints) of 2–4 weeks.

mindmap
  root((Agile Methodology))
    Scrum["Scrum Framework"]
      Sprint["2-4 week cycles"]
      Daily Standups["15-min updates"]
      Product Backlog["Prioritized tasks"]
    Kanban["Visual workflow"]
      Columns["To Do → In Progress → Done"]
      WIP Limits["Max tasks per column"]
    Extreme Programming["XP"]
      Pair Programming["Two devs on one task"]
      Continuous Integration["Automated testing"]
    Advantages["Fast delivery, flexibility"]
    Disadvantages["Requires discipline, less documentation"]

Agile vs. SDLC: A Comparison

How Daraz Uses Agile

Daraz’s mobile app team uses Scrum:

• Sprints: Every 2 weeks, they release a new feature (e.g., "one-click checkout").
• Daily Standups: Developers discuss blockers (e.g., payment gateway delays).
• Backlog: Prioritizes features like "cash-on-delivery for rural areas" based on customer complaints.

3. Feasibility Studies: Is the Project Worth It?

Before investing, organizations assess four types of feasibility:

flowchart TD
    A["Feasibility Study"] --> B["Technical Feasibility"]
    A --> C["Operational Feasibility"]
    A --> D["Economic Feasibility"]
    A --> E["Schedule Feasibility"]

Types of Feasibility with Examples

Feasibility Study Worked Example: Khalti’s UPI Integration

Problem: Khalti wants to add UPI (Unified Payments Interface) to its app. Feasibility Check:

1. Technical: UPI requires NPCI certification (feasible, but takes 6 months).
2. Operational: Users love UPI (operational feasibility: high).
3. Economic: Cost = $200K; Expected transactions = 500K/month → $100K/month revenue (economic feasibility: high).
4. Schedule: NPCI certification takes 6 months; Khalti can launch in 9 months (schedule feasibility: medium).

Decision: Proceed, but outsource NPCI certification to a fintech consultant.

4. Change Management: Overcoming Resistance

Even the best systems fail if users resist change. John Kotter’s 8-Step Change Model helps manage this:

flowchart TD
    A["1. Create Urgency"] --> B["2. Build Coalition"]
    B --> C["3. Form Vision"]
    C --> D["4. Communicate Vision"]
    D --> E["5. Empower Action"]
    E --> F["6. Generate Short-Term Wins"]
    F --> G["7. Consolidate Gains"]
    G --> H["8. Anchor Change"]

Why Pathao’s Nepal Launch Failed (Change Management Lesson)

Pathao entered Nepal in 2017 but shut down in 2019. Key mistakes:

• Lack of driver training: Drivers didn’t know how to use the app → operational failure.
• Poor communication: No clear benefits for drivers (e.g., "earn 2x more than taxis").
• No pilot test: Launched in Kathmandu without testing in smaller cities.

Solution for Future Projects:

• Pilot first: Test in Pokhara before Kathmandu.
• Incentivize users: Offer drivers bonuses for 5-star ratings.
• Training: Partner with Nepal Truck Drivers Association for workshops.

5. Maintenance: Keeping Systems Alive

After launch, systems need ongoing care. There are four types of maintenance:

Maintenance Costs Over Time

graph LR
    A["Launch"] --> B["Year 1: High corrective maintenance"]
    B --> C["Year 2-3: Mix of adaptive & perfective"]
    C --> D["Year 4+: Mostly preventive"]

Key Insight: Maintenance costs decrease over time if planned well.

6. Outsourcing vs. Insourcing: Make or Buy?

Organizations decide whether to build in-house (insourcing) or hire external vendors (outsourcing).

When to Outsource?

• Lack of in-house skills (e.g., AI for eSewa’s chatbot).
• Cost savings (e.g., NTC outsourcing network maintenance to Huawei).
• Speed (e.g., Pathao using Google Maps API instead of building its own routing system).

7. Ethical and Social Issues in System Building

Building IS raises ethical dilemmas:

• Privacy: eSewa collects biometric data—is it secure?
• Job displacement: Nabil Bank’s automated loan system may replace tellers.
• Digital divide: Rural users may be excluded if systems require smartphones.

Case Study: Nabil Bank’s AI Loan Officer

Scenario: Nabil Bank replaces 50 loan officers with an AI chatbot. Ethical Concerns:

1. Transparency: Can customers trust an AI’s decision?
2. Bias: Does the AI favor urban applicants over rural ones?
3. Job loss: 50 employees lose jobs—how to retrain them?

Solutions:

• Explain AI decisions: Show customers the logic behind loan rejections.
• Audit for bias: Test AI with rural vs. urban data to ensure fairness.
• Reskill employees: Train laid-off officers for AI monitoring roles.

In the Real World

1. Nabil Bank’s Loan Processing System

   • Idea Used: SDLC + Feasibility Study
   • How: Nabil Bank used a 6-phase SDLC to build a digital loan system, conducting a feasibility study that showed a 60% ROI within 3 years. The system now processes 80% of loans online, reducing approval time from 5 days to 2 hours.

2. Daraz’s Agile App Development

   • Idea Used: Agile (Scrum) + Outsourcing
   • How: Daraz’s mobile team uses 2-week sprints to release features like "one-click checkout." They outsource backend hosting to AWS but insource UI/UX design to maintain brand consistency.

3. Pathao’s Failed Nepal Launch (Change Management)

   • Idea Used: Poor Change Management
   • How: Pathao ignored driver training and pilot testing, leading to app crashes and driver walkouts. Lesson: Always test in a small market first (e.g., Pokhara before Kathmandu).

Exam Tip

This unit is heavily tested in PU exams through:

1. Scenario-based questions (e.g., "NTC wants to upgrade its billing system. Advise them on SDLC phases and feasibility studies.").
   • Tip: Always structure answers using SDLC phases or feasibility types (technical/operational/economic).
2. Comparisons (e.g., "Compare Agile and SDLC with examples from Daraz and Nabil Bank.").
   • Tip: Use tables (like above) to score full marks.
3. Case studies (e.g., "How would you manage resistance to eSewa’s new biometric login?").
   • Tip: Apply Kotter’s 8-step model or ADKAR framework (Awareness → Desire → Knowledge → Ability → Reinforcement).
4. Short-answer definitions (e.g., "Define perfective maintenance with an example.").
   • Tip: Memorize all four maintenance types and link them to NEPSE or NTC.

Common Mistakes to Avoid:

• Ignoring feasibility: Always check all four types (technical, operational, economic, schedule).
• Mixing Agile and SDLC: Agile is iterative; SDLC is sequential. Don’t confuse them!
• Skipping real-world examples: Exams love Nabil Bank, Daraz, or NTC—use them.
• Overlooking ethics: Questions may ask "What are the ethical concerns of automating Khalti’s customer service?"—always mention privacy, bias, and job loss.

Final Visual Summary

mindmap
  root((Building & Managing IS))
    SDLC["Waterfall Model: Planning → Analysis → Design → Implementation → Testing → Maintenance"]
    Agile["Scrum/Kanban: Iterative, sprints, daily standups"]
    Feasibility["Technical, Operational, Economic, Schedule"]
    Change Management["Kotter’s 8 Steps: Urgency → Vision → Action"]
    Maintenance["Corrective, Adaptive, Perfective, Preventive"]
    Outsourcing["Pros/Cons: Cost vs. Control"]
    Ethics["Privacy, Bias, Job Displacement"]