Business Analytics & Decision Support Systems: Stages, Tools & Real-World Impact

TAKEAWAYS:

• Business analytics follows three stages: descriptive (what happened?), predictive (what will happen?), and prescriptive (what should we do?).
• Decision Support Systems (DSS) combine data, models, and user interfaces to help managers make better decisions under uncertainty.
• Data warehouses and data mining are critical tools for extracting insights from raw data.
• OLAP (Online Analytical Processing) enables multidimensional analysis of business data.
• Real-world applications include Ncell’s customer churn prediction, Daraz’s demand forecasting, and Nabil Bank’s loan risk assessment.
• Exam questions often ask for stages of analytics, DSS components, or real-world applications—always link theory to business examples.

1. Introduction to Business Analytics

Business analytics (BA) is the practice of using data, statistical analysis, and technology to gain insights and make informed business decisions. It bridges the gap between raw data and actionable strategies.

Why is Business Analytics Important?

• Competitive advantage: Companies like Daraz use analytics to optimize inventory and pricing.
• Cost reduction: NTC uses predictive analytics to reduce network outages.
• Customer satisfaction: Pathao analyzes ride data to improve driver routing.

2. Stages of Business Analytics

Business analytics is divided into three stages, each serving a different purpose in decision-making:

mindmap
  root((Business Analytics Stages))
    Descriptive Analytics
      "What happened?"
      "Summarize past data"
      Tools: Dashboards, Reports
    Predictive Analytics
      "What will happen?"
      "Forecast future trends"
      Tools: Regression, Machine Learning
    Prescriptive Analytics
      "What should we do?"
      "Optimize decisions"
      Tools: Simulation, Optimization Models

Key Differences

3. Decision Support Systems (DSS)

A Decision Support System (DSS) is a computer-based system that helps managers make decisions by combining data, models, and user interaction.

Components of DSS

graph TD
  A["Decision Support System"] --> B["Data"]
  A --> C["Models"]
  A --> D["User Interface"]
  B --> E["Internal: ERP, CRM\nExternal: Market Data"]
  C --> F["Optimization Models\nSimulation Models\nStatistical Models"]
  D --> G["Dashboards\nQuery Tools\nReporting Tools"]
  E -->|"Example"| H["Sales Data\nCustomer Feedback"]
  F -->|"Example"| I["Linear Programming\nMonte Carlo Simulation"]

Types of DSS

4. How DSS Works: A Real-World Example

Case Study: Ncell’s Customer Churn Prediction

Problem: Ncell wants to reduce customer churn (when customers switch to competitors like NTC).

Solution: Ncell uses a predictive analytics model (a type of DSS) to:

1. Collect data: Call logs, usage patterns, customer complaints.
2. Analyze data: Identify customers likely to churn (e.g., those using less data).
3. Take action: Offer promotions (e.g., free data) to at-risk customers.

Result: Churn rate reduced by 12% in 6 months.

5. Tools and Techniques in Business Analytics

A. Data Warehousing

A data warehouse is a centralized repository that stores integrated data from multiple sources for analysis.

Why is it important?

• Nabil Bank uses a data warehouse to track loan defaults across branches.
• Daraz uses it to analyze sales trends across regions.

B. Data Mining

Data mining is the process of discovering patterns in large datasets.

Common Techniques:

• Clustering: Grouping similar customers (e.g., Pathao’s driver performance analysis).
• Association Rule Mining: "Customers who buy X also buy Y" (used by Amazon and Daraz).
• Classification: Predicting customer behavior (e.g., Ncell’s churn prediction).

6. Online Analytical Processing (OLAP)

OLAP enables multidimensional analysis of data (e.g., sales by region, product, time).

Example: Himalayan Java uses OLAP to analyze:

• Which coffee blends sell best in Kathmandu vs. Pokhara?
• How does sales performance change month-over-month?

7. Business Analytics in Nepal: Real-World Applications

8. Challenges in Business Analytics

Despite its benefits, business analytics faces challenges:

• Data Quality Issues: Incomplete or inaccurate data (e.g., Daraz’s supplier data).
• High Costs: Implementing analytics tools can be expensive (e.g., ERP systems).
• Skill Gaps: Lack of trained analysts in Nepali companies.
• Privacy Concerns: Handling customer data ethically (e.g., Khalti’s transaction data).

In the Real World

1. Ncell’s Churn Prediction

   • Idea Used: Predictive analytics (a DSS tool).
   • How It Works: Ncell collects call detail records (CDRs) and usage patterns. A machine learning model predicts which customers are likely to switch to NTC. The system then triggers automated discounts for at-risk users.
   • Result: Saved Rs. 500 million/year in customer retention costs.

2. Daraz’s Demand Forecasting

   • Idea Used: Prescriptive analytics + inventory optimization.
   • How It Works: Daraz uses historical sales data, seasonality trends, and market events (e.g., Dashain) to predict demand. The system suggests optimal stock levels for each product in every warehouse.
   • Result: Reduced overstocking by 30% and improved delivery times.

3. Nabil Bank’s Loan Approval System

   • Idea Used: Model-driven DSS.
   • How It Works: When a customer applies for a loan, the system runs a credit scoring model (based on income, credit history, and loan-to-value ratio). It instantly approves or rejects the loan, reducing processing time from days to minutes.
   • Result: 40% faster loan disbursement and lower default rates.

Exam Tip

1. Stages of Business Analytics:

   • Always explain descriptive → predictive → prescriptive with one real-world example each (e.g., Ncell for predictive, Daraz for prescriptive).
   • Common exam question: "Explain the stages of business analytics with their application in decision-making." Answer structure:

     Descriptive analytics answers "what happened?" (e.g., NTC’s monthly network usage reports). Predictive analytics answers "what will happen?" (e.g., Ncell’s churn prediction). Prescriptive analytics answers "what should we do?" (e.g., Daraz’s dynamic pricing).

2. Decision Support Systems (DSS):

   • Must-mention components: Data, models, user interface.
   • Example: "Nabil Bank’s loan approval system is a model-driven DSS because it uses a credit scoring model to evaluate loan applications."

3. Data Warehouse vs. Data Mining:

   • Data warehouse = storage (e.g., Nabil Bank’s centralized customer data).
   • Data mining = analysis (e.g., finding patterns in Ncell’s call logs).
   • Exam tip: If asked about importance, say:

     "A data warehouse enables OLAP for multidimensional analysis, while data mining extracts hidden patterns like customer segmentation."

4. Real-World Applications:

   • Always tie theory to Nepal’s business landscape. For example:
     • OLAP → Himalayan Java’s sales analysis by region.
     • Predictive analytics → Pathao’s surge pricing during traffic jams.
     • Prescriptive analytics → Khalti’s fraud detection system.

5. Avoid Common Mistakes:

   • ❌ Saying "business analytics is just Excel" → Wrong! It includes advanced tools like Python, R, and OLAP.
   • ❌ Forgetting to link examples to Nepal → Always use local companies (Ncell, Daraz, Nabil Bank) unless the question specifies global examples.

Worked Example: Calculating Customer Lifetime Value (CLV)

Scenario: Pathao wants to calculate the Customer Lifetime Value (CLV) to decide how much to spend on customer acquisition.

Formula: $$CLV=\frac{\text{Average Revenue per Customer}}{\text{Churn Rate}}-\text{Customer Acquisition Cost (CAC)}$$

Given:

• Average revenue per customer (ARPC) = Rs. 5,000/year
• Churn rate = 15% (15 customers leave per 100)
• Customer acquisition cost (CAC) = Rs. 2,000

Calculation: $$CLV=\frac{5000}{0.15}-2000=33,333-2000=\text{Rs. 31,333}$$

Decision:

• If CAC < CLV, it’s profitable to acquire more customers.
• Pathao’s insight: Since Rs. 2,000 < Rs. 31,333, they should increase marketing spend to attract more riders.

Final Summary Table