MODULE 3

FUNCTIONAL ANALYTICS
Comprehensive Study Notes
9 Hours | BTL Levels 3 & 4

Topics Covered:
HR Analytics | Supply Chain Analytics | Marketing Analytics | Customer Analytics
Business Process Analytics | Financial Analytics | Big Data Analytics | Web & Social Media
Analytics
1. HR Analytics
1.1 Definition and Overview
HR Analytics (also called People Analytics or Workforce Analytics) is the application of data
analysis techniques to human resource data to improve employee performance, engagement,
retention, and overall workforce productivity. It transforms traditional HR functions—recruiting,
training, performance management—into data-driven, evidence-based decisions. Companies like
Google, IBM, and Infosys have pioneered HR analytics to gain a competitive edge in talent
management.

1.2 Key Concepts (Sub-Points)

• Talent Acquisition Analytics: Uses predictive models to identify ideal candidate profiles,
optimize job descriptions, reduce time-to-fill, and improve quality of hire. Machine learning
ranks resumes and predicts candidate success rates.
• Employee Retention Analytics: Identifies at-risk employees likely to leave by analysing
factors like engagement scores, compensation gaps, manager relationships, and career
growth. Enables proactive interventions.
• Performance Management Analytics: Links individual KPIs to business outcomes,
identifies high-potential employees, and reveals patterns between training investments and
productivity improvements.
• Workforce Planning Analytics: Forecasts future workforce needs based on business
growth plans, retirement trends, and skill gaps. Ensures the right people are available at the
right time.
• Learning & Development (L&D) Analytics: Measures ROI of training programs by
correlating training completion with performance improvements, promotion rates, and
employee satisfaction.
• Diversity, Equity & Inclusion (DEI) Analytics: Tracks diversity metrics across hiring,
promotions, and pay equity. Identifies systemic biases in HR processes.
• Employee Engagement Analytics: Analyses pulse surveys, sentiment data, absenteeism,
and productivity metrics to gauge morale and predict disengagement.
• Compensation Analytics: Benchmarks salaries against market data, identifies pay inequities,
and models the impact of compensation changes on retention and performance.
• HR Process Efficiency Analytics: Measures time, cost, and quality of HR processes like
onboarding, payroll processing, and benefits administration to identify optimization
opportunities.
1.3 HR Analytics Framework / Flowchart
┌─────────────────────────────────────────────────────────────┐
│ HR ANALYTICS PROCESS FLOW │
└─────────────────────────────────────────────────────────────┘
│
▼
┌─────────────────┐
│ Data Collection │ ← HRIS, ATS, LMS, Surveys, Payroll
└────────┬────────┘
│
▼
┌─────────────────┐
│ Data Cleaning │ ← Remove duplicates, handle missing values
└────────┬────────┘
│
▼
┌─────────────────────┐
│ Descriptive │ ← Headcount, turnover rate, absenteeism
│ Analytics │
└────────┬────────────┘
│
▼
┌─────────────────────┐
│ Diagnostic │ ← Why is turnover high in Dept X?
│ Analytics │
└────────┬────────────┘
│
▼
┌─────────────────────┐
│ Predictive │ ← Who will leave in next 6 months?
│ Analytics │
└────────┬────────────┘
│
▼
┌─────────────────────┐
│ Prescriptive │ ← What actions reduce attrition?
│ Analytics │
└────────┬────────────┘
│
▼
┌─────────────────────┐
│ HR Decision & │ ← Retention programs, promotions
│ Intervention │
└─────────────────────┘
1.4 Key Performance Indicators (KPIs)
KPI Formula / Measure Sector Target Benchmark
Employee Turnover Rate (Employees Left / Avg IT / BPO < 15% annually
Employees) × 100
Time to Fill Days from job posting to All Sectors < 30–45 days
offer acceptance
Cost per Hire Total Hiring Cost / Number Manufacturing < ₹50,000
of Hires
Training ROI (Training Benefit – Cost) / BFSI > 150%
Cost × 100
Employee Engagement Pulse survey average (scale Healthcare > 7.5/10
Score 1–10)
Absenteeism Rate (Days Absent / Total Retail < 3%
Workdays) × 100
Revenue per Employee Total Revenue / Total FTEs Tech Maximize
Offer Acceptance Rate (Accepted Offers / Total Consulting > 85%
Offers) × 100

1.5 Case Study: IBM's Employee Attrition Prediction

Background:
IBM faced significant attrition in its technology division, losing skilled engineers at a rate of 25%+
annually. Each departing employee cost an estimated $40,000–$60,000 in replacement costs.
Analytics Approach:
• IBM built a predictive attrition model using 50+ variables including: job satisfaction scores,
years since last promotion, distance from home, work-life balance ratings, and overtime
hours.
• Used machine learning (Random Forest + Logistic Regression) on 1,470 employee records
(publicly available IBM HR dataset).
• Identified top attrition predictors: OverTime, JobLevel, MonthlyIncome, YearsAtCompany,
WorkLifeBalance.
Key Findings:
• Employees working overtime were 3.2x more likely to leave within 6 months.
• Employees with < ₹3,000 monthly income had 40% higher attrition probability.
• Job Level 1 employees had the highest attrition (26.4% vs. 10.1% average).
Actions Taken:
• Introduced 'Stay Conversations' for high-risk employees flagged by the model.
 • Restructured overtime compensation and introduced flexible work policies.
• Created accelerated career pathways for Level 1 employees.
Results:
• Attrition reduced by 18% within 12 months, saving IBM approximately $300M annually.
• Model accuracy: 87% in predicting employees who would leave within 6 months.

1.6 Practical Relevance

Organizations like TCS, Wipro, and Amazon use HR analytics dashboards (Workday, SAP
SuccessFactors) to make real-time workforce decisions. During COVID-19, HR analytics helped
companies model the impact of remote work on productivity and identify employees facing burnout.

1.7 Chapter-End Cases

BTL Level 3 – Application Case:
Zeta Technologies, an IT firm in Bengaluru, has noticed that 30% of its software engineers leave
within 18 months of joining. The HR team has collected data on exit interview reasons, performance
scores, manager ratings, and compensation. As the HR Analytics Manager, design a data collection
strategy, identify the top 5 variables you would analyze, choose an appropriate predictive model,
and describe two interventions you would implement based on model findings.

BTL Level 4 – Analysis Case:

After implementing an HR analytics solution, Zeta Technologies' model shows: (a) engineers
mentored by senior staff have 60% lower attrition; (b) employees who didn't receive a promotion in
3 years have 80% attrition probability; (c) team size above 12 correlates with lower engagement.
Critically evaluate which of these three findings is most actionable, estimate the ROI of addressing
each finding, and prioritize an action plan. Justify your decisions with analytical reasoning and
identify what additional data you would collect.
2. Supply Chain Analytics
2.1 Definition and Overview
Supply Chain Analytics involves the use of quantitative methods, statistical tools, and data mining
techniques to analyse supply chain data—demand, inventory, logistics, supplier performance, and
distribution—to optimize operations, reduce costs, and improve service levels. It spans the entire
value chain from raw material procurement to last-mile delivery.

2.2 Key Concepts (Sub-Points)

• Demand Forecasting Analytics: Uses historical sales data, market trends, and external
signals (weather, events, economic indicators) to predict future demand. Methods include
ARIMA, exponential smoothing, and machine learning models like XGBoost.
• Inventory Optimization Analytics: Balances holding costs against stockout risks by
analyzing reorder points, safety stock levels, and Economic Order Quantity (EOQ). Prevents
both over-stocking and under-stocking.
• Supplier Performance Analytics: Scores suppliers on delivery reliability, quality defect
rates, cost competitiveness, and responsiveness. Enables supplier segmentation (strategic
vs. tactical vs. commodity).
• Logistics & Transportation Analytics: Optimizes routes, carrier selection, load
consolidation, and delivery schedules using GPS data, traffic patterns, and fuel consumption
metrics.
• Warehouse Analytics: Analyses order picking efficiency, storage utilization, throughput
rates, and labour productivity to improve warehouse layout and operations.
• Supply Chain Risk Analytics: Identifies and quantifies risks—single-source dependencies,
geopolitical risks, natural disasters—and models the financial impact of supply disruptions.
• Procurement Analytics: Analyses spending patterns, contract compliance, maverick
spending, and category management to reduce procurement costs.
• Reverse Logistics Analytics: Tracks returns, refurbishments, recycling rates, and warranty
claims to minimize costs and improve sustainability.
• Network Design Analytics: Uses simulation and optimization models to determine the
optimal number and location of warehouses, distribution centres, and manufacturing plants.
2.3 Supply Chain Analytics Flowchart
┌───────────────────────────────────────────────────────────┐
│ SUPPLY CHAIN ANALYTICS FLOW │
└───────────────────────────────────────────────────────────┘

[Suppliers] → [Raw Material Data] → [Procurement Analytics]

│
▼
[Manufacturing / Production Analytics]
← Yield, OEE, Downtime Analysis
│
▼
[Inventory Analytics] ← Demand Forecast
← EOQ, Safety Stock, ABC Analysis
│
▼
[Logistics Analytics] ← Route Optimization
← Carrier Performance, Freight Cost
│
▼
[Customer Delivery Analytics]
← OTIF, Last-mile efficiency
│
▼
[Supply Chain Dashboard & Decision Support]
← Real-time KPI Monitoring

2.4 Key Performance Indicators (KPIs)

KPI Formula Sector Benchmark
On-Time In-Full (OTIF) (On-time & complete deliveries FMCG / Retail > 95%
/ Total orders) × 100
Inventory Turnover COGS / Average Inventory Manufacturing > 8x/year
Perfect Order Rate Orders without errors, E-Commerce > 98%
damage, delays / Total orders
× 100
Supplier OTD Rate (On-time deliveries / Total Auto / Electronics > 97%
deliveries) × 100
Freight Cost per Unit Total Freight Cost / Units Logistics Minimize
Shipped
Days Inventory (Inventory / COGS) × 365 Pharma < 45 days
Outstanding (DIO)
Demand Forecast 1 – |Actual – Forecast| / Actual All Sectors > 85%
Accuracy × 100

2.5 Case Study: Amazon's Supply Chain Analytics

Background:
Amazon handles over 1.6 million packages per day. Managing inventory across 175+ fulfilment
centres globally requires sophisticated analytics. The challenge was to minimize delivery time and
cost while maintaining 99%+ availability.
Analytics Applied:
• Anticipatory Shipping (Predictive Analytics): Amazon's algorithm predicts what customers
will order before they place the order, pre-positions inventory in the nearest fulfillment
centre. Uses browsing history, purchase patterns, wish lists, and demographic data.
• Inventory Positioning (Network Optimization): Uses linear programming to decide which
warehouse should stock which product, balancing storage costs against shipping speed.
• Dynamic Pricing Analytics: Adjusts prices 2.5 million times per day based on demand
signals, competitor pricing, and inventory levels.
Results:
• Average delivery time reduced from 5 days to 1–2 days for Prime members.
• Inventory holding costs reduced by 35% through predictive positioning.
• Demand forecast accuracy exceeds 95% for top 10,000 SKUs.

2.6 Practical Relevance

Indian companies like Flipkart, Reliance Retail, and Mahindra Logistics use supply chain analytics
to compete in the e-commerce era. During the COVID-19 pandemic, supply chain analytics helped
pharmaceutical companies like Cipla reroute API sourcing when Chinese suppliers were disrupted.

2.7 Chapter-End Cases

BTL Level 3 – Application Case:
Bharat Grocers, a large retail chain with 200 stores across India, faces frequent stockouts of 15 key
products and simultaneously over-stocking 30 slow-moving products. You have been given weekly
sales data for 52 weeks, supplier lead time data, and warehouse capacity data. Apply ABC-XYZ
analysis, calculate EOQ for the top 5 products, and design a demand forecasting approach. Specify
what technology platform you would recommend.

BTL Level 4 – Analysis Case:

Bharat Grocers' analytics reveals: stockouts cost ₹2.5 Cr/month in lost sales; overstock incurs ₹1.2
Cr/month in holding costs; the current demand forecast accuracy is 68%. If a new ML-based
forecasting system costs ₹80 lakhs to implement and improves accuracy to 88%, analyse the
break-even timeline, identify which product categories offer the highest ROI from better forecasting,
and assess supply chain risks that analytics alone cannot mitigate.
3. Marketing Analytics
3.1 Definition and Overview
Marketing Analytics is the practice of measuring, managing, and analysing marketing performance
data to maximise effectiveness and optimise return on investment (ROI). It connects marketing
activities to business outcomes—revenue, market share, customer acquisition—by using statistical
models, attribution techniques, and data visualization to understand what works and why.

3.2 Key Concepts (Sub-Points)

• Marketing Mix Modelling (MMM): A statistical technique that quantifies the impact of each
marketing channel (TV, digital, print, promotions) on sales. Separates base sales (brand
equity) from incremental sales (marketing-driven). Widely used in FMCG and automotive
sectors.
• Digital Marketing Analytics: Tracks performance of online channels—SEO, SEM, email,
social media, display ads—using metrics like CTR, CPC, conversion rate, and bounce rate.
Tools: Google Analytics 4, Adobe Analytics.
• Customer Segmentation Analytics: Divides customers into homogeneous groups using
clustering algorithms (K-means, RFM analysis) to enable targeted marketing. Each segment
receives tailored messaging and offers.
• Attribution Modelling: Assigns credit to various touchpoints in the customer journey for
driving a conversion. Models include first-touch, last-touch, linear, time-decay, and data-
driven attribution.
• A/B Testing and Experimentation: Scientifically tests marketing hypotheses by randomly
splitting audiences and measuring differential outcomes. Determines optimal ad creatives,
landing pages, email subject lines, and pricing.
• Campaign Performance Analytics: Measures the effectiveness of individual campaigns
against predefined KPIs. Compares planned vs. actual outcomes across reach,
engagement, conversions, and revenue.
• Brand Analytics: Monitors brand health metrics—awareness, consideration, preference, Net
Promoter Score (NPS)—using survey data, social listening, and search trend analysis.
• Price Elasticity Analytics: Estimates how demand changes with price changes. Guides
pricing decisions and promotional strategy. Uses regression models on historical sales and
price data.
• Content Analytics: Measures the performance of marketing content—blog posts, videos,
infographics—by engagement depth, lead generation, and SEO ranking improvement.
3.3 Attribution Modelling Flowchart
┌────────────────────────────────────────────────────────────┐
│ CUSTOMER JOURNEY & ATTRIBUTION FLOW │
└────────────────────────────────────────────────────────────┘

Customer Journey:
[Social Ad] → [Google Search] → [Email] → [Website Visit] → [PURCHASE]
#1 #2 #3 #4

Attribution Models:

First-Touch: [100%] → [0%] → [0%] → [0%]

Last-Touch: [0%] → [0%] → [0%] → [100%]
Linear: [25%] → [25%] → [25%] → [25%]
Time-Decay: [10%] → [20%] → [30%] → [40%]
Data-Driven: [18%] → [22%] → [35%] → [25%] (ML-based)

↓ Data-Driven Attribution is Most Accurate ↓

→ Reallocate budget from low-ROI to high-ROI channels

3.4 Key Performance Indicators (KPIs)

KPI Formula Sector Target
Return on Ad Spend Revenue from Ads / Ad E-Commerce > 4:1
(ROAS) Spend
Customer Acquisition Cost Total Marketing Cost / New SaaS / D2C < LTV/3
(CAC) Customers
Marketing ROI (Revenue – Marketing All > 200%
Cost) / Marketing Cost × 100
Click-Through Rate (CTR) (Clicks / Impressions) × 100 Digital > 2%
Conversion Rate (Conversions / Visitors) × Retail > 3%
100
Email Open Rate (Opened / Sent – Bounced) B2B / B2C > 20–25%
× 100
Brand NPS % Promoters – % Detractors Consumer > 50
Goods
Cost per Lead (CPL) Total Spend / Number of Real Estate Minimize
Leads

3.5 Case Study: Hindustan Unilever (HUL) – Marketing Mix Modelling

Background:
HUL, India's largest FMCG company, spends over ₹3,500 Cr annually on marketing across 35+
brands. The challenge: How to allocate this budget optimally across TV, digital, outdoor,
promotions, and activations?
Analytics Approach:
 • Deployed Marketing Mix Modelling (MMM) using 5 years of weekly sales data correlated
with media spends, pricing, promotions, and competitor activity.
• Used Bayesian regression models to separate base sales from marketing-driven sales for
each brand.
• Modeled interactions between channels (TV+Digital synergy effect).
Key Findings:
• For Surf Excel: Digital advertising showed 2.8x higher ROI than traditional TV for the 18–35
age segment.
• For Lifebuoy: Rural activation (below-the-line) generated 40% of incremental sales with only
15% of budget.
• Trade promotions had diminishing returns beyond 3 weeks—extending to 4+ weeks actually
reduced brand equity.
Actions and Results:
• Reallocated ₹400 Cr from TV to digital platforms for urban brands → 22% improvement in
overall marketing ROI.
• Shortened promotional windows from 4 weeks to 2 weeks, saving ₹150 Cr annually.
• Overall portfolio revenue grew 11% with same marketing budget.

3.6 Chapter-End Cases

BTL Level 3 – Application Case:
Sunrise Sports, a D2C athletic wear brand, spends ₹50 lakhs/month on digital marketing across
Instagram (₹15L), Google Ads (₹20L), Email (₹5L), and YouTube (₹10L). Current monthly revenue
is ₹1.2 Cr. Use the ROAS formula for each channel, apply first-touch and last-touch attribution, and
recommend how to reallocate budget to achieve ₹1.8 Cr revenue without increasing total spend.

BTL Level 4 – Analysis Case:

Sunrise Sports' analytics shows Google Ads has ROAS of 6:1, Instagram 3.2:1, YouTube 2.1:1,
Email 9:1. However, a customer survey reveals that 70% of buyers first discovered the brand
through Instagram or YouTube. Critically analyse the conflict between attribution model data and
survey findings. What does this reveal about attribution model limitations? How would you design a
more holistic measurement framework? Evaluate the strategic risk of optimizing purely for ROAS.
4. Customer Analytics
4.1 Definition and Overview
Customer Analytics involves collecting, processing, and analysing customer data to understand
behaviour, preferences, needs, and lifetime value. It enables organizations to create personalized
experiences, predict churn, optimize the customer journey, and maximize customer profitability. It
draws from CRM data, transaction history, behavioural data, and feedback to build a 360° view of
each customer.

4.2 Key Concepts (Sub-Points)

• Customer Lifetime Value (CLV/LTV): Predicts the total net profit a business can expect from
a customer relationship. Uses purchase frequency, average order value, retention rate, and
profit margin. Guides acquisition budget allocation and retention investment decisions.
• RFM Analysis (Recency, Frequency, Monetary): Segments customers based on when they
last purchased (R), how often they purchase (F), and how much they spend (M). Creates
segments like 'Champions', 'At-Risk', 'Lost Customers' for targeted engagement.
• Customer Churn Prediction: Uses machine learning to identify customers likely to stop doing
business. Variables include usage decline, support tickets, reduced purchase frequency,
and competitor activity. Enables proactive retention campaigns.
• Customer Segmentation & Persona Building: Groups customers by demographics,
psychographics, behavioural patterns, and value tiers. Enables personalized product
development, pricing, and communication strategies.
• Customer Journey Analytics: Maps every touchpoint from awareness to purchase to loyalty,
identifying friction points and moments of delight. Uses funnel analysis, path analysis, and
session replay tools.
• Sentiment Analysis & Voice of Customer (VoC): Applies NLP to customer reviews,
support chats, social media mentions, and survey responses to quantify satisfaction drivers
and pain points.
• Next Best Action / Recommendation Analytics: Predicts the most relevant product, offer, or
content for each customer at each point in their journey using collaborative filtering and
content-based filtering algorithms.
• Net Promoter Score (NPS) Analytics: Analyses survey responses to segment promoters,
passives, and detractors. Links NPS scores to revenue outcomes and identifies root causes
of dissatisfaction.
• Cross-sell and Upsell Analytics: Identifies patterns of complementary purchases (market
basket analysis) and readiness for premium upgrades using association rule mining and
propensity models.
4.3 Customer Segmentation Flowchart (RFM)
┌──────────────────────────────────────────────────────────────┐
│ RFM CUSTOMER SEGMENTATION FLOW │
└──────────────────────────────────────────────────────────────┘

Customer Transaction Data

│
▼
Score R, F, M on scale 1–5 for each customer
R=5: purchased yesterday R=1: purchased > 1 year ago
F=5: 20+ purchases/year F=1: 1 purchase ever
M=5: top 20% spenders M=1: bottom 20% spenders
│
▼
┌─────────────────────────────────────────────────────┐
│ RFM Score │ Segment │ Strategy │
├─────────────┼──────────────────┼────────────────────┤
│ 5-5-5 │ Champions │ Reward, Advocate │
│ 4-4-4 │ Loyal │ Upsell / VIP │
│ 3-3-X │ Potential │ Nurture │
│ 5-1-1 │ New Customer │ Onboard well │
│ 1-4-4 │ At Risk │ Win-back offer │
│ 1-1-1 │ Lost │ Re-engage/cut │
└─────────────────────────────────────────────────────┘

4.4 Key Performance Indicators (KPIs)

KPI Formula Sector Target
Customer Lifetime Value (AOV × Purchase Freq × E-Commerce / Maximize
(CLV) Gross Margin) / Churn Rate BFSI
Churn Rate (Customers Lost / Customers Telecom / SaaS < 5% monthly
at Start) × 100
Customer Satisfaction Score (Satisfied respondents / Total) Retail / Banking > 80%
(CSAT) × 100
Net Promoter Score (NPS) % Promoters – % Detractors All B2C > 50
Retention Rate ((End – New) / Start) × 100 Subscription > 90%
Average Revenue per User Total Revenue / Active Users Telecom / OTT Maximize
(ARPU)
First Contact Resolution (Issues resolved in 1st Customer > 70%
(FCR) contact / Total) × 100 Service

4.5 Case Study: Airtel's Churn Prediction Model

Background:
Bharti Airtel, India's second-largest telecom operator with 350M+ subscribers, faced monthly churn
rates of 2.5–3%, translating to millions of customers and billions in lost revenue annually. Winning
back a lost customer costs 5–7x more than retaining one.
Analytics Approach:
 • Built a churn propensity model using 90 days of customer behaviour: data usage trends, call
drop frequency, recharge patterns, customer service interactions, and competitor SIM
activity (detected via dual-SIM usage patterns).
• Used Gradient Boosting (XGBoost) classifier trained on 50M+ customer records. Features
included: % change in data usage (last 30 vs. 60 days), number of complaints in last 60
days, recharge amount decline, and tenure.
• Scored all 350M customers weekly and flagged top 5% at-risk customers for intervention.
Interventions:
• High-value at-risk customers received proactive outbound calls from retention specialists
with personalized offers (bonus data, discounted plans, device upgrades).
• Medium-risk customers received targeted SMS/app notifications with loyalty benefits.
Results:
• Churn rate reduced from 2.8% to 1.9% monthly—a 32% reduction.
• Customer retention campaign ROI exceeded 400%.
• NPS improved by 12 points as proactive engagement enhanced satisfaction.

4.6 Chapter-End Cases

BTL Level 3 – Application Case:
Nova Bank has 2 lakh credit card customers. You have data on monthly spend, payment history
(on-time vs. late), product holding (savings account, loans, insurance), and customer service
contacts. Perform RFM segmentation for the top 5 customer tiers, identify which tier represents the
highest CLV, and design a next-best-action strategy for 'At-Risk' customers showing spend decline
over 3 consecutive months.

BTL Level 4 – Analysis Case:

Nova Bank's churn model predicts that 12,000 customers (6%) will close their accounts in the next
90 days. The top predictors are: (1) no new product added in 18 months, (2) 3+ complaints in last
quarter, (3) salary credit stopped for 2 months. The cost of retention intervention is ₹500/customer;
the average CLV of these customers is ₹45,000. Analyse the financial business case for the
retention campaign, determine the minimum retention rate the campaign must achieve to be
profitable, identify ethical considerations in predictive customer analytics, and critique limitations of
the model.
5. Business Process Analytics
5.1 Definition and Overview
Business Process Analytics (BPA) is the application of data analysis techniques to understand,
monitor, and improve an organization's business processes. It bridges operational data with
strategic decision-making by identifying inefficiencies, bottlenecks, compliance gaps, and
automation opportunities. BPA uses process mining, workflow analytics, and performance
dashboards to make processes transparent and optimizable.

5.2 Key Concepts (Sub-Points)

• Process Mining: Extracts process models from event logs in IT systems (ERP, CRM, BPM).
Automatically discovers actual process flows versus designed flows, revealing deviations,
bottlenecks, and compliance violations. Tools: Celonis, UiPath Process Mining.
• Process Performance Monitoring (PPM): Tracks KPIs of business processes in real-time—
cycle time, throughput, error rate, cost per transaction—against SLAs and targets using
dashboards and alerts.
• Root Cause Analysis (RCA) Analytics: Uses statistical techniques (Pareto analysis, fishbone
diagrams, regression) to identify the underlying causes of process failures, defects, or
delays rather than treating symptoms.
• Lean and Six Sigma Analytics: Applies DMAIC (Define, Measure, Analyse, Improve, Control)
methodology with statistical tools—process capability, control charts, hypothesis testing—to
eliminate waste and reduce process variation.
• Workflow Automation Analytics: Measures the efficiency gains from RPA (Robotic Process
Automation) implementations by comparing pre- and post-automation metrics: time, cost,
error rate, and employee utilization.
• Compliance and Audit Analytics: Continuously monitors transactions and process
executions for compliance with internal policies and external regulations. Flags anomalies,
segregation of duty violations, and audit exceptions automatically.
• Service Level Agreement (SLA) Analytics: Tracks service performance against contractual
SLA commitments in real-time. Identifies patterns of SLA breaches and their root causes
across customer segments and process types.
• Process Simulation and What-If Analysis: Models the impact of proposed process changes
before implementation using simulation tools. Allows stress-testing of processes under
different demand scenarios.
• Cost-to-Serve Analytics: Calculates the end-to-end cost of delivering a product/service to a
customer, including all process steps. Identifies high-cost processes that can be automated
or eliminated.

5.3 Business Process Analytics – DMAIC Flow

┌───────────────────────────────────────────────────────────────┐
│ BUSINESS PROCESS ANALYTICS – DMAIC FLOW │
└───────────────────────────────────────────────────────────────┘

┌──────────┐ ┌──────────┐ ┌──────────┐

 │ DEFINE │ → │ MEASURE │ → │ ANALYSE │
│ │ │ │ │ │
│ Define │ │ Collect │ │ Find root│
│ problem │ │ baseline │ │ causes │
│ & scope │ │ process │ │ (Pareto, │
│ │ │ data │ │ fishbone)│
└──────────┘ └──────────┘ └──────────┘
│
▼
┌──────────┐ ┌──────────┐
│ CONTROL │ ← │ IMPROVE │
│ │ │ │
│ Monitor │ │ Implement│
│ & sustain│ │ solutions│
│ gains │ │ & test │
└──────────┘ └──────────┘

Process Mining Layer (beneath all phases):

Event Logs → Actual Process Discovery → Conformance Checking
→ Enhancement Recommendations

5.4 Key Performance Indicators (KPIs)

KPI Description Sector Target
Process Cycle Time Average time to complete a BPO / Banking Minimize
process end-to-end
First Pass Yield (FPY) (Units processed without Manufacturing > 95%
rework / Total units) × 100
SLA Compliance Rate (Processes completed within IT Services > 98%
SLA / Total) × 100
Process Cost per Unit Total process cost / Number Finance / HR Minimize
of transactions
Error / Defect Rate (Defective outputs / Total Healthcare / < 1%
outputs) × 100 BFSI
Automation Rate (Automated steps / Total BPO / Insurance Maximize
steps) × 100
Process Compliance Rate (Compliant executions / Total Regulated > 99%
executions) × 100 Industries

5.5 Case Study: Siemens – Process Mining with Celonis

Background:
Siemens, the German industrial conglomerate, processed millions of purchase orders annually
across global operations. The Order-to-Cash (O2C) and Procure-to-Pay (P2P) processes had
significant inefficiencies—delayed payments, incorrect invoices, and manual rework—costing
hundreds of millions of euros annually.
Analytics Approach:
 • Deployed Celonis Process Mining to extract and analyse event log data from SAP ERP
covering 10 million+ process events.
• Discovered actual P2P process had 47 variants versus the 3 standard variants defined in
the process manual—indicating massive process complexity and non-compliance.
• Identified top 5 process deviations: late purchase orders, invoice price mismatches, missing
goods receipts, duplicate invoice entries, and unauthorized vendor additions.
Results:
• Reduced P2P cycle time by 30%, saving €50M in working capital by accelerating invoice
processing.
• Identified and corrected 12,000+ duplicate invoices worth €3.2M.
• RPA deployed on 15 repetitive process steps, reducing manual effort by 40%.
• Process compliance improved from 68% to 94% within 18 months.

5.6 Chapter-End Cases

BTL Level 3 – Application Case:
Metro Insurance processes 5,000 claims per month. The standard process has 8 steps and should
take 7 days. However, average cycle time is 18 days, and 30% of claims are returned for re-
submission. Using the DMAIC framework, define the problem statement, identify which step is likely
the bottleneck (given that the document verification step takes 4 days on average), and propose
three process improvement interventions. Design a KPI dashboard for the Claims Processing
Manager.

BTL Level 4 – Analysis Case:

After process mining analysis, Metro Insurance discovers three major insights: (i) 40% of delays
occur when a claim requires senior underwriter approval for amounts above ₹5 lakhs; (ii) claims
submitted digitally (vs. paper) process 65% faster; (iii) auto-approved claims have a 12% higher
fraud incidence. Critically evaluate the trade-off between speed and fraud risk in process
automation. Should the approval threshold be raised to ₹10 lakhs? Develop a business case with
risk-adjusted ROI and identify which processes should NOT be automated.
6. Financial Analytics
6.1 Definition and Overview
Financial Analytics is the application of data analysis, statistical modelling, and business
intelligence to financial data for purposes of understanding past performance, monitoring current
health, and predicting future financial outcomes. It transforms raw financial data into actionable
insights for CFOs, investors, regulators, and business managers across budgeting, forecasting, risk
management, and investment decisions.

6.2 Key Concepts (Sub-Points)

• Financial Performance Analytics: Analyses income statements, balance sheets, and cash
flow statements using ratio analysis (profitability, liquidity, solvency, efficiency ratios) to
assess financial health and benchmark against peers.
• Budgeting and Forecasting Analytics: Uses rolling forecasts, driver-based planning, and
scenario modelling to predict revenues, costs, and cash flows. Replaces static annual
budgets with dynamic financial plans.
• Risk Analytics: Quantifies financial risks—credit risk, market risk, liquidity risk, operational
risk—using models like Value at Risk (VaR), Monte Carlo simulation, and stress testing.
Required by Basel III/IV for banks.
• Fraud Detection Analytics: Uses anomaly detection, network analysis, and supervised ML to
identify unusual transactions, account takeovers, payment fraud, and financial statement
manipulation in real-time.
• Credit Scoring Analytics: Builds models to assess the creditworthiness of individuals and
businesses using financial history, behavioral data, and alternative data (utility payments,
mobile usage). Powers lending decisions.
• Working Capital Analytics: Optimizes Days Sales Outstanding (DSO), Days Payable
Outstanding (DPO), and Days Inventory Outstanding (DIO) to minimize working capital
requirements and maximize free cash flow.
• Investment Portfolio Analytics: Applies Modern Portfolio Theory (Markowitz), factor models
(Fama-French), and risk-adjusted return metrics (Sharpe Ratio, Alpha, Beta) to optimize
investment portfolios.
• Financial Scenario and Sensitivity Analysis: Models the impact of key variable changes
(interest rates, forex, commodity prices, demand) on financial outcomes. Identifies which
variables have the highest financial impact.
• Revenue Leakage Analytics: Identifies unbilled services, contract non-compliance, pricing
errors, and discount abuse that lead to revenue being lost without showing up as a direct
expense.

6.3 Financial Analytics Dashboard Flow

┌──────────────────────────────────────────────────────────────┐
│ FINANCIAL ANALYTICS ARCHITECTURE │
└──────────────────────────────────────────────────────────────┘

DATA SOURCES:
 [ERP/SAP] [Banking Systems] [Excel/Legacy] [Market Data]
│ │ │ │
└──────────────┴────────────────┴───────────────┘
│
[Data Warehouse / Lake]
│
┌──────────────┼──────────────┐
│ │ │
[Descriptive] [Predictive] [Prescriptive]
Historical Revenue Optimal Capital
P&L Reports Forecast Allocation
Ratio Analysis Risk VaR Tax Strategy
│ │ │
└──────────────┴──────────────┘
│
[CFO / Finance Dashboard]
Real-time KPIs + Alerts

6.4 Key Performance Indicators (KPIs)

KPI Formula Sector Benchmark
Return on Equity (ROE) Net Income / Shareholders' BFSI / > 15%
Equity × 100 Manufacturing
EBITDA Margin (EBITDA / Revenue) × 100 All Sectors > 20%
Current Ratio Current Assets / Current All > 1.5
Liabilities
Debt-to-Equity Ratio Total Debt / Total Equity Capital Intensive < 1.5
Free Cash Flow (FCF) Operating Cash Flow – Tech / FMCG Positive & growing
Capital Expenditure
Days Sales Outstanding (Accounts Receivable / B2B / < 45 days
(DSO) Revenue) × 365 Manufacturing
Revenue Forecast 1 – |Actual – Forecast| / All > 90%
Accuracy Actual × 100
Value at Risk (VaR 95%) Maximum loss at 95% Banking / Within limits
confidence level Insurance

6.5 Case Study: HDFC Bank – Credit Risk & Fraud Analytics
Background:
HDFC Bank, India's largest private sector bank with a loan portfolio exceeding ₹14 lakh crore,
needed to improve credit risk assessment while simultaneously detecting real-time payment fraud
across its digital channels processing 3+ crore transactions daily.
Credit Analytics Approach:
• Replaced traditional CIBIL score-only credit assessment with a multi-variable credit scoring
model incorporating: banking transaction patterns, GST filing history, utility payment
regularity, digital footprint (for SME lending), and social network signals.
 • Used Random Forest classifier and logistic regression ensemble model trained on 5 years of
loan performance data. Model improved NPA prediction accuracy from 72% to 91%.
• Reduced decision time for SME loans from 21 days to 59 minutes using analytics-driven
automated underwriting.
Fraud Analytics Approach:
• Deployed real-time fraud detection using graph analytics (identifying fraud rings) and
behavioral biometrics (typing patterns, device fingerprinting).
• ML models analysed 150+ transaction features including: transaction amount, time, location,
merchant type, device change, and velocity.
Results:
• Credit NPA (Non-Performing Assets) ratio improved from 1.47% to 1.12%—₹4,900 Cr in
avoided bad loans annually.
• Fraud losses reduced by 65%, saving ₹800+ Cr per year.
• Digital loan approval boosted retail loan disbursements by 40%.

6.6 Chapter-End Cases

BTL Level 3 – Application Case:
Meridian Manufacturing Ltd. has provided you with 3 years of financial data: Revenue: ₹250 Cr
(Yr1), ₹285 Cr (Yr2), ₹310 Cr (Yr3); EBITDA Margins: 18%, 15%, 13%; Current Ratio: 2.1, 1.8, 1.4;
DSO: 38, 45, 58 days. Calculate key financial ratios, identify three financial health concerns, build a
forecast for Yr4 using trend analysis, and recommend two analytics-based interventions.

BTL Level 4 – Analysis Case:

Meridian's CFO is considering taking on ₹100 Cr debt at 11% interest rate to fund expansion.
Simultaneously, working capital is deteriorating (DSO rising, inventory days increasing). Conduct a
sensitivity analysis showing: how a 5% revenue decline would affect debt service coverage; the
break-even point of the expansion; the risk-adjusted NPV under optimistic, base, and pessimistic
scenarios. Should the company proceed? What financial analytics capabilities would you build to
continuously monitor this investment?
7. Big Data Analytics
7.1 Definition and Overview
Big Data Analytics refers to the process of examining large and varied data sets—known by the 5
Vs (Volume, Velocity, Variety, Veracity, Value)—to uncover hidden patterns, correlations, market
trends, customer preferences, and other useful business information. It goes beyond traditional data
analytics tools and databases to handle petabyte-scale data using distributed computing
frameworks like Hadoop and Spark.

7.2 Key Concepts (Sub-Points)

• The 5 V's of Big Data: Volume (terabytes to exabytes of data); Velocity (data generated and
processed in real-time—IoT sensors, social media streams); Variety (structured tables,
semi-structured JSON/XML, unstructured text/images/video); Veracity (data quality,
accuracy, and trustworthiness challenges); Value (deriving business insights from data that
justify storage and processing costs).
• Batch vs. Stream Processing: Batch processing (Hadoop MapReduce) processes large
historical datasets periodically. Stream processing (Apache Kafka, Apache Flink, Spark
Streaming) processes data in real-time as events occur. Both are used in complementary
ways: batch for deep analysis, stream for real-time decisions.
• Distributed Storage and Processing: Hadoop HDFS distributes storage across commodity
hardware. Apache Spark (in-memory computing) processes data 100x faster than
MapReduce. Cloud equivalents: AWS S3/EMR, Google BigQuery, Azure Data Lake.
• NoSQL Databases: Non-relational databases designed for big data: MongoDB (document),
Cassandra (column), Redis (key-value), Neo4j (graph). Each optimized for different data
types and query patterns.
• Machine Learning at Scale: Platforms like Apache Spark MLlib, [Link], and cloud ML
services (AWS SageMaker, Google Vertex AI) enable training ML models on massive
datasets. Auto-scaling handles demand spikes.
• Data Lake Architecture: A centralized repository storing raw data in native format until
needed. Unlike data warehouses (structured), data lakes accommodate all data types.
Governance layers (Delta Lake, Apache Iceberg) add ACID transactions.
• Real-Time Analytics Platforms: Combines Kafka (ingestion) + Flink/Spark Streaming
(processing) + Elasticsearch/Druid (query) + Kibana/Grafana (visualization) for sub-second
analytics. Used in financial trading, fraud detection, and IoT monitoring.
• Big Data in IoT: Connected devices (sensors, smart meters, wearables) generate
continuous streams of time-series data. Analysed using edge computing (local analysis at
device) + cloud analytics (deep pattern recognition).
• Data Governance and Privacy in Big Data: Managing data quality, lineage, access control,
and compliance (GDPR, India's DPDP Act) across petabyte-scale environments. Tools:
Apache Atlas, Collibra, Alation.

7.3 Big Data Architecture Flowchart

┌──────────────────────────────────────────────────────────────┐
│ LAMBDA ARCHITECTURE (Big Data) │
└──────────────────────────────────────────────────────────────┘

DATA SOURCES: IoT / Apps / Social / Transactions / Logs

│
┌───────────┴─────────────┐
│ │
[BATCH LAYER] [SPEED / STREAM LAYER]
Hadoop HDFS Apache Kafka → Spark Streaming
Historical Data Real-time Events
Deep Analysis Sub-second Decisions
│ │
└───────────┬─────────────┘
│
[SERVING LAYER]
Hive / Presto / BigQuery
Query Engine for Combined Views
│
[ANALYTICS & VISUALIZATION]
Tableau / Power BI / Kibana
Business Users & Data Scientists

7.4 Key Performance Indicators (KPIs)

KPI Description Sector Target
Data Processing Throughput GB/TB processed per hour Tech / Telecom SLA-dependent
Data Pipeline Latency Time from data ingestion to Finance / Retail < 5 seconds
availability (stream)
Data Quality Score % of data meeting quality All > 95%
rules
Query Response Time Time to execute analytics E-Commerce < 2 seconds
query
Storage Cost per TB Monthly cloud storage cost All Minimize
per terabyte
Model Training Time Time to retrain ML model on AI Companies Minimize with GPU
full dataset
Data Freshness Age of most recent data in Fraud / Trading < 60 seconds
analytics system

7.5 Case Study: Netflix – Big Data for Personalization

Background:
Netflix serves 270M+ subscribers globally, streaming 165M+ hours of content daily. Every second
generates millions of data points: what users watch, pause, rewind, skip, search for, and rate. The
challenge: use this data to personalize every user's experience and make content investment
decisions.
Big Data Architecture:
 • Netflix uses AWS cloud infrastructure with custom big data stack: Apache Kafka ingests 1.3
trillion events/day; Apache Spark processes viewing behaviour; Cassandra stores 1.3
petabytes of member data.
• The recommendation engine (Cinematch + collaborative filtering) analyses 30+ signals per
user including: time of day, device used, browsing patterns, completion rates, and similar-
user preferences.
Key Analytics Applications:
• Content Recommendation: Personalized rows on homepage use A/B tested ML models.
80% of hours streamed come from recommendations.
• Content Investment Decisions: Big data analysis of viewer preferences, regional tastes, and
content completion rates guided the ₹100 Cr investment in original Indian content like
Sacred Games.
• Thumbnail Optimization: A/B tests 20+ thumbnail images per title and shows each user the
thumbnail they're most likely to click based on viewing history.
Results:
• Personalization saves Netflix $1B annually in reduced churn.
• Original content (guided by analytics) has 3x higher engagement than licensed content.

7.6 Chapter-End Cases

BTL Level 3 – Application Case:
SmartCity Bengaluru has deployed 10,000 IoT sensors across the city measuring: traffic flow (per
minute), air quality (per 5 minutes), water consumption (per hour), electricity usage (per 15
minutes), and incident reports (real-time). Total data: ~500GB/day. Design a big data architecture
(tools and layers) to: ingest this data in real-time, store it efficiently for 5 years, run daily traffic
optimization models, and provide a real-time dashboard for city administrators.

BTL Level 4 – Analysis Case:

After 6 months of operation, SmartCity's data reveals: traffic congestion on 3 corridors costs ₹8
Cr/day in productivity loss; air quality in 5 zones consistently exceeds safe limits; 15% of water
distribution is lost to leakages. Evaluate whether a Lambda or Kappa architecture is more
appropriate for each use case. Critically assess data privacy concerns when citizens are tracked,
analyse trade-offs between edge analytics (on sensor) versus cloud analytics, and build a priority
roadmap for analytics investments with estimated ROI.
8. Web and Social Media Analytics
8.1 Definition and Overview
Web Analytics is the collection, measurement, analysis, and reporting of website data to understand
and optimize web usage. Social Media Analytics extends this to social platforms—tracking brand
mentions, sentiment, influencer reach, and community engagement. Together, they provide digital
intelligence that informs marketing strategy, product development, customer service, and brand
management.

8.2 Key Concepts (Sub-Points)

• Website Traffic Analytics: Analyses sources of traffic (organic, paid, referral, direct, social),
volume trends, geographic distribution, device types, and browser usage. Tools: Google
Analytics 4 (GA4), Adobe Analytics. Key insight: identifies which channels drive quality traffic
vs. vanity metrics.
• User Behaviour Analytics (UBA): Studies how users navigate websites—click patterns, scroll
depth, heatmaps, session recordings, form abandonment. Tools: Hotjar, Crazy Egg,
Microsoft Clarity. Reveals UX problems invisible in aggregate metrics.
• Conversion Rate Optimization (CRO) Analytics: Analyses the conversion funnel from landing
to purchase. Identifies drop-off points, tests hypotheses through A/B testing, and
implements changes to improve conversion rates. A 1% CRO improvement can mean
millions in revenue.
• Search Engine Optimization (SEO) Analytics: Measures organic search performance—
keyword rankings, impressions, click-through rates, Core Web Vitals, backlink profile. Tools:
Google Search Console, SEMrush, Ahrefs. Informs content strategy and technical SEO
investments.
• Social Media Listening & Monitoring: Tracks brand mentions, hashtags, and conversations
across platforms (Twitter/X, LinkedIn, Instagram, Facebook, Reddit, YouTube) in real-time.
Detects emerging crises, competitor moves, and trending topics. Tools: Brandwatch,
Sprinklr, Meltwater.
• Sentiment Analysis on Social Data: Applies NLP (VADER, BERT-based models) to classify
social media mentions as positive, negative, or neutral. Tracks sentiment trends around
product launches, campaigns, and issues. Goes beyond volume to understand emotional
tone.
• Influencer Analytics: Identifies high-impact influencers by analysing reach, engagement rate,
audience authenticity (fake follower detection), content relevance, and conversion
attribution. Moves beyond follower counts to true influence metrics.
• Social Commerce Analytics: Measures the revenue impact of social media—Instagram
Shopping, Pinterest buyable pins, YouTube product links. Tracks the social-to-purchase
journey and calculates Social ROAS (Return on Ad Spend).
• Competitive Intelligence Analytics: Monitors competitor web traffic (SimilarWeb), social
share of voice, search rankings, and content strategy. Enables benchmarking and identifies
competitive threats and opportunities.

8.3 Web & Social Media Analytics Funnel

┌───────────────────────────────────────────────────────────────┐
│ WEB & SOCIAL ANALYTICS: AWARENESS TO ADVOCACY │
└───────────────────────────────────────────────────────────────┘

[SOCIAL MEDIA] [SEARCH/SEO] [PAID DIGITAL]

Impressions, Reach Organic Traffic Impressions, Clicks
│ │ │
└──────────────────────┴─────────────────────┘
│
WEBSITE ENTRY
Sessions, New vs. Returning
│
┌─────────┴──────────┐
│ │
ENGAGE BOUNCE
Pages/Session > 2 Bounce Rate > 70%
Session Duration > 2min → Fix landing page
│
▼
MICRO-CONVERSION
Email signup, Add to cart
Content download
│
▼
MACRO-CONVERSION
Purchase / Lead / Registration
│
▼
RETENTION & ADVOCACY
Return visits, Reviews, Shares
NPS, Referrals

8.4 Key Performance Indicators (KPIs)

KPI Formula / Measure Platform Benchmark
Bounce Rate (Single-page sessions / Total Website < 40–55%
sessions) × 100
Average Session Duration Total session time / Total Website > 2 minutes
sessions
Engagement Rate (Social) (Likes+Comments+Shares) / Instagram / > 3–5%
Reach × 100 LinkedIn
Share of Voice (SOV) Brand mentions / Total Social Media > 30% in category
category mentions × 100
Net Sentiment Score % Positive – % Negative All Platforms > 60%
mentions
Domain Authority (DA) Moz score 1–100 based on SEO > 50 for large
backlink profile brands
Social ROAS Revenue from social ads / E-Commerce > 3:1
Social ad spend
Cost per Click (CPC) Total ad spend / Number of Paid Search Sector-specific
clicks
8.5 Case Study: Zomato – Social Media & Web Analytics
Background:
Zomato, India's leading food delivery platform, serves 20M+ monthly active users and operates in
500+ cities. Its brand is built significantly through digital engagement—witty social media posts,
data-driven content, and real-time marketing. Zomato uses analytics extensively to drive growth and
manage brand reputation.
Social Media Analytics Approach:
• Zomato monitors 50,000+ daily brand mentions across Twitter/X, Instagram, and food
review forums using Brandwatch and custom sentiment models trained on food/restaurant
domain data.
• Real-time sentiment dashboards alert the social team within 5 minutes of a negative
sentiment spike (e.g., a quality incident or delivery delay complaint going viral).
• Tracks Share of Voice versus Swiggy across 12 dimensions: delivery time perception,
customer service sentiment, app usability, food quality, and pricing value.
Web Analytics Application:
• GA4 funnel analysis revealed 35% of users who viewed a restaurant page added items to
cart but didn't complete payment. A/B testing of the checkout page reduced cart
abandonment by 22%.
• Search analytics showed 'biryani near me' had 10x more searches on Friday evenings—
Zomato now promotes biryani restaurants in notifications sent Thursday evening, increasing
conversion by 18%.
Influencer Analytics:
• Replaced reach-based influencer selection with engagement-rate and conversion-attribution
model. Micro-influencers (50K–500K followers) showed 6x higher ROI than mega-
influencers (1M+).
Results:
• Social-driven customer acquisition increased by 40% with 15% lower CAC.
• Crisis response time improved from 4 hours to 18 minutes, preventing 3 potential viral
reputation incidents.
• Overall digital revenue contribution grew from 22% to 38% of total orders.

8.6 Chapter-End Cases

BTL Level 3 – Application Case:
Organic Harvest, a premium organic skincare brand, has launched a new face serum. It has ₹20
lakh/month digital budget. Website traffic: 80,000 sessions/month, bounce rate 72%, conversion
rate 0.8%, average order value ₹1,200. Using web and social analytics, identify the three biggest
issues to fix, design an A/B testing plan for the landing page, propose an Instagram influencer
strategy, and calculate the revenue impact of improving conversion rate to 2.5%.

BTL Level 4 – Analysis Case:

Six months later, Organic Harvest's web analytics shows: conversion rate improved to 1.9%;
bounce rate down to 55%; Instagram engagement rate is 8.2% with 3.5M reach; however,
sentiment analysis reveals 23% of brand mentions include complaints about 'greenwashing'
(product claims don't match customer experience). Critically analyse how unaddressed negative
sentiment could impact future conversion rates and brand value. Develop an analytics-driven
reputation management strategy. Evaluate whether the web analytics improvements are
sustainable if the underlying product perception issue is not addressed. What additional data
sources would strengthen this analysis?
9. Quick Reference: Module 3 Summary

Analytics Type Primary Goal Key Tools / Methods Example Company

HR Analytics Optimize workforce HRIS, ML Churn Models, IBM, Google, Infosys
decisions: hiring, Pulse Surveys
retention, performance
Supply Chain Reduce cost, improve EOQ, ARIMA, Celonis, Amazon, Flipkart,
Analytics OTIF, forecast demand SAP SCM Maersk
Marketing Analytics Maximize marketing MMM, Attribution, A/B HUL, P&G, Zomato
ROI and campaign Testing, GA4
effectiveness
Customer Analytics Improve retention, RFM, Churn Models, Airtel, HDFC, Netflix
CLV, and NPS, CRM
personalization
Business Process Eliminate waste, Process Mining, DMAIC, Siemens, Accenture,
Analytics improve compliance, RPA, Celonis TCS
reduce cycle time
Financial Analytics Risk management, VaR, Monte Carlo, Credit HDFC Bank, JPMorgan,
forecasting, fraud Scoring, ML Fraud Deloitte
detection
Big Data Analytics Derive insights from Hadoop, Spark, Kafka, Netflix, Jio, Google
high volume, velocity, NoSQL, Cloud ML
variety data
Web & Social Media Optimize digital GA4, Hotjar, Zomato, Swiggy,
Analytics presence, brand, and Brandwatch, NLP Flipkart
conversion Sentiment

BTL Level Framework Reference

BTL Level Cognitive Skill Verb Examples Applied in This Module
Level 1 – Recall facts Define, List, Name, State Define HR Analytics, List
Remember 5Vs
Level 2 – Explain concepts Explain, Describe, Explain RFM Segmentation
Understand Summarise
Level 3 – Apply Use in new situations Calculate, Design, Chapter-end application
Implement, Apply cases
Level 4 – Analyse Break down & Analyse, Compare, Chapter-end analysis
evaluate Evaluate, Critique cases
Level 5 – Evaluate Judge with criteria Justify, Assess, Case study
 Recommend recommendations
Level 6 – Create Build new solutions Design, Formulate, Analytics framework design
Construct

— END OF MODULE 3: FUNCTIONAL ANALYTICS —

Prepared with industry case studies, KPI tables, process flowcharts, and BTL 3 & 4 chapter-end cases.