Post Graduate Diploma in Management (2023–25)

MARKETING ANALYTICS

MARKET BASKET ANALYSIS OF E-COMMERCE

Submitted to: Professor Niharika Gupta

Submitted by: Group 7

Aastha Singh 23PGDM001

Amshu Jain 23PGDM004

Geetika Sharma 23PGDM252

Payal Sharma 23PGDM

Smiti Joshi 23PGDM166

 Naba Nazneen 23PGDM093
Problem Statement
Analyze transaction data to identify product purchase patterns using association rules
(e.g., Apriori algorithm).
Recommend cross-selling or bundling strategies.

What is Market Basket Analysis

Market basket analysis is a data mining technique that businesses employ to uncover
purchasing patterns by examining which items customers frequently buy together, thereby
enabling them to optimize sales strategies through cross-selling, targeted promotions, and
improved product placement; it also aids in inventory management by forecasting demand
based on these identified associations, and provides valuable market insights for informed
decision-making regarding product development and marketing, ultimately enhancing customer
experience and driving revenue growth.

Where have we collected the data from ?

The data has been collected from blinkit, Zepto and other food commerce website for their
commonly sold produce

How are we approaching the problem ?

By leveraging association rule mining, we analyze key metrics like support, confidence, and
lift to uncover strong relationships between items frequently bought together. This insight
enables us to develop strategic recommendations for product placement, bundling,
promotions, and inventory management to maximize sales and enhance customer
experience. Additionally, we employ upselling strategies such as discounted bundles and
tiered pricing to increase the average order value (AOV). Our approach ensures that decisions
are backed by real transactional data, making them highly actionable and effective in
improving revenue and customer engagement.

Overview of Transaction Data

The dataset consists of 999 transactions involving 16 grocery items, including Banana,
Wheat, Lentil, Potato, Chips, Pulses, Mushroom, Ice Cream, Noodles, Milk, Nutella, Jam,
Sugar, Coffee, Paneer, and Jaggery. Each transaction is represented in a binary format
(True/False), indicating whether an item was purchased. This structure enables the
identification of frequently purchased items and their associations, allowing for data-driven
strategies in product placement, promotions, and bundling.
Based on the bar chart, there is the list of the Top 10 Most Frequently Occurring Items,
ranked from highest to lowest support:

1. Jaggery
2. Lentil
3. Paneer
4. Ice Cream
5. Sugar
6. Noodles
7. Chips
8. Milk
9. Potato
10. Jam

Inference from Market Basket Analysis Results

The dataset represents association rules derived from transactions, highlighting the
relationships between products based on confidence and lift values. Here’s a structured
interpretation of the findings:

1. Key Product Associations

● Strong associations with Jaggery

○ (Milk → Jaggery) with a confidence of 52.09% and a lift of 1.32

→ When Milk is purchased, there is a 52% probability that Jaggery will also
be bought.
○ (Jaggery → Milk) with a confidence of 50.11% and a lift of 1.32
→ This is a bidirectional relationship, meaning customers tend to buy
these items together.
○ (Paneer → Jaggery) with a confidence of 47.14% and a lift of
1.11 suggests Paneer buyers might also purchase Jaggery, but the lift is lower,
indicating a weaker association.
● Ice Cream and Lentil Show a Strong Correlation

○ (Ice Cream → Lentil) with a confidence of 49.28% and a lift of

1.16 suggests that Ice Cream buyers have a relatively high tendency to also buy
Lentils.
○ (Lentil → Ice Cream) has a similar confidence of 49.28%,
reinforcing this bidirectional relationship.
○ These insights could be leveraged for discount bundling or placing Ice Cream
and Lentils near each other in a store.
● Noodles and Lentil Have a Strong Connection

○ (Noodles → Lentil) and (Lentil → Noodles) both have a

confidence of ~49% and a lift of 1.17, meaning people who buy one of
these items have a higher-than-random likelihood of purchasing the other.
 ○ This indicates a strong cross-selling opportunity, where bundled offers could
work well.

2. Top Rules Based on Lift (Strongest Associations)

● (Jaggery → Pulses) has a lift of 1.88, meaning this combination is highly likely
to be bought together.
● (Milk → Noodles) with a lift of 1.20 is another key finding, meaning Milk buyers
have an increased tendency to purchase Noodles.
● (Nutella → Lentil) with a lift of 1.17 suggests a moderate relationship.

3. Business Implications
A. Product Placement and Promotions

● Place Milk and Jaggery together as they have a high confidence and lift. Promotions
like "Buy Milk, get 10% off on Jaggery" could work well.
● Pair Noodles and Lentils in marketing campaigns (e.g., combo packs or shelf
placements).

B. Cross-Selling Opportunities

● Jaggery and Pulses should be bundled for better upselling, as they show a strong
relationship.
● Ice Cream and Lentils can be targeted together through ads and product placement.

C. Inventory Optimization

● Ensure sufficient stock of high-lift pairs like Milk & Jaggery, Noodles & Lentils to
avoid stockouts and maximize revenue.

Upselling Strategies Based on Market Basket Analysis 🚀

Given the strong product associations observed in your data, here are some effective
upselling strategies to increase revenue:

1. Bundle High-Lift Items Together

Strategy: Create value bundles with products that frequently appear together in transactions.
🔹 Examples:

● Milk + Jaggery Combo: Since this pair has a high confidence (52%) and lift (1.32), offer
a "Buy 1L Milk, Get 10% Off on Jaggery" promotion.
 ● Noodles + Lentils Pack: Customers already buy these together (lift = 1.17). Offer a
"Healthy Meal Combo" featuring these items at a slightly discounted price.
● Ice Cream + Lentils Deal: Since they have a strong connection, promote "Sweet &
Healthy Combo" with Ice Cream and Lentils at a discounted rate.

Bundling increases average order value (AOV) and ensures customers buy more than they
initially intended.

2. Upsell Through Tiered Pricing

Strategy: Encourage customers to buy larger quantities or premium versions of associated
products.
🔹 Examples:

● When a customer adds Milk, show a pop-up: "Upgrade to Organic Milk & Get Free
Jaggery!"
● If they purchase Noodles, suggest a higher-end pasta brand with a message:
"Upgrade your meal with premium whole wheat pasta!"

The Market Basket Analysis reveals strong product associations, such as Milk & Jaggery, Ice
Cream & Lentils, and Noodles & Lentils, highlighting valuable cross-selling and bundling
opportunities. High-lift pairs like Jaggery & Pulses (1.88) and Milk & Noodles (1.20) suggest
strategic promotions and product placements to maximize sales. Businesses can leverage
these insights by optimizing inventory, offering bundled discounts, and implementing
tiered pricing strategies to drive higher revenue. Ensuring high-association items are well-
stocked and positioned together can enhance customer experience and increase average order
value. Ultimately, these data-driven strategies improve both customer satisfaction and
profitability, making Market Basket Analysis a crucial tool for retail success.