Customer Segmentation and Marketing Strategy using Online Retail Dataset

Dataset Overview:

The Online Retail dataset consists of the following columns:

1. InvoiceNo: Unique identifier for each invoice.
2. StockCode: Unique identifier for each product.
3. Description: A description of the product.
4. Quantity: The number of units purchased.
5. InvoiceDate: The date when the purchase was made.
6. UnitPrice: The price per unit of the product.
7. CustomerID: Unique identifier for each customer.
8. Country: The country where the customer made the purchase.

---

Project Steps:

1. Exploratory Data Analysis (EDA):

   • The first step was to analyze the dataset using EDA. This helped me understand the structure of the data, check for missing values, and identify any outliers or patterns.
   • I found that there were some missing CustomerID values, which I handled by either removing those rows or imputing values where necessary.

2. Feature Engineering using RFM:

   • The main objective of the project was to identify customer segments and suggest marketing strategies. To achieve this, I used the RFM (Recency, Frequency, Monetary) technique, which helps in segmenting customers based on their purchase behavior:
     • Recency (R): How recently a customer made a purchase.
     • Frequency (F): How often a customer makes a purchase.
     • Monetary (M): How much money a customer spends.

3. Clustering with DBSCAN:

   • I initially applied DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to identify clusters. Using this method, I found 57 clusters, but this was too many for my goal, which was to have only 6 clusters.
   • Although the Silhouette Score for DBSCAN was 99%, I decided to switch to a more controlled clustering technique that would give me exactly 6 clusters.

4. Clustering with KMeans:

   • I used KMeans clustering to reduce the number of clusters to 6. After applying KMeans, the Silhouette Score dropped to 48, which was acceptable for this type of analysis, indicating a decent clustering result.

5. Cluster Profiling:

   • For each of the 6 clusters, I calculated the average Monetary, Recency, and Frequency values. This helped me understand the characteristics of each cluster, such as whether they consisted of high-spending or frequent buyers, or if they were new or less engaged customers.

6. Marketing Strategy:

   • Based on the RFM values and the characteristics of each cluster, I devised personalized marketing strategies to suggest to each group. For example:
     • For clusters with high Monetary and Frequency, I suggested loyalty programs or special offers to reward frequent shoppers.
     • For customers with low Recency but high Monetary, I recommended re-engagement campaigns to encourage them to make another purchase.
   • I stored these strategies in a DataFrame called marketing_strategy.

7. Visualization:

   • I visualized the number of customers in each cluster using a bar chart, which showed the distribution of customers across the 6 clusters.
   • I also created a mean RFM chart for each cluster to visualize how each cluster performed on the Recency, Frequency, and Monetary metrics.
   • Finally, I created a pie chart to show the percentage of customers in each cluster, providing a clear overview of the distribution of customers across the segments.

---

Conclusion:

In this project, I successfully performed customer segmentation using the RFM model and applied clustering algorithms such as DBSCAN and KMeans to identify 6 customer segments. Based on the characteristics of each cluster, I developed personalized marketing strategies for each group. The visualizations helped provide a clear picture of the customer distribution and their purchasing behavior. The final marketing strategies were stored in a DataFrame and ready for implementation.

---

Skills Demonstrated:

1. Data Cleaning and Preprocessing: Handling missing values, outliers, and transforming the dataset for clustering.
2. Feature Engineering: Using the RFM model to create relevant features (Recency, Frequency, Monetary) for customer segmentation.
3. Clustering Algorithms: Applying DBSCAN and KMeans to perform customer segmentation.
4. Cluster Profiling: Analyzing and interpreting the characteristics of each cluster based on RFM values.
5. Marketing Strategy Development: Creating targeted marketing strategies based on customer segmentation.
6. Data Visualization: Using bar charts, pie charts, and other visualizations to communicate cluster characteristics and customer distribution.

This project showcases a strong understanding of customer segmentation, clustering techniques, and marketing strategy formulation, all of which are essential for businesses looking to improve customer engagement and retention.