Purchase Behaviour of Users in Online Retail

Introduction:

• This data is about the users who actually purchased some items after browsing the website and visited certain parts of website

Problem Statement:

• Customer segmentation is the problem of uncovering information about a firm’s customer base, based on their interactions with the business.
• In most cases this interaction is in terms of their purchase behavior and patterns. We explore some of the ways in which this can be used.

Approach Taken:

The approach taken is to divide the customer base in different segments, which will help in the understanding of following:

• Customer Understanding.
• Target Marketing.
• Finding Latent Customer Segments.
• Higher Revenue.

For the Customer Segmentation I used Clustering Technique.

Here is one of the plots:

The steps include:

• Exploratory Data Analysis
• Deciding the Clustering Strategy:
• Recency
• Frequency
• Monetary Value
• Data Cleaning
• Data Pre-processing.
• K-means Clustering
• Cluster Analysis
• Cluster Description.

Interpretation of Results:

Initial analysis tells us that not all the users who browsed, purchased items.

• Some of them bought.

• Others didn’t

• We see the maximum sale is from 22 to 3 hours (11 PM to 3 AM.)
• Comparatively less sales from 4:00 AM to 1:00 PM.
• The customers purchasing during 10:00 PM to 12:00 AM buy more costly items than the users purchasing from 12:00 AM to 3:00 AM.
• We have 41,008 unique customers but almost 10% of total sales are contributed by only 1000 customers (based on the cumulative percentage aggregation in the preceding output).
• The next thing we want to determine is how many unique items the firm is selling.
• Looking at 3-D plot of Recency, Frequency and Monetary values:
• People who buy with a higher frequency and more recency tend to spend more based on the increasing trend in Monetary value with a corresponding increasing and decreasing trend for Frequency and Recency, respectively.
• By looking at the boxplots of different clusters we see the difference in their Amount of purchase with maximum, minimum and median amount.

More plots are: