Background
Our client is a company manufacturing consumer electronic products like mobile devices, printers, computer monitors and so on, who is leading the electronic goods merchant wholesalers industry for many years. Their advanced data analytics team connected to WeCloudData for the machine learning solution on predicting their top merchandize sales and marketing strategies on their key customers.
The project of E-commerce Customer Segmentation will be discussed in this article.
Problem Statement
The client wants to analyze the e-commerce customer spending patterns on certain product so as to guide the online promotions.
- Segment customers based on spending behaviour, time between multiple purchases.
- Identify the loyal customers and gain insights on how to improve the customer retention.
Tools used: SQL Server, Python, Tableau
Milestones
- Data Consolidation
- Transactional data
- Customer ID, Date of purchase, Transaction id of purchase, Total amount spent, Quantity of product ordered, etc.
- Project details
- Product Code, Product SKU, Product Description, Specifications of a device, Name of the device etc.
- Customer journey information
- 3rd party demographic data
- Postal code of the delivery address, City & Province on the delivery address, Average Household Income, Population etc.
- Transactional data
- Model Development
RFM (Recency, Frequency & Monetary) model has been used for this case and the RFM score for each customer were calculated and recorded as new features to the dataset.
Recency (in number of days): How recently a customer has purchased
Frequency (in number of purchases): How often they purchase
Monetary (in CAD): How much the customer spends
K-means clustering was then applied to come up with a RFM clusters separately with a compiled RFM score correspondingly of which the cluster number indicating the high, mid and low value customers.
Snapshot of the key findings
- Customer segments
- Purchase Trends by Segments
- Registered vs. Guest Customer
Achievements & Conclusion
WeClouddata has achieved several outcomes through an advanced analysis and machine learning methodologies:
- Targeted marketing for the identified customer clusters as per their product preferences and purchase behavior saved 10% of the overall marketing costs.
- Targeted marketing improved the loyalty and retention of respective client segments, majorly the high-value segment.
- The overall customer buying experience improves by understanding the purchase behavior. It offers the right product at the right time, which lead to enhanced retention and shortening of repeat purchase cycles.
- Identified key geographies from where the clusters emerged and improved the marketing and service activity in those areas.
This project expanded the company’s ability in mastering the e-commerce customer segmentation and retail industry use cases.
