Undressing the Customer Mind: Deep Dive into Spending Behavior and Campaign Response

---

1. Business Overview

In today’s competitive retail market, understanding customer behavior is essential for effective marketing strategies. Since not all customers respond the same way to campaigns, segmentation and response analysis are critical.

This analysis focuses on two priorities:

1. High-Value Customers (HVC): Customers who generate significant revenue and demonstrate strong loyalty. Identifying their profiles enables tailored retention strategies.

2. Campaign Responsiveness: Analyzing how different customer segments respond to campaigns provides insights into engagement levels, helping optimize targeting, improve conversion rates, and allocate resources more effectively.

By distinguishing between customer segments and their behaviors, businesses can design precise strategies to strengthen loyalty, maximize lifetime value, and enhance overall marketing performance.

Key objectives:

• Identify the characteristics of high-value customers and develop data-driven strategies to retain them.
• Identify the characteristics of customer responses to marketing campaigns and determine appropriate strategies to enhance effectiveness.

2. Data Sources

The dataset used in this project is the Customer Personality Analysis.
It contains 2,240 customers with 29 variables describing demographics, lifestyle, and purchasing behavior.
The features are grouped into five categories:

👤 People

Feature	Description
ID	Customer's unique identifier
Year_Birth	Customer's year of birth
Education	Customer's education level
Marital_Status	Customer's marital status
Income	Customer's yearly household income
Kidhome	Number of children in the household
Teenhome	Number of teenagers in the household
Dt_Customer	Date of customer enrollment with the company
Recency	Days since the customer's last purchase
Complain	1 if the customer has complained in the last 2 years, 0 otherwise

---

🛍️ Products

Feature	Description
MntWines	Amount spent on wine in the last 2 years
MntFruits	Amount spent on fruits in the last 2 years
MntMeatProducts	Amount spent on meat products in the last 2 years
MntFishProducts	Amount spent on fish products in the last 2 years
MntSweetProducts	Amount spent on sweets in the last 2 years
MntGoldProds	Amount spent on gold products in the last 2 years

---

🎯 Promotion

Feature	Description
NumDealsPurchases	Number of purchases made with a discount
AcceptedCmp1	1 if customer accepted the offer in the 1st campaign, 0 otherwise
AcceptedCmp2	1 if customer accepted the offer in the 2nd campaign, 0 otherwise
AcceptedCmp3	1 if customer accepted the offer in the 3rd campaign, 0 otherwise
AcceptedCmp4	1 if customer accepted the offer in the 4th campaign, 0 otherwise
AcceptedCmp5	1 if customer accepted the offer in the 5th campaign, 0 otherwise
Response	1 if customer accepted the offer in the last campaign, 0 otherwise

---

🏪 Place

Feature	Description
NumWebPurchases	Number of purchases made through the company’s website
NumCatalogPurchases	Number of purchases made using a catalogue
NumStorePurchases	Number of purchases made directly in stores
NumWebVisitsMonth	Number of visits to the company’s website in the last month

3.Technologies Used

Category	Tools / Libraries
Programming Language	Python (Pandas, NumPy)
Data Visualization	Matplotlib, Seaborn, Plotly
Interactive Dashboard	Tableau
Development	Jupyter Notebook

4. Project Structures

The repository is organized as follows:
├── README.md          <- The top-level README for developers using this project.
│
├── data               <- Raw and processed datasets
│
├── notebooks          <- All stages of analysis (data cleaning, EDA, and reporting) were conducted within Jupyter Notebooks
│
├── slide             <- Presentation including background, objective, visuals and business strategies
│
├── dashboard          <- Tableau workbook and related files
│
└── requirements.txt   <- Python dependencies

5. Methodology / Approach

The project follows these main steps:

1. Data Cleaning & Preprocessing – Handling missing values, outliers, and encoding categorical data.
2. Exploratory Data Analysis (EDA) – Understanding demographics, income, and purchasing behavior patterns.
3. Feature Engineering – Creating new features such as total spend, average order, or customer tenure.
4. Segmentation Analysis – Identifying clusters of high-value customers using Length, Recency, Frequency and Monetary (LRFM) analysis.
5. Campaign Response Analysis – Measuring responsiveness across segments to improve targeting.
6. Visualization & Dashboarding – Presenting findings using Tableau for better decision-making.

6. Key Insigth

6.1 Characteristics of High-Value Customers (HVC)

• Champion & Loyal Customer (High Value Customer)

  • Adults (30–45), high income, well-educated.
  • Mostly in relationship, no children.
  • Spending concentrated on Wine & Meat.
  • Heavy multichannel users (web, catalog, store).
  • Responsive to high-value campaigns (Cmp1, Cmp5, Response).

• Big Spender (Growth Potential)

  • Adults, high income, no children.
  • High spending across all categories, store & catalog heavy.
  • Similar to HVC but not yet loyal.

• Potential Loyalist (Growth Potential)

  • Adults / Middle-aged, medium income, usually with 1 child.
  • Family-oriented spending (Meat, Fruits, Sweets).
  • Low campaign engagement.

6.2 Customer Responses to Campaigns

• Overall Acceptance

  • Generally low (6–7%), highest in the last Response campaign (15%).
  • Cmp2 failed (1.3%) → irrelevant design/targeting.

• By Segment

  • Most responsive: Loyal Customers, Big Spenders, Champions (25–37% acceptance in Cmp5/Response).
  • Moderate: Potential Loyalists (10–12%), Occasional Shoppers (13% in Response).
  • Low: At Risk, Loyal Lost, Dormant → mostly 0 acceptance.
  • New Customers: low, but start engaging in Response (~8%).

• Distribution

  • 70%+ customers never accepted any campaign (0 acceptance).
  • Only ~5% accepted ≥3 campaigns.
  • No segment consistently accepted all campaigns → no one-size-fits-all campaign works.

  7. Actionable Recommendation

  7.1 Retention Strategies

• Champion & Loyal Customer

  • VIP programs (tiers, reward points, early access).
  • Personalized campaigns → less frequent, more relevant.
  • Focus on premium & lifestyle positioning, not just discounts.

• Big Spender

  • Push conversion into Loyal Customers.
  • Personalized service, luxury bundles, cross-selling.
  • Store events & premium catalog campaigns.

• Potential Loyalist

  • Loyalty starter program (bonus points for repeat purchase).
  • Family bundles (Meat + Fruits + Sweets).
  • Nurturing campaigns → education & affordable offers to drive repeat purchases.

  7.2 Campaign Strategies

• Segmented Campaigns

  • Champions & Loyal → VIP perks, premium bundles.
  • Big Spenders → personalized offers, push to loyalty tier.
  • Potential Loyalists → nurturing family bundles + starter loyalty program.
  • Occasional & New → welcome/activation campaigns.
  • At Risk, Loyal Lost, Dormant → selective reactivation only.

• Channel Optimization

  • HVC → omnichannel (web, catalog, store).
  • Growth Potential → store-heavy, supported by digital.
  • New/Occasional → start with store promotions, then move to web/app.

• Benchmark Campaign

  • The last Response campaign was the most successful (15%).
  • Its design (product, channel, targeting) should be replicated in future campaigns.

  9. Brief Summary

In a competitive retail world, not all customers are created equal, some are loyal champions driving revenue, while others remain hesitant to engage. Through the Customer Personality Analysis dataset, this project uncovers the patterns behind who spends, how they shop, and what makes them respond. High-Value Customers emerge as adults with high income, well-educated, and often without children. They favor wine and meat, shop across every channel, and prove the most responsive when campaigns speak their language.

Yet, the story is not the same for everyone. Big Spenders and Potential Loyalists hold great promise but need the right push, luxury bundles for one, family packages for the other. Most campaigns fell flat, with acceptance rates as low as 6%, except for one that resonated strongly, achieving almost 15%. The lesson is clear: blanket strategies don’t work. By crafting personalized offers, rewarding loyalty, and learning from the campaigns that succeed, businesses can not only win more responses but also build lasting relationships that fuel longterm growth.

8. Contact

• Author: Naufal Fajar Revanda
• GitHub: nrevanda
• LinkedIn: Naufal Fajar Revanda
• Email: nrevanda@gmail.com