The launch of the Ford Ka marked a bold move by Ford into a rapidly evolving small-car market in France—one where consumer preferences were splintering across lifestyle, design, and emotional appeal. My goal was to help Ford identify the right target segments for the Ka using data-driven segmentation, and recommend how to best position the Ka to win market share from dominant competitors like Renault Twingo.

Using k-means clustering on both demographic and psychographic data, I developed a robust segmentation framework that uncovered distinct buyer personas, aligned with Ford's brand strengths, and informed targeted marketing recommendations.

The business challenge required:

• Moving beyond traditional automotive segmentation based primarily on product class and demographics
• Understanding the emotional and psychological drivers behind small car purchases
• Identifying segments where Ford Ka's unique design and personality would resonate most
• Developing actionable marketing strategies to reach and convert these segments

1. Traditional vs. Modern Segmentation

Historically, Ford and other automakers segmented the market by car size and product class—e.g., Type-A (<360 cm), Basic-B, Trend-B, etc.—and aligned these with basic demographics like age and income. But with the Twingo's runaway success, it became clear that small car buyers wanted more than value and compactness. Consumers were now making purchase decisions based on design, personality, and brand image.

I reframed the problem: instead of matching cars to fixed buyer categories, I used data mining to uncover what actually drives purchase decisions today.

2. Data Structure & Clustering Setup

I used the provided FordKaData.xlsx dataset, which included:

• Demographic data: Age, gender, marital status, income, first-time buyer flag, number of children
• Psychographic data: Responses to 62 attitudinal statements (e.g., "I want a trendy car")

I ran:

• K-means clustering on demographic data (k=7 optimal)
• K-means clustering on psychographic data (k=6 optimal)

I evaluated solutions using scree plots, R-squared measures, and centroid interpretation. I also used cross-tabulations with PreferenceGroups (Ka Choosers, Non-Choosers, and Neutrals) to evaluate each segment's alignment with Ka's perception.

Exploratory Insight: "Trendiness" Drives Preference

Q1 ("I want a car that is trendy") emerged as a key differentiator. Ka Choosers scored significantly higher than others. This insight helped validate the Ka's personality as a bold, youthful, trend-forward vehicle.

I used ANOVA tests and descriptive statistics to confirm the relationship. Visuals like boxplots and balloon plots made these insights easier to communicate to stakeholders.

Cluster Analysis Results

Demographic Clustering (k = 7)

These two clusters had the strongest Ka preference overlap, but lacked nuance on emotional drivers.

Psychographic Clustering (k = 6)

Clusters 2 & 3 stood out as the most promising segments for Ford Ka. Of the two, Cluster 3 was my top recommendation based on size, profitability, and alignment with Ford's values.

Messaging Strategy

This message appeals to urban buyers who care about:

• Practical design for city driving
• Aesthetic appeal that stands out
• Fuel efficiency and maneuverability

Other message pillars:

• "Compact size. Big statement."
• "For drivers who want more than a car—they want a vibe."

• Psychographics beat demographics for segmentation—two buyers may look the same on paper but seek entirely different things. This project confirmed that attitudinal data provides much richer insights for market segmentation.
• Data-driven marketing strategy doesn't just segment—it tells you how to win. Ka's uniqueness lies in its emotional appeal, not specs. The analysis clearly showed that the car's character was its main selling point for target segments.
• The car's polarizing design is a feature, not a bug—it helps capture attention in a crowded market. The segmentation analysis revealed that Ka's distinctive styling was precisely what attracted key buyer groups.

Tools & Techniques Used

• R (ggplot2, kmeans, openxlsx, CrossTable, balloonplot)
• Cluster centroid analysis
• Scree plot (elbow method) for k-value tuning
• Cross-tabulations with preference labels
• Psychographic cluster profiling

This approach can be extended to other product categories and markets where consumer preferences are increasingly driven by emotional and lifestyle factors rather than traditional demographic categories.