Predicting High-Revenue Product Categories

Executive Summary

This case study focuses on predicting which product categories are likely to generate the highest revenue in the upcoming quarter. By leveraging logistic regression modeling and historical sales data, the analysis identifies key product categories expected to exceed a $25,000 revenue threshold. The insights from this analysis enable businesses to target high-revenue categories for marketing and inventory optimization.

Problem Statement

Retail businesses often struggle to forecast which product categories will drive significant sales, limiting their ability to optimize marketing strategies and inventory management. This case study aims to predict which product categories will generate higher sales based on historical transaction data, guiding businesses to make data-driven decisions.

Approach

Dataset Overview: The dataset includes transactional data from July to September 2023, capturing various product categories and factors such as quantity sold, price, discounts applied, and customer location.

Study Framework: The objective is to predict which product categories will exceed the $25,000 revenue threshold using a logistic regression model. Key features include quantity sold, price, discount applied, and total revenue.

Logistic Regression Model: A logistic regression model was used to predict whether a product category will generate more than $25,000 in revenue for the next quarter.

Results

Revenue Performance by Product Category

• High-Revenue Categories:
• Home Decor: $28,451.61
• Clothing: $26,460.83
• Groceries: $25,246.71
• Low-Revenue Categories:
• Books: $24,858.80
• Electronics: $20,359.13

Insight: Home Decor, Clothing, and Groceries generated higher revenue, making them the top-performing categories, while Books and Electronics fell below the $25,000 threshold.

Quantity Sold and Discount Analysis

Home Decor had the highest quantity sold (530 units), followed by Clothing and Groceries, each selling more than 499 units. Despite having the highest discount applied (16.22%), Books remained a low-revenue category, indicating that high discounts did not significantly boost revenue.

Predictive Model Performance

ROC AUC Score: 1.0, indicating perfect prediction accuracy in identifying product categories that would exceed or fall below the $25,000 revenue threshold.

Model Confidence: The high predictive power of the model allows businesses to confidently rely on it for future planning and decision-making.

Customer Behavior Insights

Home Decor had the highest number of customers (101), with an average order value of $281.70, suggesting that customers spent significantly more in this category. Electronics had the lowest average order value ($214.31), contributing to its low revenue generation despite consistent customer demand.

Visualization

Explore the complete interactive visualization here:

Key Insights

• High Revenue Drivers: Home Decor, Clothing, and Groceries consistently generate high revenue, driven by strong customer demand and higher average order values.
• Low Revenue Challenges: Despite discounts, Books and Electronics remain under the $25,000 threshold, largely due to lower average order values.
• Predictive Model Confidence: The logistic regression model's accuracy, with an ROC AUC of 1.0, ensures reliable predictions for product category performance.

Recommendations

• Focus on High-Revenue Categories: Businesses should prioritize marketing and inventory for Home Decor, Clothing, and Groceries, as these categories are predicted to generate the highest revenue in the next quarter. Customizing promotions for these categories will likely drive sales growth.
• Reevaluate Discount Strategies for Low-Revenue Categories: While Books and Electronics saw higher discounts, they remained in the low-revenue segment. Consider revising discount strategies or bundling these products with higher-performing categories to boost overall sales.
• Target High-Value Customers: Focus marketing efforts on attracting high-spending customers in Home Decor, which has both the highest customer base and average order value. Understanding customer preferences can help businesses design better loyalty programs.

Conclusion

By using predictive modeling, businesses can accurately identify which product categories will generate the most revenue in the next quarter. Home Decor, Clothing, and Groceries are clear drivers of high revenue, while Books and Electronics require strategic adjustments to improve performance. The logistic regression model’s high accuracy offers businesses a reliable tool for data-driven decision-making, ensuring that marketing and inventory efforts are focused on maximizing profitability.