Understanding E-Commerce Sales Performance: Insights for Strategic Business Decision-Making (Published)
This study examines sales performance in a synthetic e-commerce dataset and evaluates the business relevance of an order-level revenue model. The analysis covers 5,000 transactions with information on product category, region, quantity, price, discount, payment method, delivery time and customer rating. Descriptive statistics, grouped revenue summaries, visualisations and Pearson correlations are combined with an ordinary least squares regression model. Recorded total revenue is 5,109,775.74 monetary units, and average revenue per order is 1,021.96. Electronics accounts for the largest revenue share at 35.81%, while Beauty has the highest average order value at 1,059.28. Card payments represent 45.40% of transactions. Revenue has positive correlations with unit price and quantity, whereas delivery time and customer rating have little linear association with revenue. The recorded regression results indicate a test R² of 0.870 and mean absolute error of 219.76. However, displayed transactions suggest that revenue is derived from quantity, price and discount, limiting the model’s independent predictive value. The study demonstrates how managers can interpret sales dashboards while distinguishing descriptive patterns, accounting relationships and causal claims. Its managerial recommendations are proposals for validation with real business data rather than evidence of proven commercial effects.
Keywords: E-Commerce, Linear Regression, Sales performance, average order value, business analytics, synthetic data