Retail Performance Analytics

I framed the project around a question that executives actually care about: are stronger sales creating sustainable profit? The dataset had enough revenue growth to look encouraging at first glance, but the profit trend was less stable. That tension became the center of the analysis and kept the case from becoming a generic sales dashboard.

The Sample Superstore file contained 9,994 order-line records and 21 variables, including order dates, customer segments, regions, categories, sub-categories, discounts, sales, and profit. I preserved the order-line level because the business questions depended on product, discount, and regional detail. This also let the Tableau workbook support multiple levels of aggregation without losing operational context.

The first dashboard summarizes the business in one place: total sales, total profit, profit ratio, average discount, and sales-profit movement over time. I wanted the opening view to make the main tension visible immediately. Sales reached $2.30M and profit reached $286K, but the trend chart shows why revenue alone is not enough to judge performance.

The category view made the first structural issue clear. Technology generated $145,454.95 in profit, while Office Supplies produced $122,490.80. Furniture, despite being a major revenue category, produced only $18,451.27 in profit. That difference changed the recommendation: the business should not simply push every category equally, because categories do not convert sales into margin at the same rate.

Discounting was the most important warning signal in the case. Across the full dataset, average discount was 15.6%, but high-discount records were destructive to margin. Records with 40% to 59% discounts had a -25.47% profit ratio, and records with 60% to 100% discounts had a -119.2% profit ratio. The dashboard translates that into a practical management rule: discounts need thresholds, not just campaign enthusiasm.

Regional analysis showed that performance was not evenly distributed. The West region led profit at $108,418.45 and a 14.94% margin, while Central delivered $39,706.36 and only a 7.92% margin. State-level and regional product views helped move the discussion from national averages to local decisions: product emphasis, pricing, and discount controls should differ by market.

The dynamic Top N and Bottom N product views were useful because they combined sales ranking with profitability color. A high-sales product is not automatically a good product if it has weak margin, and a low-sales product is not automatically a problem if it serves a profitable niche. This part of the case helped me present product portfolio decisions with more nuance.

The recommendations followed directly from the dashboards: reduce aggressive discounting, protect margins with approval thresholds, prioritize Technology, review Furniture pricing and cost structure, retain the Consumer segment, and use regional strategies rather than one national playbook. I also recommended that dashboards emphasize Profit Ratio alongside Sales so performance conversations do not reward volume at the expense of margin.

This project pushed me to act like both analyst and decision translator. I did the dashboarding work, but I also had to decide which charts deserved executive attention and which findings should become strategy. The final case is personal to me because it shows the way I like to use BI: not as decoration, but as a structured argument about what the business should do next.