Inventory & Procurement Optimization for an E-Commerce leader, at the scale of hundreds of thousands of SKUs

The Opportunity

Operating at the scale of hundreds of thousands of SKUs, spanning private label, marketplace, and third-party brands, the company's procurement function was under immense pressure. Demand volatility driven by flash sales, seasonal peaks, and promotional campaigns made traditional reorder-point methods unreliable. Working capital was trapped in slow-moving and dead stock, while fast-moving SKUs faced recurrent stock-outs that directly impacted GMV and customer satisfaction. Vendor quality was inconsistent, with no systematic way to track sellable-vs-unsellable ratios at the supplier level. Inventory ageing costs were estimated but never precisely measured, making it impossible to quantify the true cost of overstock decisions.

• 01Hundreds of thousands of active SKUs across private label and third-party brands, each with distinct demand patterns, lead times, and margin profiles.
• 02Working capital locked in slow-moving and dead inventory with no visibility into ageing cost at the SKU or vendor level.
• 03Stock-outs on high-velocity SKUs directly eroding GMV and customer lifetime value during peak periods.
• 04Vendor quality inconsistencies, no structured framework to evaluate sellable-vs-unsellable ratios or hold suppliers accountable.

The Solution

We architected a comprehensive Inventory Ageing and Purchase Order analytics platform on AWS, ingesting transactional data from the company's OMS, WMS, and vendor systems into a purpose-built data warehouse. The engineering layer handled deduplication, SKU normalization across catalog variants, and time-series alignment for accurate ageing calculations. A Power BI front-end delivered real-time visibility across procurement, enabling category managers and procurement leads to move from reactive firefighting to strategic, data-informed purchasing decisions.

• 01Sellable vs. non-sellable inventory at vendor and SKU level.
• 02Inventory ageing by quantity and cost, bucketed to prioritize action on slow-moving stock.
• 03Historically non-performing SKUs to inform smarter future purchases.
• 04Pareto analysis of the long-tail SKU base, vendor consolidation and PO rationalization.
• 05Automated reorder recommendations based on velocity, lead time, and demand seasonality, replacing manual spreadsheet-based PO planning.
• 06Vendor scorecard integration, tracking fill rate, defect ratio, and lead-time adherence at the supplier level.

The Impact

Real-time visibility across procurement reshaped how the team made buying decisions, freeing significant working capital while protecting availability across the long tail. Category managers moved from weekly review cycles to continuous, exception-driven monitoring. Dead stock was identified and liquidated systematically rather than discovered during quarterly audits. Vendor negotiations became data-backed, with fill-rate and quality metrics replacing anecdotal feedback.

• 01Stock-outs reduced by 24% across the portfolio.
• 02Holding inventory value reduced by 17%, significant working capital freed.
• 03Procurement operations streamlined through automation, reactive firefighting replaced with proactive strategy.
• 04Vendor negotiation leverage strengthened through transparent, data-backed performance scorecards.
• 05Dead stock identification accelerated from quarterly audits to continuous, automated flagging, enabling faster liquidation and markdown decisions.