PBS

 The Predictive Buying System (PBS) is Amazon’s machine learning–driven replenishment engine used primarily in the Vendor Central (1P) model. It analyzes historical sales, seasonality, promotions, pricing, availability, and external trends to predict future demand and automatically create Purchase Orders (POs) to keep inventory levels optimal across fulfillment centers.

Key functions of PBS:

• Forecasts weekly customer demand per ASIN
• Triggers POs based on P-level forecasts (e.g., P50, P70, P90)
• Adjusts predictions for events like Prime Day, Black Friday, or Lightning Deals
• Considers inventory levels, lead times, capacity constraints, and vendor performance

Why PBS matters to vendors:

• Directly controls how much and when Amazon orders your products
• Impacts your Net Receipts, weeks of cover, and sell-in performance
• Affects whether your products stay in stock or go NR (non-replenishable)
• Can be influenced through manual overrides by Amazon Retail buyers in limited cases
• Vendors can monitor PBS behavior through tools like Demand Forecasts, Flow Reports, and Forecast Recast Events (FRE)

💡 Example: PBS forecasts a demand spike for a toy ahead of Q4. Based on its P90 projection, it issues a PO for 15,000 units across regional fulfillment centers.

In short:
PBS (Predictive Buying System) is Amazon’s automated demand forecasting and replenishment engine that predicts customer needs and issues POs to vendors - playing a central role in 1P inventory planning.