Home > Docs > Industries > Retail & CPG

🛒 Retail & CPG — Demand Forecasting & Customer 360¶

Last Updated: 2026-05-05 | Version: 1.0.0

"Retailers that act on data in minutes instead of days convert 23% more shoppers and waste 18% less inventory."

📑 Table of Contents¶

• Scenario Overview
• Regulatory Landscape
• Data Flow Architecture
• Why Fabric for Retail and CPG
• Getting Started
• References

🎯 Scenario Overview¶

📋 Regulatory Landscape¶

🏗️ Data Flow Architecture¶

flowchart LR
    subgraph Sources["🛒 Data Sources"]
        POS["POS Terminals<br/>(Event Hub)"]
        ERP["ERP / Inventory<br/>System"]
        LOYALTY["Loyalty &<br/>CRM Platform"]
        SUPPLY["Supply Chain<br/>/ EDI Feeds"]
        WEB["E-Commerce<br/>Clickstream"]
    end

    subgraph Bronze["🥉 Bronze Layer"]
        B1["POS Transactions<br/>(Eventstream)"]
        B2["Inventory Snapshots<br/>(daily extract)"]
        B3["Loyalty Events<br/>(append-only)"]
        B4["Purchase Orders &<br/>Shipments (EDI)"]
        B5["Clickstream<br/>(Eventstream)"]
    end

    subgraph Silver["🥈 Silver Layer"]
        S1["Sales Transactions<br/>(deduplicated, card masked)"]
        S2["Inventory Positions<br/>(reconciled)"]
        S3["Customer Master<br/>(PII pseudonymized)"]
        S4["Supply Chain<br/>Events (validated)"]
    end

    subgraph Gold["🥇 Gold Layer"]
        G1["Customer 360<br/>Star Schema"]
        G2["Demand Forecast<br/>Feature Store"]
        G3["Promotion<br/>Effectiveness"]
        G4["Supply Chain<br/>KPIs"]
    end

    subgraph BI["📊 Consumption"]
        DL["Direct Lake<br/>Semantic Model"]
        PBI["Power BI<br/>Dashboards"]
        RTD["Real-Time<br/>Dashboard"]
        DA["Data Activator<br/>Stock Alerts"]
    end

    POS --> B1
    ERP --> B2
    LOYALTY --> B3
    SUPPLY --> B4
    WEB --> B5

    B1 --> S1
    B2 --> S2
    B3 --> S3
    B4 --> S4
    B5 --> S3

    S1 --> G1
    S1 --> G2
    S1 --> G3
    S2 --> G4
    S3 --> G1
    S4 --> G4

    G1 --> DL --> PBI
    G2 --> DL
    G3 --> DL
    G4 --> DL
    B1 --> RTD
    S2 --> DA

💡 Why Fabric for Retail and CPG¶

POS streaming and batch analytics on one platform. Point-of-sale events flow through Eventstreams for real-time dashboards while the same data lands in Lakehouse Bronze for daily aggregation — no duplicate pipelines, no data reconciliation issues.

Demand forecasting without a separate ML platform. AutoML model endpoints train time-series forecasting models directly on Gold-layer sales data, producing daily or weekly demand predictions that feed inventory planning dashboards via Direct Lake.

Customer 360 without a CDP. By unifying POS, loyalty, e-commerce, and CRM data in the Lakehouse medallion pattern, retailers build a governed Customer 360 without purchasing a separate customer data platform. Sensitivity labels and pseudonymization protect PII.

Supply chain visibility across partners. Fabric's data sharing and Mirroring capabilities let retailers share filtered views of inventory and shipment data with CPG suppliers — enabling collaborative planning without raw data exposure.

Sub-second dashboards for store operations. Direct Lake semantic models give district and store managers Power BI dashboards that perform at import-mode speed while reading live from OneLake, eliminating overnight refresh windows.

🚀 Getting Started¶

1. Build the medallion Lakehouse — Follow Medallion Deep Dive to set up Bronze/Silver/Gold layers for POS and inventory data.
2. Stream POS transactions — Configure Eventstreams to ingest POS events from Event Hub for real-time sales dashboards.
3. Apply PCI and privacy controls — Enable OneLake Security CLS for card masking and configure GDPR deletion workflows for loyalty data.
4. Build Customer 360 — Create a Gold-layer star schema joining POS, loyalty, and clickstream data, then connect Direct Lake for unified customer analytics.
5. Deploy demand forecasting — Use AutoML to train a time-series model on sales history and publish predictions to the Gold layer.
6. Enable supply chain sharing — Use Data Sharing to provide filtered inventory views to CPG partners.

📚 References¶