Retail Supply Chain

**Overview**

CPG is an exciting space for analytics, but analysts, data engineers and data scientists at CPGs are often lacking easy-to-access and up-to-date data. The data needed does exist, but it's often locked away in clunky retailer reporting and file sharing systems that are hard to integrate with modern data stacks.

Crisp connects to over 40 leading U.S. retailers and distributors. Our self service data pipeline ingests a live-feed of data from retail vendor portals to bring you real-time data directly into Databricks that is normalized, reliable, and ready to use — so you never have to download a report again. Of course, Crisp also integrates modern retailer and distributor systems such as Walmart Luminate and UNFI Insights.

**Types of supply chain data provided via Crisp:**

- **DC fill rates.** Delivery performance from DCs to stores and inbound to DCs.
- **DC Inventory.** DC on hand and on order inventory.
- **DC to store shipments.** Shipments from distributor or retailer DCs to stores
- **Store Inventory.** On Hand inventory and in-stock levels.
- **Much more.** Crisp also provides many other datasets, depending on source availability. For example, age of inventory and spoilage risk, chargebacks and deductions.

Not only does Crisp automate the ingestion of retailer and distributor data, we also can take care of cleaning and organizing the data so you can spend less time on data prep and more time on actual analysis.

**The Crisp Platform provides three different types of data, with different levels of pre-processing:**

**Source data.** For when you are developing a replacement for existing data pipelines and / or want to do your own data modeling. The source data schemata closely resembles what the upstream data source provides, so you can leverage any pre-existing modeling experience with this data.
**Normalized data.** For when you are developing retailer or distributor specific solutions and want to leverage source-specific attributes. The data model is a fully normalized schema and naming conventions are consistent across all data sources, making it easy to mix and match retailers in analyses.
**Harmonized data.** For when you are developing cross-retailer solutions, and it’s acceptable to use a narrower subset of data that is common across multiple data sources. The data model is built on the normalized model and combines common data elements, such as POS and inventory into a harmonized model across multiple retailers and distributors.

With Crisp, you can easily mix and match data from the models as you see fit.

**The applications of the data are vast, but some common use cases are:**

- **Reduce out of stock.** Use retailer inventory and sales data to easily detect out of stock and customer availability issues.
- **Increase distribution.** Use shipment data from distributors or retailers to identify Gaps in distribution.
- **Optimize DC Inventory.** Use DC order and inventory data to optimize DC replenishment and maximize fill rates.
- **Improve demand forecasting.** Combine retailer inventory and sales to generate demand forecasts and inventory allocation plans.

**概述**
快消包装品（CPG）行业是数据分析的优质赛道，但CPG企业内的分析师、数据工程师与数据科学家往往难以获取便捷可用且实时更新的数据。所需的数据确实存在，但往往被封锁在笨重的零售商报告与文件共享系统中，难以与现代数据栈集成。

Crisp已对接超过40家美国头部零售商与分销商。我们的自助式数据管道可从零售供应商门户获取实时数据流，将经过标准化处理、可靠且可直接使用的实时数据直接接入Databricks平台，让您无需再手动下载报告。此外，Crisp还集成了沃尔玛Luminate、UNFI Insights等现代零售商与分销商系统。

**Crisp提供的供应链数据类型如下：**
- **DC补货履约率**：配送中心（Distribution Center，DC）向门店的配送绩效，以及运入配送中心的入库绩效。
- **DC库存**：配送中心的在手库存与在途订单库存。
- **DC至门店发货量**：分销商或零售商配送中心向门店的发货量。
- **门店库存**：门店在手库存与现货库存水平。
- **更多数据类型**：根据数据源的可用情况，Crisp还提供诸多其他数据集，例如库存账龄、损耗风险、扣款与扣减费用数据。

Crisp不仅可自动化对接零售商与分销商的数据采集流程，还能协助完成数据清洗与整理工作，让您无需在数据预处理上耗费过多精力，从而将更多时间投入到真正的数据分析中。

**Crisp平台提供三类不同预处理程度的数据：**
**原始数据**：适用于您需要替换现有数据管道，或希望自行开展数据建模的场景。该数据的数据模式（schema）与上游数据源提供的格式高度贴合，您可依托此前针对该类数据的建模经验开展工作。
**标准化数据**：适用于您需要开发针对特定零售商或分销商的解决方案，且希望利用数据源专属属性的场景。该数据采用完全标准化的模式，命名规范在所有数据源中保持一致，可轻松实现跨零售商的数据混合与匹配分析。
**统一化数据**：适用于您需要开发跨零售商解决方案的场景，此时可使用多数据源共有的窄子集数据。该数据模型基于标准化模型构建，整合了销售点（Point of Sale，POS）与库存等通用数据元素，形成覆盖多家零售商与分销商的统一化数据模型。

借助Crisp平台，您可根据自身需求灵活混合与匹配各类模型中的数据。

**该数据的应用场景十分广泛，常见的使用案例包括：**
- **降低缺货率**：借助零售商库存与销售数据，可轻松识别缺货及顾客购买受阻问题。
- **优化铺货效率**：利用分销商或零售商的发货数据，可定位铺货缺口。
- **优化DC库存管理**：通过DC订单与库存数据，优化DC补货流程并最大化补货履约率。
- **提升需求预测精度**：整合零售商库存与销售数据，生成需求预测与库存分配方案。

搜集汇总

背景与挑战

背景概述

该数据集通过Crisp平台整合美国40余家零售商的实时供应链数据，提供原始数据、标准化数据和跨零售商统一数据三种层级，支持库存管理、分销优化和需求预测等供应链分析场景。数据经过清洗处理可直接用于分析。