An American Department Store
Retail stores have lots of moving parts and activities. People need real-time KPIs and metrics to have a full view of their performance. To unlock these capabilities for a major department store, CI&T worked together with them on several initiatives, such as campaign tracking, data warehousing, and data processing at scale using Hadoop and Spark in Google Cloud Platform. From cost reduction to better performance, the company was committed to move more workloads to the Google Cloud Platform (GCP).
The relationship between CI&T and the department store began with three projects. First, one that involved creating an email dashboard using the Google Cloud Platform to provide better and faster analysis. Another project included using BigQuery’s capabilities for housing Big Data and comparing it with Teradata, which was their original data warehouse. Finally, another experiment involved demonstrating the migration process of current Hadoop and Spark workloads to the cloud
After seeing a successful demonstration of GCP’s capabilities and how it could benefit the department store’s business, they invited CI&T to design a new analytics platform to provide the most accurate reports. These reports would be sent to managers across all stores, based on specific sets of KPIs, such as receiving merchandise in stores, measurement of lead times, unload times, merchandise transfers, delivery accuracy, and more—in real time.
The Solution and Result
The department store chain was looking for a way to consolidate its metrics and data, thus the development of the analytics platform known as the "Omni-Channel Reporting" (OCR).
They had several metrics. For example, its logistics process metrics included measurements involving merchandise that moved between warehouses and stores. When a customer makes an online purchase, there are several ways by which it can be fulfilled—either by a warehouse or by a store, via processes known as "buy online/pick-up in store" and "ship from store" or the traditional "ship from the warehouse" approach. Then there are processes specific to sales in stores, such as inventory tracking, order tracking, etc.
They wanted to have visibility of all of metrics and data on lead times, returns, trailer unload times, cartons sent to wrong stores, minimum inventory quantities, etc, in order to optimize their business. The idea was to reduce errors on delivery, reducing lead times, returns, and being more efficient on the order fulfillment.
The platform was built completely on Google Cloud, leveraging the computing power of Google Dataproc to crunch the numbers and prepare the data for reports. Google Cloud Storage was used as intermediary storage before sending data to Google BigQuery.
The OCR started in two pilot stores that then extended to all 1,200 stores across the U.S. Initially, it generated 18–23 reports per KPI, set once a day, that then turned to once every hour for all the stores. Currently, the CI&T team continues to work on adding more KPI sets to the OCR.
CI&T along with the marketing analytics, big data, and logistics departments have been able to develop a solution that provides more omnichannel reports and a seamless and smart replenishment for all stores.
Google Cloud Platform Products Used:
BigQuery, Dataproc, Dataflow, Google Cloud Storage (GCS), IAM, VPC, GKE, StackDriver, GCE