A retail giant approached JK Technosoft (JKT) with a problem that their order fulfillment distribution network was performing sub-par. There were delays in order fulfillment, and the source of origin of these delays was unknown. Customer satisfaction was dipping. Their predicted churn rate (a machine learning engine was in place to predict churn) had spiked. The client was worried that if the problem was not addressed soon, it would result in revenue and reputation loss.
- Delays in Distribution Network.
- Impacting Customer Satisfaction.
- Could potential impact the top-line.
The diagram below gives an overview of the current mode of operation. A traditional hub & spoke is in play. With each participant (at times a participant is referred as a node) operating their own system of record with limited shared data. Transparency was missing.
A permissioned blockchain network was designed to include grocery stores, the central hub, and warehouses as the participating nodes. The existing order management system remains unchanged. An endpoint of the blockchain system was exposed. All the current systems at the nodes trigger the endpoint. Timestamps at each hand-off is recorded.
The digtial team at JKT evaluted platforms like Ethereum, Multichain, and Hyperledger Fabric on the trade-off parameters:
- Cost of data storage.
- Computation requirement.
- Access control levels.
After due deligence Hyperledger Fabric was selected as the preferred platform to build the solution. Client was already using AWS EC2 services for hosting. Hence we deployed the Blockchain network on AWS. The following architecture diagram gives a high-level overview of the system described below.