Streamlining Telecom Supply Chain using a Data Quality Management Model
Overview
In the age of rapidly increasing volumes of information – good quality data has become one of the topmost priorities of business executives. The Supply Chain segment in every organization regularly sees a large influx of unstructured data. Compromising on the quality of data can hamper high level business decisions from Strategic Supply Chain Management to Sales Planning. This demands the need for an effective and evolved Data Quality Management Model.
Mu Sigma collaborated with a Telecom giant and helped them build an innovative data quality management model that streamlines their supply chain.
The Problem
The Supply chain team of a leading telecom company was facing difficulty in capturing huge inconsistencies occurring in their data. They approached Mu Sigma for building a solution that could improve their Data Quality Management (DQM).
The Supply chain team utilizes multiple reports such as inventory, promotional data and sales reports for decision making. However, the quality of these reports was inconsistent. They faced multiple issues from blank reports to missing KPI’s used to track the inventory in stock.
Our clients were looking for a business solution that would help them generate error free reports and devise an alert mechanism that could diagnose any abnormalities in their data.
Current State
• Manual resolution of abnormalities in data
• Reports generated through telemetry i.e recorded through a device
• Data issues catered to after data capture
Gap
• Lack of a framework that could identify data errors in advance
• Heavily rule-based current framework was limited to syntax checks
• There was a tendency to miss seasonality and trend in data resulting in false alerts
• Constant manual intervention was time-consuming and challenging
Desired State
• A proper architecture that could govern the data quality
• Prompt Alerts/checks for any data issues
The Solution
Mu Sigma approached the problem by performing a 360-degree analysis of the data issues faced by the organization. It performed a thorough data quality assessment, devised Business rules and Math based techniques to detect data anomalies.
We used Machine Learning techniques for predictive correction of the data issues i.e. correction of data in advance by devising a set of rules.
A Data Governance Framework was set up which automated alerts. In addition to this, a user interface was added to monitor and report any discrepancies.
Solution Design
A streamlined DQM framework was built to cater to all the client’s current challenges. The model implemented suitable data governance processes by leveraging extensive machine learning techniques,
• Rules customized for key data Quality KPIs such as inventory, sales, purchase orders, insurance and warranty
• Automated and continuous Model monitoring framework
• To ensure future data accuracy – 20 Data Quality Management modules were built on parameters such as accuracy, validity, consistency, data distribution, statistical summary of the data etc.
The Impact
Our collaboration helped the client completely automate the data quality management process and lead to an 80% reduction in the occurrence of data issues. The advanced email alerts reduced resolution time by 70%. There are currently more than 120 tables being monitored by the team. Our next goal is to further scale the DQM system across all divisions of the company by building modules that can accommodate higher volumes of data.
Want to know more?
Write to us at TheSherpa@mu-sigma.com.