Challenges

A large communications leader in Asia-Pacific wanted to drive more targeted, data-driven marketing campaigns based on customer insights harnessed from various data sources to generate business income. They had organization-wide data silos, a high dependency on IT for analytic purposes, and no real data management strategy.

Our solutions

Celsior was brought in to develop an enterprise data lake and predictive models to support more targeted marketing campaigns. Celsior implemented the data lake using the open-source Apache Hadoop, which facilitates solving problems involving massive amounts of data and computation. The predictive models were implemented using R and Python. This ecosystem encouraged the use of information to support a variety of business purposes. We also established a data management system and mechanism to maintain the quality of data.

Business outcomes

Through the implementation of the data lake and predictive models, Celsior was able to give the communications organization:

• An enterprise view of its customer data for greater insight into customer experience and increased profitability.
• 80% predictive accuracy in analytical models for customer behavior, helping influence business decision-making for campaign programs.
• Higher performance than the legacy data marts.
• Reduction in operation and maintenance time.

Also Read – Explainable AI – A Key to the Practical Success of AI and ML