Marketing team achieves 30% reduction in campaign costs with the implementation of Data Quality & Customer Segmentation

Our Client

A conglomerate established in UAE during 1979, with portfolio extending across several industries, including automotive, retail, hospitality, real estate etc. The group entities operate in 6 countries in the Middle East, including nearly 200 stores and 23 showrooms in multiple markets in the Middle East.

Problem Statement

Having a customer base exceeding 1 million, the group acknowledged a lack of comprehensive understanding about their customers, hindering their ability to provide a holistic service. Given the diversity of customers across their retail, auto, and real estate sectors, the group identified a missing element: a unified perspective on customers and their relationships within the entire group. Additionally, the absence of a consistent customer profile across various systems collecting customer data prompted the leadership to consider a data quality and consistency assessment. This evaluation was deemed necessary to initiate the process of establishing a consolidated customer portfolio spanning the entire group.

The Solution

A two-stage solution was executed to tackle the identified issue through a collaboration between Amazon Web Services (AWS) and DataPhi.

In the first stage, a data quality assessment and sanitization exercise were conducted. The six different applications responsible for capturing customer data were integrated into the AWS data lake using S3, facilitated by AWS Glue for data engineering and transformation. Business rules for analyzing customer data quality were implemented within Glue, and the outcomes of the data quality assessments were transferred to AWS Redshift with a dimensional model to facilitate convenient visualization. Business stakeholders were provided with Data Quality dashboards based on Power BI, offering a comprehensive view of data quality. This empowered them to act on missing parameters or address issues related to poor-quality data within the respective applications capturing these elements.

In the subsequent phase, following the refinement of customer data, a unified customer view involving the treatment of duplicate customers across multiple applications was implemented. Leveraging various business rules to amalgamate customer information, a singular “Golden” customer record was generated to uniquely identify customers across all applications. AWS tools, particularly Glue, played a pivotal role in transforming and creating this Golden customer record. Through Power BI-based visualization, business users gained the ability to scrutinize customer relationships throughout the group. Customer segments were also formulated to classify customers based on factors such as spending patterns, purchase frequency, and transaction recency. The execution of customer segmentations was accomplished using AWS Sagemaker as a machine learning tool.


By implementing the solution, the group achieved a consolidated view of customers across their divisions, such as Retail and Auto. Through the application of machine learning-based segmentation, the group’s marketing team created dynamic segments tailored to specific campaigns, enabling a more targeted approach to customers compared to generic mass marketing methods. This focused segmentation resulted in a 30% reduction in campaign reach costs for the marketing team. The unified customer view facilitated by the dashboard contributed to a 22% improvement in performance related to customer profiling KPIs, offering enhanced visibility into their overall customer performance.

Also See

We'd like to hear from you

Contact Us