Case Study
Retailer Enhances Customer Interactions Using Data Management
A large retail company operating online and in-person struggled to deliver a seamless customer experience across sales channels. Customer data was siloed in different systems: a point-of-sale (POS) system, an e-commerce platform, a CRM tool, and a customer feedback dataset. Correspondingly, if a customer called with an issue, customer service could not draw on customer-specific data and had to re-direct the call to other departments. Similarly, when the marketing team tried to analyze customer behavior, they realized they were operating with a limited point of view. While they could see how customers interacted with marketing emails in the CRM system, they could not access detailed transaction histories from the POS tool. After management caught wind of these problems, the company began to revamp its data management practices.
Implementing the Solution(s)
First, it centralized all data from the fractured systems into a unified storage environment: Azure Data Lake. Subsequently, the company could use the Azure Data Catalog, a searchable, organized catalog of the stored data. The firm established rules for adding metadata in Azure Data Catalog to ensure ease of search. For example, every dataset must have a metadata element indicating its status (Active, Deprecated, Archived), allowing employees to assess dataset quality and relevance quickly.Additionally, the company used Informatica Data Quality tools to set up automatic validation rules, as Azure could track, but not enforce data quality rules. For example, the company set up a rule that if a customer's “Account Status” is “Active,” then the “Last Purchase Date” must not be older than 12 months. Records failing this validation are flagged for review.
The company then set up workflows in Apache NiFi to automate real-time data transfer from existing systems (e.g., POS) to Azure. The firm could see customer data in real-time, integrated across physical and online sales. The business also noticed that using Apache NiFi reduced manual errors during transfers and allowed for simple automatic data cleaning (e.g., converting John Smith to Smith, John).
However, Apache NiFi is not a specialized cleaning tool. It could not handle the complex cleaning needed to make the company's customer data consistent, complete, and without duplicates. Therefore, the firm continued to rely on Informatica to standardize customer information and enrich incomplete profiles using third-party demographic datasets.
Results
With centralized, standardized, and accessible data, the company improved operational efficiency. Customer service, for example, could now easily answer customers' queries, regardless of which department worked with the relevant data. Moreover, the firm began to take advantage of data-driven opportunities. Now that the data was all in one place, the marketing department deployed Azure Synapse Analytics to analyze customer behavior. Then, using Azure Machine Learning, they trained machine learning models to segment customers based on preferences and purchase history. Compared to manual segmentation processes, the new method allowed for quicker, more granular analysis. Correspondingly, the firm's marketing campaigns became more agile and personalized.To monitor future trends, the marketing department built a real-time dashboard visualizing customer purchase trends using Power BI. This data analytics tool integrates easily with Azure, as does Tableau. Subsequently, the marketing team could be more responsive to even the most minute changes in customer behavior.
Ultimately, the company achieved a 25% increase in repeat customer purchases, using segmentation to tailor email promotions based on past shopping behavior. It also improved first- contact resolution rates, decreasing the time non-customer service staff spent answering customer service questions. Finally, the company saw a major reduction in manual data management tasks and associated errors.