The Customer Management Portal was built to make customer information more accessible to the business’s customer experience team.

Problem Statement

Prior to the creation of this report, the customer service (CX) team could not access customer information on the fly. This presented a challenge as customers calling in with complaints could not be attended to with real time data. As a stop-gap, the in-house Database administration team had created data repositories for the team, however these were sometimes difficult for non-technical staff on the CX team to decipher.

Proposed Solution

The CX team requested a dynamic report containing the following customer information:

1. Customer demographic data including Age, Gender, Contact Information to help agents ID customers in real time when dealing with complaints and requests;
2. Customer account information including Account ID, Fund ID, Account Status and Account Balance to equip CX agents with the right information when trying to understand and address customer concerns;
3. Customer account history including payments and withdrawals to verify progress of payment requests;
4. Customers’ previous interaction history with the CX team to identify prior issues and resolution progress;
5. Historical Application records - records of all withdrawal applications highlighting their most recent application stage.
6. Churn Analytics - The team would go on to use this information to monitor churn rates and manage client onboarding for incoming customers.
7. Customer clusters - The team also specified pre-determined criteria for classifying customers into industry relevant clusters such as HNI, millennial, centennial, et al.

Tool Used: Microsoft Power BI

Data Sourcing

The data was sourced from multiple Microsoft SQL Server databases (DB). Due to the volume of data generated for > 1million customers, bringing in the data as they appeared in the DBs would have made the power BI dataset too large and negatively impacted the performance of the data models. Dataquery mode could not be used in this scenario as the DBs in question served the organization’s live channels and applications. Therefore, Import mode was used.

Dimensional data like customer demographics and account info were imported with priority given to columns relevant to the team’s need whereas for fact data such as interactions and transactions, a more nuanced approach was required.