Table 1 uses a RACI matrix to discuss how different organizations might be engaged in data stewardship for customer data at a retail bank. The RACI matrix stands for the following:
- Responsible – The person who has delegated responsibility to manage an attribute. There may be multiple responsible parties for one attribute.
- Accountable – The person who has ultimate accountability for the data attribute. The accountable person may delegate the responsibility to manage an attribute to a responsible party. There should only be one accountable party.
- Consulted – The person or persons who are consulted via bi-directional communications.
- Informed – The person or persons who are kept informed via uni-directional communications.
Table 1: RACI matrix for stewardship of customer data at a retail bank
We discuss the responsibilities for each category of attributes below:
The bank created a customer information management (CIM) department reporting into operations. For the most part, CIM had overall accountability for stewardship of customer data within the master data management system. For the most part, finance was kept informed of the policy decisions around stewardship of customer data, which affected reports such as customer profitability.
This information included name, type of party, and social security number. The type of party was important because the bank was using incorrect matching rules when a party was wrongly classified as a person as opposed to an organization. Although CIM had overall accountability for identity information, operations had delegated responsibility to fix any data quality issues. For example, operations would run periodic checks with government sources to validate social security numbers. If operations discovered any discrepancies, then they would typically review a copy of the driver’s license or passport, or invite the customer to visit the branch. Compliance was consulted due to the implications associated with Know Your Customer (KYC) regulations. Marketing was also consulted because high-quality information was critical to the elimination of customer duplicates, which reduced wasteful spending on multiple marketing mailings to the same customer. Finally, credit risk was consulted because they needed to quantify the overall exposure to the same individual across multiple, fragmented records.
This information included date of birth and gender. CIM had overall accountability for these attributes. Compliance and marketing were both kept informed for the same reasons as mentioned above. In addition, compliance and marketing needed to know that information regarding date of birth was accurate so that the bank was not sending marketing mailings to minors. In addition, about 0.5 percent of the gender attributes were incorrect and the internet team was at the receiving end of complaints when a customer such as Mary Jane logged on to her online account and was greeted with “Hello, Mr. Jane.” In addition, the bank had a policy of driving low-value customer traffic to the internet as much as possible. However, customers had to provide a date of birth in order to be authenticated for online access. Inaccuracies in the date of birth caused a number of headaches for the internet department because they could not easily set the customer up for internet banking. Table 2 lays out the sample business benefits associated with an improvement in the quality of dates of birth.
Table 2: Business benefits associated with an improvement in the quality
of dates of birth for retail banking customers
This information included net worth, income, and segmentation. The marketing team had overall accountability for these attributes that were maintained only in the data warehouse, which was a target of the master data management system. CIM was kept informed of any changes because its data was a source for the data warehouse.
This information included relationships such as “spouse of,” “child of,” “parent of,” and “co-signer.” CIM had overall accountability to maintain this information. Operations had delegated responsibility to address any data quality issues. Marketing was a key stakeholder because of the potential for householding, or treating multiple individuals as part of one household. With the help of the insight from householding, marketing could tailor offers at the household level and eliminate multiple mailings to the same household. In addition, credit risk was able to quantify the overall exposure to a household.
This information included email address, mailing address, phone number, preferred mode of contact, and the do not contact indicator. There were several data quality issues. There are only 50 states in the United States, but the bank profiled its data and found that were actually 53 states. For example, the developers had no way to account for customers who lived in the state of New Hampshire (“NH”) but worked in the neighboring state of Massachusetts (“MA”), and they created a new state called “NHM” for tax purposes. CIM had overall accountability for this information. Customer service had delegated responsibility to fix any data quality issues. Marketing was a key stakeholder because any improvement in the quality of contact information had a direct correlation with campaign response rates. In addition, the channels department was a key stakeholder because it needed accurate phone numbers. Existing customers would reach out to the bank via the branch, call center, or internet to purchase a new product. Ever so often, the channels team felt they needed more information. However, if the phone number was inaccurate, it made it difficult to contact the customer, and increased the probability that the product sale would not be completed. Debt collections was also a key stakeholder for contact information, especially phone numbers. Finally, security and privacy was kept informed about CIM policies regarding the appropriate stewardship of customers’ privacy preferences.