Streamlining Customer Success: The Benefits of a Centralized Identity Management System

Traditionally, organizations manage user identities within each application or service, leading to customer data being siloed and losing the value that can be extracted by correlating customer behaviors across systems. However, having a centralized IMS is critical for customer success as it enables organizations to understand and target users better, set preferences across all products/channels with a single click, and utilize derived and aggregated data points better through AI and ML systems.?????????

At ADF, a data analytics-first organization that offers services through multiple product lines and channels, customer identity is crucial. Therefore, we needed a centralized IMS to extract value from customer behavior and data, serve them better, and evolve for the future.

We identified several key roles and responsibilities for our IMS, including:

• Being the source of truth for anything related to a customer across the entire organization.
• Storing not only customer identity but also preferences such as communication preferences and "Do Not Contact" information.
• Seamlessly cross-matching and identifying the same customer across multiple channels and products.
• Automatically building and maintaining a behavioral profile of each customer based on predefined algorithms to enhance our understanding of them.
• Providing a globally unique reference ID for each customer to be used by multiple systems across ADF; and
• Establishing a valid identity for a new or existing customer, thereby eliminating fraudulent activities.

?Key differences between a legacy system vs ADF's IMS

Building a centralized Identity Management System (IMS) does not mean moving all customer/identity-related data to a common database or service and being done with it. Such systems, often built a couple of decades ago, used to rely on a single match key, such as a government-issued ID like a Social Security Number (SSN), to identify customers and store all their data in one place. However, this approach has limitations, especially with the increase in mobile usage and the potential for errors like typos.

To address these limitations, ADF's IMS offers advanced functionalities designed to meet the demands of the current era. Here are some key differences between a legacy system and ADF's IMS:

• ADF's IMS offers progressive and fuzzy matching, allowing for more accurate customer identification across multiple channels and products.
• The system is agile and adaptive, supporting the use of multiple matching algorithms simultaneously and allowing for updates to the matching algorithm as needed.
• ADF's IMS collects not just basic identity data but also customer behavior data, maintaining an immutable, chronological log of all critical events to help make sense of the data.
• The system includes a graph of all customer touchpoints and relations, providing deeper insights into customer behavior and journey and detecting potential fraud or suspicious activity.
• Capturing more data points, in addition to plain customer data, can also improve fraud mitigation, allowing for more informed decisions in identifying fraudulent customers based on the level of matching between key identity fields and their similarities.

?Exact vs. Fuzzy Matching

IMS aims to provide a central source for customer identities with the ability to accurately match the same customer across multiple channels or products. Traditionally, this is done using a government-issued ID like a Social Security Number (SSN). However, with the increase in mobile usage, relying on a single field is prone to errors such as typos. To mitigate these errors, IMS uses a combination of input data fields and fuzzy matching algorithms. These algorithms recognize names and addresses as specialized data types rather than plain strings, making them less prone to errors.

Progressive + Multiple Customer Matching Algorithms

Modern systems are agile and adaptive and should embrace changes. IMS is no exception. The current matching algorithm may need to be revised as more data is gathered, and a better approach is identified. The system should support updating the matching algorithm without impacting client systems that rely on the reference IDs provided by IMS. This will ensure the system remains relevant to current needs.

Additionally, the IMS system should be capable of supporting multiple matching algorithms simultaneously. For example, some product lines may have limited information about a customer, such as only a name and address, while others may have more information, such as an SSN. The IMS system should be able to accommodate different algorithms depending on the available data and the needs of the client system.

Behavioral Profile of Customers

IMS should collect more than just basic identity data; it should also gather customer behavior data. The chronological log of all critical events in a specific channel or touchpoint is as important, if not more important, than just knowing the customer's identity. IMS should maintain an immutable, append-only chronological log of all events to help make sense of the data. Collecting the lowest granular level of data possible, it will allow for slicing and dicing to generate advanced aggregated/derived variables for better AI/ML-based analysis. These variables can range from a few hundred to tens of thousands and can be computed both in real-time and in the background batch as needed for performance.

Graph of All Customer Touchpoints and Relations

Having a relationship graph of touchpoints in IMS can give deeper insights into customer behavior and journey, which can inform better decision-making. Additionally, this graph can also be useful in detecting potential fraud or suspicious activity by flagging unusual patterns in customer behavior. The IMS can be a powerful tool in providing a holistic view of the customer’s journey and their interactions with the company.

Fraud Mitigation

Capturing more data points, in addition to plain customer data, can also improve fraud mitigation. Fraud often occurs when a portion of a customer's identity data, such as their Social Security Number (SSN) or address, is stolen and combined with a malicious actor's bank account information. By incorporating bank account data into the IMS, we can detect if the same bank account is linked to multiple customers or if the same SSN is associated with unrelated contacts.

Furthermore, we can make more informed decisions in identifying fraudulent customers based on the level of matching between key identity fields and their similarities.

Modern organizations need more from their identity management systems than simply centralizing data.        

ADF's IMS offers fuzzy matching, multiple customer matching algorithms, customer behavior data collection, touchpoint relationship graphs, and fraud mitigation capabilities. These advanced features allow us to gain deeper insights, mitigate fraud, and make informed decisions based on key identity fields.

?