Expert Interview with David Borean about Master Data Management Solutions
What is the next biggest step organizations need to take in their customer data management programs?
Master Data Management (MDM) has become mainstream and accepted as the set of technologies and processes required to get a trusted single view of the customer. This is then used in real-time and analytics processes.
The single view of the customer is somewhat myopic, however. It contains kernel information about the customer managed within the MDM hub, such as name, address, etc, and references to account and transactional data stored in an ODS or maintained in other systems. The one type of information it is missing is interaction data such as call center, dot-com and social interactions. Evolving the customer data management program to include interaction information is the next biggest step.
Why is managing interaction information so important?
The number of ways an individual can interact with a company is drastically higher than a decade ago. You have a number of channels you can go through to make a purchase or for customer care. Examples include by phone, dot-com, social platforms, webchat, email and mobile.
Customers now expect a seamless transition from one channel to the next. For example, if you are on a webchat seeking help and the agent is unable to help you and directs you to a 1-800 number, then why shouldn’t your call be intelligently routed and prioritized with a call center agent that has the ability to read the webchat first?
Likewise, a company can market to you through many different channels. You can get a promotion through email, mobile, social or websites. But these tend to be disjointed and not tied back to the customer, so they are anonymous users being marketed to. Customers expect personalized and targeted promotions.
Omni-channel service and marketing is possible only with a Customer 360 that contains all the customer interactions and is current.
What does this mean for existing MDM technologies?
I don’t think it means that much for existing MDM technologies. They weren’t built to handle this type of requirement, much like they weren’t built to store and manage transactional data. Instead, there needs to be something that complements and integrates with the MDM hub to get a single view of customer interactions. In other words, a “Customer 360” hub.
It needs to handle a large scale of unstructured data. A large company will have millions or tens of millions of interactions each month, and it is not good enough to record the fact that there was an interaction; the full details and content of the interaction need to be stored as well. Real-time business processes require very fast retrievals of the interaction data for decision making, so it is important that it is all centralized.
It also needs to be able to match interaction data to customer records, which is very different than matching customer records to customer records. For example, picture a webchat where the customer entered their name, a subject and then the chat text with the web chat agent. If there is no mention of an account or transaction number, then you have to apply “contextual matching” techniques that use the unstructured text of the chat during the matching process. This is very different than what MDM hubs do today.
What are the other features of a Customer 360 hub?
Apart from having to manage a big scale of interaction data, it needs to be MDM-aware. The kernel of a Customer 360 record must come from MDM as that is the trusted source. Interaction data and any other type of data connects into it. So it must be able to react to MDM actions such as adding, updating and de-duplicating customers. It cannot be an island, so it must support integration, visualization and support for real-time analytics.
Also, it would be a shame to have all of that customer data centralized and in a clean trusted state and not be able to search on it. We’ve come to expect “Google-like” searching now, and so text searching on any and all attributes of a customer is important.
Last and not least, it needs to support governance. So this includes gathering and reporting on metrics such as data change trends, data quality trends and consumer usage.
Does Hadoop come into play, given that you mention “big scale” of data and unstructured data?
Yes it does. And for more reasons than just the technology infrastructure. The Hadoop vendors like Cloudera and Hortonworks have had a strong focus on Hadoop for the enterprise and positioning Hadoop as the “Enterprise Data Hub.” Customer data for many organizations is key to real-time business processes and analytics, and therefore it is difficult for me to think of an Enterprise Data Hub without a trusted Customer 360 hub within it.