Customer Data Platforms: A New Tool to Unify Customer Data
Customer Relationship Management is a grandiose term, but systems bearing that label have largely been limited to helping salespeople and service agents manage one-on-one interactions. That has recently begun to change as CRM is used to describe all systems working with identified individuals. The change happened almost by default: as new interaction channels are added–think mobile, video, voice, Internet of Things, and smart TVs, among others–no umbrella term has appeared to cover them all. CRM, which already sounded as if it should include all customer data and interactions, has been drafted into service.
This shift in meaning is mostly harmless except when organizations expect their old CRM systems to provide the much broader set of functions that new meaning implies. A CRM that included all customer data and interactions would be at the center of an organization’s customer-facing processes, executing or at least feeding Web interactions, e-commerce, mobile apps, email, marketing automation, display advertising, loyalty programs, and retail point of sale. But today’s CRM systems simply can’t do this because they were designed for the much narrower needs of sales and service agents. Rather than being the hub that connects all these systems, CRM is on edge—just one spoke in the wheel.
What does sit at the center? In some companies, nothing at all: each customer-facing system is separate, working only with the data it generates internally. But this fragmentation is increasingly unacceptable to customers who have been trained by companies like Amazon and Netflix to expect consistent, personalized service across all channels. Pressure from those customers has led many businesses to connect their systems with a unified, shared customer database. Some have built this database for themselves, but it’s a daunting prospect: imagine your most complicated data warehouse project and then add twice as many source systems, dirtier data, constant schema changes, endlessly shifting use cases, and privacy regulators second-guessing your every move. Many other companies have turned to a new class of packaged software designed specifically to unify customer data: Customer Data Platforms (CDPs).
CDPs are defined by the CDPI Institute as “packaged software that builds a unified, persistent customer database that can be shared with other systems.” They are more structured than a data lake, more flexible than a data warehouse, keep data longer than a Data Management Platform, and hold more detail than a CRM. In an architecture diagram, their position as connectors between other systems makes them resemble an enterprise data bus or integration platform. But those systems only move data from a source to a destination, while the CDP keeps its own copy of the data it ingests. This matters because source systems often discard or archive data such as past addresses or details of completed transactions. That’s fine for operational processing but not for personalization, which needs historical data to identify trends, predict future behaviors, and track individuals over time. Creating a separate, comprehensive database is thus a fundamental requirement of CDPs.
Other core CDP features include retaining full detail of inputs; building a unified profile for each customer; and making all data accessible to other systems. Technically, most CDPs use NoSQL data stores such as Hadoop or MongoDB, enabling them to store both structured and unstructured data such as Web browser logs. Nearly all CDPs are cloud-based although about half also allow on-premises or private cloud deployment. Some have large libraries of connectors to common CRM, e-commerce, Web analytics, business intelligence, reporting, and other systems. Many offer analytical capabilities such as predictive model building and personalization features such as message selection.
The sheer variety of CDP configurations can be confusing. It’s best to think of CDPs as platforms that build a unified customer database. All CDPs do this. Some CDP vendors supplement their platform with applications for analytics and personalization; other vendors rely on third party developers for those apps. It’s no different from the way that Apple and Android build some apps for their smartphone platforms while also supporting apps built by others.
CDPs are not the only packaged systems that build a unified, persistent customer database. Some CRM, Web personalization, email, marketing automation, and other customer-facing systems also have this capability. In many cases, however, those systems only make their database available to their own components: it is like a smartphone that does not allow third-party apps. Some businesses are happy with this, finding the single vendor’s products adequate and preferring to avoid the cost of integrating multiple vendors’ products. Other organizations prefer an open approach that lets them use whichever apps they prefer.
The applications built by CDP vendors have focused largely on marketing uses. This is largely because marketing departments have been the primary CDP buyers. But unified customer data can help other departments: sales and service, obviously, but also compliance, risk, anti-fraud, and operations. Corporate IT groups are increasingly exploring CDPs as a smarter alternative for meeting the needs of all these groups.
There are more than 100 CDP systems on the market today. In addition to scope, they vary in the quality of their features, industries and regions served, size of typical clients, supporting services, company size, and funding. While these differences make it harder to pick the right CDP, they shouldn’t obscure the fundamental value shared by all CDPs: making it easier to connect all systems with unified, consistent customer data. As the importance of these connections increases, so will the role of the CDP.