What IT leaders need to know about customer data platforms
Customer data platforms offer a lot, but many IT decision-makers remain skeptical -- there's still plenty of confusion over what they are. Find out what companies get with them.
Marketers have been excited for the past several years by customer data platforms -- packaged software that builds unified, persistent and shareable customer databases. The appeal of CDPs is that they promise to solve a pressing problem: the inability of marketers to assemble all data for each customer into the single profile needed to deliver consistent, personalized treatments.
The problem only grows as companies add more marketing systems to manage more interaction points, resulting in more data spread across more disconnected silos. Conventional technologies such as data warehouses and data lakes are not a solution because marketing systems and non-technical users cannot easily access their contents. Existing marketing systems, such as CRM, marketing automation or data management platforms (DMPs), cannot assemble enough data from other sources to create complete profiles. The failure of these alternatives created the opening for customer data platforms to emerge.
Many in the marketing community are still confused about what CDPs are and what they should include. But it's now widely accepted that companies need a unified, persistent customer database that feeds other marketing systems. Enterprise marketing platform vendors Adobe, Salesforce, Oracle, SAP and Microsoft have all released their own customer data platform offerings.
Yet even as CDP has become established in the marketing world, it has shifted away from marketing and into the core corporate technology stack. This is driven by increased interest at the corporate level in managing customer data. This is a product of trends, including new privacy regulations, which make closer control over customer data a corporate priority; digital transformation programs, which often rely on company-wide access to unified customer data; and new standards for customer experience, which extend beyond marketing to encompass all customer-facing departments. Like marketers before them, corporate IT groups find their existing toolkit cannot meet this need and CDP promises to fill the gap.
IT staff are properly skeptical. Hype is an inevitable part of the technology product cycle and customer data platforms have attracted their share. Then again, marketers have deployed CDPs successfully for years, so it is better proven than most new products. Still, many IT groups look at CDP and feel they could build something similar by themselves. In some cases, this closes the door to any consideration of CDP. But in most, it starts a usefully pragmatic discussion of how best to meet the company's needs, with both CDP and custom development as options to consider.
The best choice depends on the resources and requirements of each business. Here is what IT leaders should know about customer data platforms.
What's in a packaged CDP
Most CDPs use off-the-shelf data management technologies, which are equally available to IT departments building their own solution. Some CDPs include proprietary tools for identity management but others incorporate commercial software or rely on the user to apply customer identifiers outside of the CDP. But all CDPs provide prebuilt process flows for data ingestion, storage, profile building and data sharing.
From a "build vs. buy" perspective, avoiding the need to reinvent these flows is a major advantage of buying a packaged CDP. The value of this depends largely on the IT organization's expertise in dealing with the peculiarities of customer data. Customer identity management and privacy regulation are common gaps. A packaged CDP also has prebuilt connectors for common data sources and marketing systems, avoiding the substantial effort of building custom connectors. Even an IT organization with deep experience with all aspects of customer data may find it more effective to use a packaged CDP than to create a new one. Among other things, buying a CDP frees scarce development resources for other projects where no commercial alternative is available.
How CDP relates to existing data warehouses and data lakes
CDP functions overlap with data lakes and warehouses, serving both to collect data (like a lake) and to refine it into useful structures (like a warehouse). But CDPs don't need to replace those systems; rather, they can use them as sources. This is a common approach that avoids the duplicate effort of extracting and storing data from source systems and running data preparation processes. The value of the CDP in these situations is that it complements the existing resources by adding new capabilities. Beyond identity and privacy management, these are likely to include extracting structured data from unstructured formats, adding derived or summary variables to customer profiles, self-service interfaces for users to create extracts and customer segments and converting profiles into formats that external systems need.
How CDP differs from CRM, marketing automation and DMP
Many marketing departments have tried to use existing marketing systems to create unified customer profiles. They invariably find those systems are not suitable for the task, which is why a separate CDP is needed. CRM systems are built with highly normalized data structures to support real-time interactions with call center or sales agents. Marketing automation systems use simple data models designed primarily to manage email lists. DMPs use a huge data table with one row per ID to enable highly efficient queries against audience attributes. Each design is optimized for the system's primary role but lacks the CDP's ability to ingest all kinds of data, retain all original details, include any data in profiles, and share the profiles with external systems. As with data lakes and warehouses, the CDP complements marketing systems rather than replaces them.
CDP and master data management
master data management systems cover many domains that are outside of a customer data platform, such as products and locations. The customer domain in MDM is similar to customer identity management in a CDP, and companies with a mature MDM customer domain may use it to populate customer IDs in a CDP. Even when this happens, the CDP likely stores many kinds of data that the MDM system does not. This data includes transaction details, semi-structured data such as Web logs and unstructured data such as text and audio. An MDM might apply tags or metadata to some of those.
CDP scope
All CDPs build unified, sharable customer profiles. Some CDP vendors provide additional capabilities including analytics, predictive modeling, personalized message selection, campaign management, cross-channel orchestration and message delivery such as email or website banners. The extended features are most appealing to marketing departments that want a single system to fill many needs. IT departments looking for a company-wide data asset are unlikely to want those other functions in their CDP. CDPs that focus exclusively on profile building are more likely to have advanced features for data management such as specialized data structures, high scalability, precise access controls, derived variable management, sophisticated APIs and large numbers of prebuilt connectors. These are most likely to be of interest to corporate IT buyers.
Self-service interface
One of the primary reasons that marketers buy CDPs is to gain self-service capabilities that reduce their dependence on technical resources. Business users through an organization have similar priorities, so self-service is likely to be important to corporate-level CDPs as well. IT departments making a build vs. buy decision should consider the cost to create and maintain such features. They may also be able to apply existing third-party tools rather than building their own. IT departments should assess the security and access controls of a CDP's self-service features, recognizing that capabilities originally designed for use in the marketing department may not be robust enough to support enterprise-wide deployments.
Consideration for customer data platforms
IT departments need to consider the details before making a decision about buying a CDP. But it's also important to look at the broader picture. Unified, sharable customer profiles are essential for success in today's business environment. Few companies have them in place and customer data platforms provide a potentially faster, cheaper, more flexible alternative to custom-built systems as a way to fill this gap.
About the Author
David Raab is founder and CEO of the CDP Institute. A long-time industry consultant specializing in evaluation and selection of marketing technology, Raab coined the term "customer data platform" in 2013.
