a group of people have separated into a pie chart
PHOTO: Shutterstock

Forrester researchers in their Digital Experience Platform Trends report (fee required) last year said the software is designed to manage, deliver and optimize digital experiences consistently across every phase of the customer life cycle. One common entrant into the Digital Experience Platform (DXP) ecosystem is Customer Data Platforms (CDP), a “marketer-managed system that creates a persistent, unified customer database that is accessible to other systems,” according to the CDP Institute

The rise in popularity of CDPs over the past year is undeniable. The CDP Institute is widely cited in literature about digital customer experience software, and Gartner, in its July 25 release of the Hype Cycle for Digital Marketing and Advertising, 2018 (fee required) has CDPs at the “peak of inflated expectations.” Beyond the hype, though, marketers need to know where, how and why these systems play into the digital experience platform (DXP) environment and what integration challenges lay ahead before they invest in the technology. 

To help readers find the answers they seek, we’ve tapped analysts and those involved with CDP implementations to help discover some key considerations for implementing CDPs into a DXP environment.

Website Integrations Are Common

CDPs on the data capture side in a DXP most often integrate with an organization’s website, either directly or via the systems that help manage marketing tags, according to David Raab, founder of the CDP Institute, which provides analysis of the CDP market. 

CDPs can also be found helping organizations manage data within their mobile apps and with email. “Some include offline data sources such as point of sale, call center, and order processing,” added Raab, who breaks down CDP vendor capabilities in his CDP Vendor Comparison report (download required). “On the messaging side, it’s pretty much the same systems plus often DMPs for advertising.”

Related Article: Understanding the CDP Landscape: Which Kind is Right for You

Most Common Integrations, Outbound Marketing 

Ryan Willette, global VP of customer success for AgilOne, which offers a CDP platform, said he sees CDPs most commonly integrating with outbound marketing technology (SMS, direct mail, email), digital advertising platforms (Facebook/Instagram, DMPs, retargeting systems), and customer experience tools (clienteling apps, call center systems, website personalization tools). “For brands first implementing a CDP, connecting to outbound marketing systems is low hanging fruit, and direct mail, SMS, mobile push, email are all the first types of use cases where brands get value with a CDP,” Willette added.  According to him, once a brand is up and running with outbound marketing, marketers will typically expand to acquisition channels in advertising and media, where they can then use a CDP for smarter ad suppression, better lookalike modeling and omnichannel targeting. 

Marketing teams need to understand the scope and investment for each type of use case before they begin to roll it out, Willette shared.  

Related Article: Why Your Customer Data Platform Is a Failure

Analytics and Insights Drive Practitioner Desires

Digital Clarity Group researchers in their Digital Experience Platforms: Buyer Trends, Preferences and Strategies report (login required) called customer data management technologies “the center of gravity for many organizations on the DX path.” And CRM is the epicenter, research found. 

DX teams often double down on CRM, their key data repository, as well as ERP, data lakes, CDPs, data warehouses, data integration, data quality software and master data management. “The most important thing for customers was analytics and the ability to be able to kind of mine through data and be able to get some insights about customer intentions or buying desires and feedback and all of those things,” said Scott Liewehr, founder and CEO of Digital Clarity Group.

Integrations with Optimization Engines, Portals

Steve Zisk, senior product marketing manager at CDP provider RedPoint Global, sees CDPs integrating with three kinds of elements in a DXP:

  • Decision components like optimization engines, which need customer data to analyze/understand what the customer is doing and how to optimize the digital experience. The optimization engine may call a CDP for a customer profile or the CDP may push a profile based on observing customer activity.
  • Experience touchpoints like user forms, portals, chatbots, etc., where timely customer information is used to personalize DX and provide richer information. Here, Zisk said, the touchpoint software will generally call the CDP or use a cached version of a profile to understand preferences and objectives and tailor a response.
  • Human interactions, where agents need accurate information about the customer/client to respond to requests for information or action. For human agents, they may be using CRM software, clienteling apps, or chat/interaction software that provides customer details, and potentially customer analytics like CLV, product offers, or customer intent, pulled from a CDP “just-in-time” to enrich the interaction.

Related Article: Customer Data Platforms: A Contrarian's View

Data-as-a-Service Connections to Third-Party Sources

For a CDP to successfully be woven into the fabric of a DXP environment, data-as-a-service connectivity to third-party sources will be critical, according to Ajay Khanna, VP of marketing at Reltio, which provides a self-learning data platform. “Customer information must be related to other data entities, such as product, store, location and site,” Khanna added. “If your CDP does not support multiple domains, then it should integrate with a multi-domain data management platform that also includes graph technology to uncover these relationships.” 

Customer information must be infused with relevant insights and intelligent recommendations to help with business decisions. “CDPs must seamlessly provision customer data to today’s advanced analytics and machine learning technologies,” Khanna said.

Related Article: Customer Data Platforms: The Truth Behind the Hype

Transforming Data in Logical Manner

With CDPs comes questions of data integration. How well do they play in the DXP environment that’s composed of content-producing, customer analytics and transactional systems such as CRM, digital asset management (DAM), web content management (WCM) and ecommerce? 

Storing data in a CDP that fits the formats of downstream applications is always a challenge, as well as retrieving all needed data from upstream environments and applications that is manipulated and imported properly for a CDP, said Jonathan Moran, director of product marketing at Earnix. Organizations must be mindful, Moran said, of how they may have to pull data out of the source systems and place it into a data warehouse (extract, transform and load process). They need to be ready to move and transform data from upstream applications in a logical manner. They also need to make sure data imports and exports are clean and complete and work for other applications in the ecosystem.

Data Mapping Challenges

CDPs are designed to make integration easier but that doesn’t mean it’s automatic, according to CDP Institute’s Raab. “Even though many CDPs can ingest data with little predefinition of its meaning,” he added, “there is still a later step where the ingested contents do need to be mapped to specific fields.” Marketers and CX professionals can also expect a need to define validation rules and to classify ingested items into useful categories for analysis, Raab added; e.g., identify the product category associated with a SKU. “Still, I’d say the main integration challenges will be dealing with systems that don’t have an API to exchange data with other systems,” Raab said. “In those cases, you’re likely to need to feed the CDP a batch file. It can also be time-consuming to write a new connector to a system that has a poorly documented API or API with limited functions.”

Related Article: What Can You Do With a Customer Data Platform?

Unstructured and Semi-Structured Data and Extracting Meaning

Identifying customers across different sources is a different kind of integration challenge — but with big implications when it comes to CDPs, according to Raab. Marketers will find it challenging identifying anonymous website visitors and connecting their profiles with other profiles related to the same person. “Probably the biggest challenge is dealing with unstructured and semi-structured data,” Raab said. “Most CDPs are equipped to ingest and process this but it can be difficult to extract meaning. This is where tags within the semi-structure data must be classified and where parsing or natural language processing must be applied to unstructured data.”

Predictive modeling or machine learning is often essential to making data actionable by adding features such as recommended products, likelihood scores for actions (e.g. churn risk score), and content classifications. 

Another consideration according to Raab is real time processing, which is often needed for specific applications such as recommending messages to display on the website or to a call center agent. “This has many nuances related to speed of loading new data, of moving new data through the transformation processes needed to make it available, and of responding to data requests. CDPs vary widely in this, so it’s especially important to understand their capabilities in detail,” Raab said.

Related Article: How to Build the Right Team for Cross-Channel CRM and Marketing

Prioritizing Data Key Part of Integration

The role of the CDP is to consolidate customer information, but sometimes just identifying which data sources to integrate can be a challenge, according to Andy Zimmerman, chief marketing officer of Evergage, which offers a CDP. “But assuming you’re able to identify and gain access to the appropriate systems in your organization,” Zimmerman said, “you then have to prioritize the data. In other words, what data is critical to your business objectives and what can be ignored? With a clear understanding of the systems and data hierarchy, the integrations process itself should be relatively straightforward.”

CDPs Involve More Than Marketers

Although marketers are the ones who will get the most use of a CDP, they are reliant on other teams, Willette said. For example, deploying a CDP webtag might involve a website service provider, or the IT team. “Resources needed to implement a CDP often get pulled into other projects,” Willette said, “and urgent fire drills always take priority.” 

Organizations must also be mindful of creating alignment across different teams that have different views of the same data when working with CDPs. “There is often no consensus within an organization about where to get data from, or which data is the most accurate,” Willette said. “Brands should try to build a partnership across teams to create data consistency.” 

Siloed Customer Data Will Remain

Recognize that CDPs don’t fix the broader problem of siloed customer data, Zisk said. Rather, they make it possible to get an up-to-date, accurate and consistent customer profile, but they don’t magically impose that profile across the data layers in DXP applications. “That means the whole set of governance considerations — privacy, security, and compliance, along with semantics, portability and reconciliation — must be addressed,” Zisk added.

Understand how to integrate CDP customer records back into the source systems, beyond just using the CDP to enhance interactions. “In an ideal situation, every source system would be able to use the CDP as its source and target, but most real-world DXP apps operate as systems-of-record for their domains, or use underlying databases to serve that role,” Zisk added. DXP practitioners will have to define processes for interacting with the CDP to update, correct and enhance customer data in the source systems.

Agility to Keep Pace with Market Changes

“New social apps, new devices and new use cases are the norm, not the exception, and the CDP must support fast changes to the DXP stack,” Zisk said. “This requires that tools for connectivity, validation, processing and analytics are both easy to use and easy to extend.” Teams of CX, IT, and brand practitioners, Zisk added, “must be allowed to try new and different components of digital experience without the CDP getting in the way of innovation.”