David Raab, CDP pioneer, discusses AI’s impact on CDPs, debunks myths, and predicts future trends. Learn about hyper-personalization and real-world success.
Customer Data Platforms (CDPs) have come a long way since their inception, but as David Raab, Founder of the CDP Institute, points out, they may be coming full circle. Initially bundled within application systems and then evolving into standalone databases, today’s CDPs are again integrating into larger ecosystems, raising new questions about their role in an AI-driven world.
In this exclusive conversation, Raab cuts through the noise, debunking myths about CDPs being “too complex,” explaining why AI-driven personalization won’t replace but reinforce them, and predicting a future where hyper-personalization meets evolving privacy norms. Plus, he shares a real-world success story of how McDonald’s leveraged a CDP to drive customer engagement and boost sales.
What’s next for CDPs, MarTech, and AI? Dive into our interview with the man who coined the term “CDP” to find out.
How has the definition of a Customer Data Platform evolved in the last decade?
It has somewhat come full circle. When CDPs first appeared, they were part of application systems such as campaign managers and analytical packages. What was new was that those systems were building a unified customer database instead of connecting to an external database. The focus quickly shifted to the customer database, and my definition shifted accordingly. Today, we see fa ew CDPs that only build a database. Most are part of a larger system, usually for analytics or customer interaction.
Although I have not formally revised the CDP Institute’s definition of a CDP, some change is needed. Still, the function of a CDP remains the same: to build unified, shareable customer profiles. This doesn’t change regardless of whether the CDP is a stand-alone system or part of something larger. What changes is the requirement (in the CDP Institute definition) for the CDP to create its database? This no longer applies when the CDP uses a database it does not control.
What are the most common misconceptions businesses hold about CDPs?
I’d like to say that CDPs are complicated and expensive, but that’s true. Gathering and putting all your customer data in a useful format is difficult. Anyone who tells you it’s possible to skip all that work is probably assuming it’s already been done in your company by some other system—which is rarely the case. One actual misconception is that it’s somehow less efficient for the CDP to build its database than to use a data warehouse.
This may be true sometimes, but most data warehouses were designed for other functions and would need extensive development to add CDP capabilities. Perhaps even more important, data warehouse technology typically lacks key features such as real-time updates and access, which are best handled in a system designed for those capabilities. There are costs to moving data from one system to another, but those are fairly small compared to running complicated CDP-type queries within a cloud data warehouse.
Also Read: AI, Storytelling and Trends: Ruchika Batra on Content Marketing
How will AI-driven personalization reshape the role of CDPs in the next five years?
The CDP will likely be one source and the primary source of customer data to support AI-driven personalization. This makes the CDP even more important than it has been in the past. However, the CDP was always intended to be the main source of customer data, so this isn’t a new role. The core requirements for accurate, current data accessible in real-time don’t change just because an AI is now consuming the data.
Are we witnessing a convergence of CDPs, CDWs, and other data management tools, and if so, where is it leading?
To the extent that the CDP is a system for building customer profiles, we now see profiles built in many places, both by data management tools and customer engagement systems. The central role of customer data in business today makes it desirable for a vendor to provide that data because it is very difficult to replace a system when so many other systems rely on it. Of course, any given company will ideally have a single customer data system or, more likely, only a few.
This means that many vendors adding CDP capabilities to their products will find they get only limited use. We may see a rise (or a continuation) of departmental CDPs at organizations where individual departments are large enough to fund such projects. This will give those departments control over the system, making it easier to meet their needs. Always remember that the CDP is a specialized system, so other data management systems must add CDP features to meet CDP requirements.
How can companies reconcile hyper-personalization with escalating privacy concerns?
Companies need to earn the customer’s trust by asking permission to use their data and showing that they will use the data in ways that are in the customer’s interest and will protect it from misuse. Hyper-personalization will not be perceived as a privacy violation if it is useful. The opposite also applies: hyper-personalization that isn’t an obvious benefit or uses data the customer didn’t knowingly provide is often seen as “creepy”.
This is a rapidly evolving area: new norms will likely emerge around how much data companies are expected to have about each customer and how much personalization they provide. As data-driven personalization becomes more widespread, many customers will probably be less concerned about what might today feel like a privacy violation.
Also Read: From Hype to Must-Have: How Products Are Stealing the Spotlight
What are businesses’ primary challenges when implementing a CDP, and can you provide a success story?
Our research consistently shows that the biggest implementation challenges are organizational, not technical. Companies that struggle with their CDP often fail to define their use cases and business requirements, secure cooperation across departments, and adequately train their users. Data access and data quality are the most common technical issues. If a CDP fails, it’s most likely because the company didn’t understand its needs and bought the wrong systems.
The CDP Institute has a library full of success stories. I don’t like picking just one, so I’ll take the most recent. This is a study of how McDonald’s (West and South India) used their CDP to understand customers personally and make personalized, real-time offers based on the customer’s current activity. The benefits included a better understanding of campaign results, labor savings from campaign automation, a 33% increase in omnichannel customers, and a 40% increase in the use of delivery services.
Beyond CDPs, what is the most significant MarTech breakthrough you foresee in the next five years?
It’s hard to look beyond AI, which has the potential to replace the existing martech framework of campaigns with true one-to-one interactions: imagine a system that picks from a menu of options and picks the right one in each situation, creating a tailored version for each customer. Also, sticking with AI, an interesting case is made that it would let companies replace existing SaaS software with zillions of custom applications. I’m not entirely convinced, but it would be a huge breakthrough if it happened.
But if we set aside AI, I think the next breakthrough will be in omnichannel customer management, which will do away with the current divisions between web, email, text, video, apps, CTV, and other channels. This would work best in a world with fewer privacy-related barriers to identifying customers within and across channels. I believe privacy will diminish (although I would rather it did not). AI will play some role in converting messages from one format to another, but that’s less important than companies having easy, central access to customer data and decisioning engines, which can pick the best message and channel in each situation. Replacing the current channel-based fragmentation with a true omnichannel, unified customer experience would fundamentally reorganize how marketing, and thus martech, works.
