Interview with David Raab of the Customer Data Platform Institute: Four Questions

August 09, 2017


A top hot topic for enterprise marketers is how to truly understand and activate massive amounts of omni-channel customer data and leverage it to optimize marketing efforts. Today's customers are expecting a seamless customer experience across channels. From clicking an email to searching online, to checking the closest store on their mobile, to going into the store -- all customers and their journeys are different. 

With all of this data and demand, enterprise companies are looking to solutions like AgilOne - to cleanse, dedupe, and stitch together all customer data sources into a single customer profile that gives insights enables action.  

AgilOne sat down for an interview with David Raab, the founder of the Customer Data Platform Institute, who has been closely tracking the evolution of the Customer Data Platform (CDP) market. We got the chance to ask him what requirements enterprise retailers should look for in a CDP, their role of machine learning in CDPs, and what direction he sees the category evolving.

What capabilities should an enterprise retailer with lots of stores look for in a CDP?  Large retailers often have a mix of technologies in different locations, so they need a CDP that’s especially good at standardizing disparate inputs.  This means making it easy to map different input formats to a standard data model and automatically identifying inputs that are not in the expected format. Retailers also need to deal with variations in product data, such as different SKUs that describe the same product in different sizes or packaging. 

Retailers may also deal with complex purchase records that split their data among several levels such as order headers, line items, tax charges, payment types, and so on.  Geographic features to understand store trading areas and distances to competitors are also important for retail. Finally, for retailers, the CDP may need to look up external data such as weather conditions in a particular location at the time of purchase.  In some cases, this requires real time access to help select appropriate customer treatments. Enterprise retailers should consider their specific use cases when evaluating a CDP as not all CDPs provide these capabilities.

CDPs traditionally focus on known first party data. How do you see CDPs evolving to integrate further into the "unknown" world of AdTech?  This is happening already.  It’s probably easier for most CDPs to support anonymous data than it is for most DMPs to add known customer identities, since the known data comes with more complex privacy rules.  Similarly, CDPs are built to handle multi-level data structures, so it’s easier for them to deploy the single table structure of a cookie-plus-attributes DMP than it is for a DMP to move from single- to multi-table data.  But adtech does have its own requirements, such as very fast data access for real time bidding and integration with other adtech systems and process flows.  So it will certainly take some work for CDPs to fully replace DMPs in particular applications.

What is the role of machine learning within a customer data platform?  The most common use of machine learning is for predictive modeling, such as assigning customers to segments or selecting the best customers for a promotion.  Machine learning can also help with other tasks such as interpreting and standardizing data inputs, recommending the best product or content to offer each individual, or even recommending new marketing programs to deploy.  From a strategic standpoint, machine learning removes manual labor and decision making, enabling marketers to run more sophisticated programs – for example, with more segments or more content variations – than is otherwise possible.  This lets them make better use of the massive data volumes assembled in the CDP.

How do you anticipate the adoption curve for CDPs will change over the next year and beyond? What factors will drive this change?  We’re seeing a sharp uptick in interest in CDPs – for example, here’s a Google Trends chart. 


Adoption is also accelerating according to our data.  I expect this to continue.  Factors driving the growth include greater awareness and understanding of what CDPs offer, more martech expertise in marketing organizations, allowing them to take advantage of CDPs, and of course greater understanding among senior management that unified customer data is essential to meeting customer expectations. It’s a perfect storm but, in this case, one that leads to good results.


Looking to get your own CDP?

CDPcover-blog.pngIn partnership with David Raab and the Customer Data Platform Institute, we present you with an executive guide to enterprise CDPs.

This white paper will walk you through the top requirements you should consider when picking the right tool for your enterprise customer data needs.