Media CoverageMay 16, 2019

Lessons learned from CDP implementations

CMSWire's Dom Nicastro recently spoke with mParticle's Head of Product, Craig Kelly, to better undestand lessons he learned when implementing a Customer Data Platform.

Forrester predicts that more than 70% of B2B marketers will invest in a customer data platform (CDP) this year (subscription required). With data management a top 5 weakness for 80% of marketers, they’ll be faced with a build or buy decision and likely will take the latter path because of pressing needs, lack of internal skills and a growing market of CDP vendor solutions, according to Forrester

Regardless of the need, teams still have work to do in order to integrate CDPs within their marketing technology (martech) ecosystem. So what should you know if you're investing in a CDP? These practitioners who have been involved in CDP rollouts share some lessons they've learned along the journey.

Get the Right Skills, Capabilities and Resources in Place First

Doing this will ensure you are starting off in a good place and will lay a solid foundation for your progress.

Jeremy Muras, SVP of digital for Lion Capital, who called himself a huge CDP fan, along with his team, created a central database that he called a data co-op with the CDP. They’re putting all the brands in the company onto the CDP to help create audiences and to help customize lifetime value. 

The first lesson he shared about the CDP rollout? “You've got to have the right capability and have the right resources in-house,” he said, “to actually do a lot of the work, but it also has to have a real high-level sponsor that really drives it through.”

Don’t Think Data Will Magically Transform

Data abounds but much of that data is unstructured and will need work to clean and integrate. What Muras has found is that most brands have data that is poorly structured and needs work. 

He also found that not all CDPs are as good as the vendors say they are in terms of really being able to transform that data structure. The lesson? It will take work regardless of the promises from the provider of the technology.

Add a Data Product Manager

When Craig Kelly served as chief digital officer at, the team rolled out a CDP (using mParticle, for whom he now works for as head of product). One of the goals was to get rid of the bad data, start to drive visibility into high data volumes and control what type of data can come through.

One big lesson they learned along the way? You need to assign a data product manager who treats data as a product that serves multiple stakeholders in the organization. "We did eventually end up instituting that, and that helped a lot,” Kelly said.

Recognize CDPs Are a Forever Project

Asked how long the CDP implementation took, Kelly called it a “forever type of project." There is no end date, he added, in terms of how you implement and use a CDP. "There are always going to be new types of data that you want to collect and activate on," Kelly said. "There is always going to be channel expansion, different ways we want to interact with customers and different ways that customers want to interact with the brand.”

That said, Kelly added it took about four to six weeks with his Overstock team’s CDP rollout, and they began with a couple of use cases.

The questions they would commonly ask include: 

  • What are all the bidirectional data flows that I want to watch and run? 
  • Do I need to upgrade other systems?
  • What are the new applications that I want to build on top of this?
  • Do we have all the data that we require?

Making the Shift to Customer Experience Focus

This recognition that CDPs will help customer expansion on different channels and touchpoints can help you win at the customer lifetime value (LTV) game, Kelly said. “What is the infrastructure we need to be able to keep up with that and allocate our resources toward building those experiences instead of just capturing the raw data?” Kelly said. “How much time do organizations spend building API connections and meeting all of the maintenance on the back end of having to build these systems to be able to shift what might be 50% or 60% of all the engineering and focus on customer experience? That's how you win in the LTV game.”

Driving Marketing from CAC to LTV

Kelly called LTV the sum of your interactions across channels. “You have to increase the surface area of where you can target somebody that's an identity problem," he said, adding, "and you have to also increase the depth of the profile to understand how you can interact with them across channels, and kind of keep up with them where they are.”

Next comes freeing up the operational side and your resources. who are currently building back-end systems and pipelines and refocusing them on building front-end, customer facing applications that are going to impact LTV. “So it's kind of an all-hands-on-deck, drive-LTV program," Kelly said, "which I think everybody's kind of shifting to as customer acquisition cost (CAC) inevitably goes up 10% to 15% a year."

Latest from mParticle

See all insights
mParticle 2.0


Deep-dive into the new mParticle: A unified platform and updated UI

The new mParticle featured image thumbnail


Welcome to the new mParticle

Mach Alliance


Leading the next generation of CDP solutions: mParticle celebrates acceptance into the MACH Alliance