CDP 2.0: The Resurgence

The CDP space is undergoing a Dionysean transformation and next generation of CDPs will deliver the highest return on capital investment of any marketing technology.

For those unfamiliar with obscure Roman mythology, Zeus saved Dionysus at birth by sewing him to his thigh, making him the “twice born God.” I still think it’s crazy that actually happened.

As I’ve written in previous posts, solutions that prioritized the “ways of doing the job” and not the “jobs to be done” have struggled amidst tighter macro conditions, limited budgets, and encroachment from adjacent categories.

The CDP category has become associated with expensive and time consuming implementations with limited measurable business value. Composable CDPs have capitalized on this trend by offering lower cost tooling that leverages the data foundations already established in a cloud data warehouse.

With agentic artificial intelligence, solutions that function as a “marketing copilot” have the opportunity to materially improve business outcomes for the teams that use them.

I remember reading a post from Martin Kihn of Salesforce which said that “you should be able to talk to your CDP.”

At the time I assumed that such product marketing involved taking ayahuasca and vomiting industry jargon in the form of things that aspirationally resonate with marketers but would never actually be built.

While Salesforce does tend to market, sell and then build software in that order, there’s something prescient in Martin’s vision.

The CDP should be the solution the marketer goes to in order to ask questions about customers, what the data shows and (most importantly) how to drive better outcomes.

This is the CDP 2.0 and I have to give credit to Michael Katz of mParticle for coining the term.

Let’s consider a practical example of how this could work.

I’ve flown Delta for over a decade and (humble brag) have had Platinum status or above every year. This past year, due to a variety of issues, I was left stranded in no fewer than half a dozen random cities overnight, staying at budget hotels that they comped and eating alone at a Chili’s or some such place.

I missed important business meetings and spent probably a hundred hours at the airport waiting for flights that were sequentially delayed by an hour each time.

At the end of the year, I received an “inconvenience bonus” of 1,000 miles, which is roughly valued at $10.

Now, either the person who made that business decision at Delta is like the CEO of Clark Griswold’s cereal varnish company and is really bad at understanding humans, or alternatively, as I believe is more likely, Delta has a data challenge.

Now, imagine an alternate scenario where the same person in charge of making the decision to offer me nothing went into a software solution connected to all of Delta’s data. They could see every delay the customer experienced, how many miles they traveled and what their drink orders were at the lounge.

This person could then ask “create me an audience of people who are most likely to decrease in loyalty this year” and then “what incentive is each individual most likely to respond to.” Some might be motivated by points, others by just a personal note of atonement and gratitude. This would be reflected in customer support interactions/emails.

Doing this without a tool connected to all data would be impossible. Connecting all of that data, including real-time, historical, loyalty, etc. is also very challenging as anyone in the CDP space knows.

The promise we as a category have delivered on is the latter: getting the data together, performing identity resolution, and integrating it across all marketing tools and systems. What we have yet to deliver on is automatically surfacing the right audiences and content based on the business goal.

With AI agents, this is now possible and can be especially powerful if approached from the lens of the business outcome.

The underlying “CDP work” of collecting, organizing and integrating the data still needs to be done in order to deliver against this vision, making CDPs the best solution the deliver this capability.

Marketing channel tools will be unable to achieve this because their data model isn’t customer centric; they view data as events, messages, journey stages, etc.

These tools will use AI to optimize send time, 1-1 content personalization and channel selection and they will be better than CDPs at some of these things due to the data they receive directly back on clicks, opens, etc. But, they wouldn’t know I spent the night on a prison mattress in Indianapolis and missed an important client meeting.

CDPs will own this next phase of customer-centric and (more importantly) business outcome-oriented AI marketing optimization.