Who Owns Your Dealership's Data? Turning a CDP Into a Real Strategy

A dealer group signs a contract for a customer data platform. The demo looked great, the sales rep answered every question, and the ROI projections made sense on paper. Six months later, nobody inside the dealership can say who's actually responsible for the thing, and the CDP is quietly becoming another tab nobody opens.

That gap between buying a CDP and actually building a real data strategy around it is the real subject of a recent episode of V20 Voices, hosted by Launch Labs CRO Mike Morgan with guests Jim Kosobucki of Morgan Auto Group and Jeremy Nowling of Rohrman Automotive Group.

What Better Data Actually Does for Dealership Marketing

Mike opened with the question most dealership managers actually have: why should CDP, data warehouse, and identity resolution matter to someone selling or servicing cars? Kosobucki answered that a structured, unified customer profile is what makes AI actually useful down the line. Nowling added that better data means better decisions, and confidence a marketing message reaches the right customer instead of a guess.

That shift is already changing how both groups market. Instead of one broad message going out to everyone, they can now speak to customers based on where they actually are in their journey—a level of precision that simply wasn't possible before.

Data Access vs. Data Ownership: Why the Difference Matters

One idea ran through the entire conversation: real value comes from owning your customer data outright, not just accessing it through a vendor's platform. As Nowling put it, the whole push toward a real data strategy was about being able to own their data to make better decisions, rather than depending on a vendor to decide what they could do with it.

Morgan drew out a distinction worth noting. Owning your data isn't just about having access to it. It's about owning it in a form that's actually portable, so it can move from one system to another without being rebuilt from scratch. That single idea, ownership as portability rather than just access, is what shapes almost every practical decision the two groups described.

Deterministic vs. Probabilistic: The Identity Resolution Layer

This is where the conversation moved from principle to mechanics. A dealership's own CDP typically includes native identity resolution, and it's highly reliable—but it's built entirely on deterministic matching: a shopper fills out a form, and the system is certain that a specific name, email, and activity all belong to the same person. The limitation is stark. Industry-wide, only about 2% of website visitors ever fill out a form. The rest of the traffic, and the intent behind it, is effectively invisible.

Nowling described the second layer his team built through their partnership with Launch Labs: probabilistic identity resolution. Instead of requiring a form fill, it assigns a confidence score based on signals like browsing behavior and device patterns. That means a shopper looking at SUVs can be identified and added to a relevant audience, or activated into an email, direct mail, or text campaign, even if they never raise their hand.

• Deterministic: high certainty, but limited to shoppers who explicitly identify themselves
• Probabilistic: confidence-scored matching that captures shoppers who never fill out a form
• Used together, they extend a dealership's addressable audience well beyond the small share of visitors who typically convert on a form

The practical result: a dealership can start marketing to someone who showed real intent on the website, without waiting for them to take an action they may never take.

Why AI Tools Depend on Clean Data

Once the data foundation is solid, AI stops being a talking point and starts doing real work. At Morgan Auto Group, Kosobucki's team is exploring a website chat feature that draws on the full depth of a customer's CDP profile, so a conversation can reference their actual purchase history, service visits, and past sentiment instead of starting from nothing. As website behavior continues shifting away from static forms toward chat, that context is what will make the difference—and no vendor can offer it without a dealership's data already being unified in one place.

Mike's framing tied it back to the bigger point: AI isn't a stand-alone solution. It performs in direct proportion to the data quality underneath it. Get that right, and the results follow.

The Measurable Revenue Behind a Dealership Data Strategy

Ask most dealership managers what a CDP is actually worth, and the honest answer is usually a hunch: marketing "feels sharper," or they're "reaching more people." Kosobucki and Nowling can point to something a lot more specific than that.

At Morgan Auto Group, probabilistic identity resolution matches anonymous activity back to known customers, surfacing what the team calls "data-ups”—leads that wouldn't exist without that added layer of matching. In March alone, that one method produced roughly 7,500 of them, and Kosobucki ties a hard outcome directly to it: "We sold over 100 cars just from those data-ups.”

Rohrman Automotive Group is running similar math from a different angle. Customers who start booking a service appointment online and never finish used to just disappear from the pipeline. Now Nowling's team follows up with them directly, and can point to the actual revenue that follow-up recovers.

How to Start Building a Dealership Data Strategy

Mike's closing summary boiled everything down to three things: put one person in charge, understand the landscape of CDP options, and remember the CDP is a foundation, not a finish line. Identity resolution and AI are separate, additional layers built on top of it.

Nowling's closing message was the most direct: don't wait, because "the industry is going to continue to pass you by." Start researching now, and talk to people who've actually gone through it.

Practical Takeaways

• Assign a single data owner. This isn't a committee exercise. Someone needs clear responsibility for the strategy and the vendor relationships.
• Hire or train for automotive fluency, not just technical pedigree. The best candidate may already be inside your organization.
• Use deterministic and probabilistic identity resolution together to extend your addressable audience beyond form-fill shoppers.
• Vet vendors on fit and reputation—not the pitch. Research first, meet vendors second.
• Structure your data for portability before you sign with any single vendor.
• Layer AI in only after the data foundation is clean. AI amplifies whatever it's given, good or bad.

Listen to the full episode of V20 Voices for the complete conversation with Jim Kosobucki and Jeremy Nowling.