One brain, every channel: the automation a CDP has to earn
Automating a channel is easy, and almost every vendor in automotive can do it. Email sends on a trigger. SMS sends on a trigger. The mail house drops a file on a schedule. The BDC works a list. Every one of those systems is automated, and together they still produce an experience that feels like four companies talking over each other. Automation is not the missing piece. Coordination is.
The coordination failure nobody owns
When each channel holds its own rules, every rule is right and the result is still wrong. The email platform caps sends at two per week and is obeying that cap. The SMS vendor has its own cap and is obeying that too. The mail house honoured the suppression file it was given on the fifteenth. Paid social is running an audience refreshed last Tuesday. Nobody broke a rule, and the customer who took delivery on Saturday still got a conquest offer on Monday, a service reminder on Wednesday and a retargeting ad for the car now sitting in their driveway.
This is the structural argument for putting suppression in the unified profile layer rather than in each channel. Channel-level suppression can only see channel-level activity. The profile layer is the only place where all of it is visible at once, which makes it the only place a global frequency cap can actually mean anything.
Move the logic upstream
The pattern emerging across the category is simple to state and hard to retrofit. Define the business logic once, in the customer data platform, and let downstream tools receive instructions rather than make decisions. The email platform stops deciding who is eligible and starts executing a send it was told to make, with the context it needs attached. The mail file stops being a query someone runs and becomes an output of the same decision engine that governs every other channel.
That inversion is what turns a set of automated channels into an orchestrated programme. It also removes the most common failure mode in these deployments, which is the handoff. Data leaves the platform, crosses an API, and arrives somewhere else late, with a dropped attribute or a mismatched field. Every handoff is latency and a chance to be wrong. Fewer decisions downstream means fewer handoffs that matter.
Define the logic once and let the channels execute. A channel that makes its own decisions is a channel that will eventually contradict you.
Where automotive is genuinely harder
Most orchestration writing assumes digital channels with instant suppression. Automotive does not get that luxury, and the differences are worth naming.
- Mail is committed before it lands. A piece entering production on Monday arrives ten days later, so suppression has to happen at the file, and the sale that happens in between has to be reflected in what the customer receives next rather than in what was already printed.
- The store is a channel. A delivery, a repair order or a declined estimate is a signal that has to travel back into the profile the same day, not in a monthly extract.
- The household is the unit. Two rooftops in one group reaching two people at one address about the same vehicle is a frequency failure, even though every individual cap was respected.
- Service and sales disagree. A retention message and a conquest offer to the same household in the same week are not two campaigns. They are one conversation, and something has to decide which one wins.
What good looks like
A useful test is to pick one event and follow it. A customer takes delivery. Inside the hour, the profile updates, the conquest audience drops them, the paid platforms receive the change, the pending mail file is amended, the service clock starts, and the follow-up sequence that begins is the ownership one rather than the acquisition one. If that sequence takes a week or a person, the platform is a reporting layer with a marketing label on it.
The second test is measurement. Orchestration is judged in appointments, repair orders and delivered units, attributed across the whole journey rather than to the last channel that touched it. Channel-level reporting will always show every channel performing well, because each one is measured against the traffic it was handed.
None of this requires a new channel. It requires one place where the customer is known, one set of rules about how often they hear from you, and the discipline to stop letting each vendor answer that question for itself.
Back to all articles