Just got the alert. The latest ruling means "right to data portability" now explicitly includes *processed* historical data—not just raw logs. Your old journey maps, lead scores, attribution weights? They're part of the dataset.
So when you migrate CDPs now:
* Your "historical event backfill" just got heavier. It's not just copying raw events. You need to port the *derived* state.
* Good luck translating your old CDP's proprietary scoring model into the new one's schema. How do you backfill a lead score from system A into system B when the logic is a black box?
* Downstream models (attribution, segmentation) that depend on that historical processed data will break if the ported data doesn't match in granularity.
Who's actually prepared for this? Most migration plans I've seen only handle raw data pipes. If you're moving from say, mParticle to Segment, or vice versa, how are you planning to export and reconstruct the *processed* customer profile state as it existed at the time of the request?
They've been selling "portability" for years. It's nonsense. Your contract says you own your data, but the format is proprietary. Exporting raw events is one thing, but processed data is the product.
Ask for the spec to export a historical lead score with full context. They'll go quiet. The ruling changes nothing if the schema and logic aren't open.
The real question is who pays for the engineering time to rebuild the derived state? Because the vendor isn't doing it for free. Your next invoice will have a new line item for "compliance data reconstruction."
Trust but verify.