I'm trying to sync customer profiles between a CDP we use for analytics (ActionIQ) and another we use for email campaigns (Simon Data). The idea is to use a single identity graph as the source of truth for both.
Has anyone actually done this? I'm curious about the technical lift. Do you pipe the raw graph data into each, or try to make one CDP feed the other? Any major pitfalls with mismatched resolution rules?
Yeah, we attempted this with Segment and mParticle as the central graph feeding into Braze and a warehouse. The biggest pitfall wasn't the plumbing - it was the *mismatched resolution rules* you mentioned.
You can pipe the raw event stream into both, but each CDP will still build its own internal graph from those events, applying its own deterministic/probabilistic logic. The "source of truth" becomes messy when ActionIQ merges two anonymous sessions that Simon Data keeps separate, or vice versa. Your segmentation starts to drift.
My take? Use the graph from the analytics CDP (ActionIQ) as the master, and *push resolved profiles* to the campaign CDP, not raw events. Treat Simon Data more as an execution layer. You'll still have some edge cases, but it keeps your merge logic in one place. The lift is in building that reliable profile export pipeline.
pipeline all the things