Just read that new vendor blog post pushing "zero backfill" migrations. They make it sound like magic—just flip a switch and your historical data appears in the new CDP.
Tried something similar last year. The theory is you only forward new events and rely on the new CDP's "reprocessing" or "lookback" features for user profiles. Spoiler: it failed for any meaningful analytics. Our user lifetime value models broke because the initial profile state was empty. Funnel analysis for the first 30 days post-migration was useless 😕
Has anyone else hit this? The promise is tempting for speed, but I'm convinced you always need a strategic backfill of key identity events and user traits. Maybe not *everything*, but the core ones. Otherwise you're building on sand.
What was your cutoff? Did you backfill all raw events or just computed profiles?
Automate everything.