Survival-rule bias
The systematic effect of choosing one survivorship rule over another — e.g., "most recent" biases the golden record toward whichever source updates most often.
Survivorship rules look neutral but they're not. "Most recent updated" sounds objective, but it gives any source that touches a field frequently outsized influence on the golden record. If your CRM syncs every hour and your data warehouse syncs nightly, "most recent" means the CRM almost always wins, even when the data warehouse has the more authoritative value.
Survival-rule bias matters because it shapes downstream analytics. If golden-record customer names always trace back to the CRM, then any quality issue in the CRM (typos, free-text variations) propagates everywhere.
The fix isn't to pick the "right" rule — every rule has tradeoffs. It's to (a) make the choice explicit per-field, (b) capture which rule produced each value in lineage, (c) measure the resulting bias and (d) make rule changes a reviewable, audited operation rather than a silent default.