Reverse ETL

The pattern exists because the warehouse became the only place where a concept like "high-value at-risk customer" is defined consistently across every source, and the sales and marketing tools that need to act on that concept cannot see it.

The alternative — recomputing the segment inside each operational tool — reproduces the definition in several places, where the versions diverge and nobody can say which is authoritative. Reverse ETL keeps one definition and syndicates its output.

Three cautions determine whether it becomes an asset or a liability. Latency: the warehouse is batch, so operational systems receive data that is hours old, which is fine for a marketing segment and wrong for anything transactional. Coupling: an operational system now depends on a warehouse pipeline, so a failed transformation becomes an operational incident rather than a stale dashboard, and the pipeline needs an operational SLO it probably does not have. Governance: this is a data export path, so it must respect the same access controls, residency rules and consent constraints as any other, and it very often does not.

Where it is genuinely the wrong tool: anything needing sub-minute freshness, which belongs to a streaming path or a direct service call, not to a warehouse round trip.