RudderStack — Warehouse-Native CDP (Product Overview)

Snippet-only — full fetch unavailable from this environment (tier-1: HTTP 403 on both rudderstack.com URLs attempted; tier-4 search snippet accepted).

Product classification. RudderStack positions itself as a composable, warehouse-native CDP: it does not store customer data in its own datastore, instead building the CDP layer on top of the customer's own warehouse (Snowflake, BigQuery, Databricks, Redshift).

Core capabilities (per search snippet synthesis):

• Event streaming and collection across the customer journey (positioned as comparable to Segment's event-collection surface).
• Transformations authored in JavaScript or Python.
• Identity stitching via a "Profiles" feature.
• Reverse ETL: syncing warehouse-computed audiences, conversions, and enriched profiles out to downstream marketing/business tools.
• "Data Apps" for attribution and propensity modeling without hand-written SQL/Python, running on top of the customer-360 warehouse tables.

Stated limitations (self-disclosed by vendor). RudderStack does not send email, SMS, or push notifications itself — it is an activation/sync layer, not a messaging platform. The vendor also explicitly cautions that data warehouses are optimized for analytical queries rather than sub-second profile lookups, and that in-session personalization, triggered messaging, and real-time AI decisioning require API-speed profile access that warehouse query latency cannot deliver — i.e., RudderStack self-positions outside the intra-page / intra-session latency tiers.

Context for KG. RudderStack is a composable-CDP peer to vendor.hightouch and vendor.census (event-streaming + reverse-ETL category) with no existing KG vendor node. Surfaced during the 2026-08-01 web-refresh seed-topic search for "Composable CDP vendors." See evolution-log/2026-08-01/web-refresh.md for the draft vendor.rudderstack candidate.