Which customer data platform owns identity? Tealium, Salesforce, Adobe, Segment and warehouse-native CDPs

What to take away

Tealium AudienceStream

Tealium AudienceStream is a packaged CDP. It keeps persistent visitor profiles, stitches identities from a configurable rule set, manages audiences, and ships connectors. Tealium advertises more than 1,300 prebuilt integrations across its EventStream, AudienceStream and DataAccess products. Real-time actions and privacy tooling come with it, and predictive scoring is sold on top.

It suits teams that want one vendor accountable from ingestion through activation and do not want to assemble that stack themselves. Pricing is quote-based, so ask for the connector count you will use and the event volume behind it.

Test it on export. Ask what leaves the platform, in what shape, and how fast.

Salesforce Data 360

Data 360 is Salesforce's data and activation layer for its own clouds: ingestion or zero-copy access, unification, workflows, decisions and analytics. It fits organizations already deep in Sales, Service, Marketing or Commerce, where activation targets are mostly Salesforce objects and Agentforce agents.

Salesforce sells the product in credits, and entry enterprise agreements typically land in the six figures a year in most markets. Model your credit burn before you sign, because the bill follows usage.

The hard test is the boundary. Ask which capabilities stop at the edge of the licensed products.

Adobe Experience Platform

Adobe Experience Platform pairs a data lake with Real-Time Customer Profile, Identity Service, segmentation and destinations into Adobe and third-party tools. Real-Time CDP sits on that foundation, with Customer Journey Analytics and Adobe Journey Optimizer sold beside it.

It suits enterprises already running Adobe Analytics, Campaign or Experience Manager, and teams with data engineering capacity to run schemas and datasets properly. Pricing is enterprise and negotiated, typically six figures a year in most markets.

Ask how much of the identity graph you can inspect and export without Adobe's tooling.

Twilio Segment

Segment leads with collection: SDKs, sources, transformations and a wide destination catalog, with Unify adding identity resolution and computed traits. Segment publishes a free tier at 1,000 monthly tracked users, and the Team tier has been listed around $120 a month for 10,000 monthly tracked users. Business sits above that and scales with volume.

It suits product and growth teams that want clean event data flowing everywhere fast, and it is the easiest of these to start on. The question is depth. Confirm how much profile logic Unify owns versus what stays in your warehouse.

Amperity, Treasure Data and BlueConic

Amperity builds its own identity resolution, called Stitch, over transaction and profile data, and it bought Custora in 2021 to add predictive modeling. Retail and consumer brands use it to unify loyalty, ecommerce and point-of-sale records. Pricing is quote-based and typically annual.

Treasure Data runs batch and streaming ingestion into its Customer Data Cloud, with a SQL audience builder and a SoftBank parent. BlueConic takes a marketer-first route: profile building, consent management and on-site dialogue in one interface. Neither publishes a rate card, so ask Treasure Data about event-based pricing and BlueConic about profile and seat pricing.

Warehouse-native builds

A warehouse-native CDP keeps data and modeling in Snowflake, BigQuery or Databricks, then layers identity, audiences and activation on top through tools like Hightouch, Census and RudderStack. ActionIQ runs its Hybrid Compute identity resolution inside the customer's Snowflake, Databricks, AWS or Google Cloud account. Simon Data runs its CDP on Snowflake.

It suits teams with strong data engineering, existing warehouse investment and a preference for owning the model. Cost is the warehouse bill plus the activation tool. Hightouch publishes a free tier and paid plans that scale by rows synced; RudderStack prices by events.

Responsibility spreads across components. Marketing data governance has to name who owns identity, consent and deletion before you commit.

The comparison table

Prove it with one scenario

Give every finalist the same scenario. An anonymous visitor becomes known. Two systems disagree on email. Consent is withdrawn. A purchase lands late. One destination fails. Then the record must be corrected and deleted.

Score accuracy, recovery time, operator effort, and data location. Also score latency, full cost, and who was accountable. The GAO data reliability guide frames reliability as fitness for an intended use with documented assessment. Apply that to your scenario, not to a vendor's benchmark.

Public performance claims carry a substantiation duty in the US under the FTC advertising substantiation policy. Ask vendors for the evidence behind customer data platform comparison claims before you repeat them.

Common questions

Which CDP is best?

None universally. Tealium suits teams wanting one accountable vendor, Data 360 suits Salesforce-heavy estates, Adobe suits Adobe estates, Segment suits fast collection, Amperity suits retail identity work, and warehouse-native stacks suit strong data teams.

Can a warehouse-native stack count as a CDP?

Yes, if the composed system clearly owns persistent unified records. It must own identity, governance and downstream activation. It must also own monitoring, correction and support. If nobody owns those, it is a pipeline, not a CDP.

How should vendor claims be compared?

Give every finalist identical data, identities, permissions, and a destination failure. Give them a correction, a deletion, the same volume, and the same operator. Record which capabilities were native, configured or custom. Then note which were partner-delivered, preview or roadmap.