How to choose a customer data platform

Every customer data platform does the same three things in principle: collect customer data from many sources, resolve it into a unified profile, and make that profile usable elsewhere. Where they differ sharply is who is meant to operate the tool day to day, and that single question eliminates more of the market than any feature checklist.

A CDP is not the right purchase for a company with one or two data sources and a small list — that's a job for the CRM or email platform you already have. It becomes worth evaluating once customer data is genuinely fragmented across systems (web, POS, CRM, loyalty, support) and no single existing tool can stitch it together.

Marketer-owned versus engineer-owned

This is the decision that should come first. Bloomreach Engagement and BlueConic are both built for marketers to run directly: BlueConic specifically emphasizes no-code, marketer-managed data collection and segmentation so day-to-day changes don't route through engineering, and Bloomreach pairs its unified profiles with a visual, omnichannel campaign builder across more than a dozen channels. mParticle and RudderStack sit at the other end: both are infrastructure that a product or data-engineering team owns, focused on schema validation, data governance and routing clean event data to downstream tools rather than building campaigns themselves. Buying an engineering-first CDP for a marketing team with no engineering support produces a powerful pipe nobody can operate; buying a marketer-facing CDP for a data team that wanted warehouse-level control produces the opposite frustration.

Warehouse-native versus its own data store

A newer architectural split is whether the CDP keeps its own copy of your data or reads directly from a warehouse you already control. RudderStack is warehouse-first by design — it loads event data into your cloud warehouse and can sync data back out (reverse ETL) from the same pipeline, and its core is open source. Salesforce Data 360 uses what it calls zero-copy integration to connect to warehouses like Snowflake and Databricks without duplicating the source data. Treasure Data and Adobe Experience Platform, by contrast, both build their own unified profile store that other tools then read from. Warehouse-native designs reduce data duplication and governance overhead if you already have a mature warehouse; a self-contained store is simpler to stand up if you don't.

How identity resolution is actually built

Identity resolution quality is the hardest thing to evaluate from a features list, because every vendor claims to do it well. A few records here describe their approach specifically enough to compare: Amperity is built around a probabilistic and deterministic identity-resolution engine designed to work without a rigid schema upfront, aimed at retail, hospitality and travel brands with especially messy, overlapping identifiers. Zeotap emphasizes deterministic identity matching and match-rate accuracy specifically, with a European, GDPR-oriented privacy posture reflected in its "InfraFlex" flexible-deployment architecture. Most of the rest of the category — Adobe, mParticle, Tealium, Treasure Data — describe identity resolution as a capability without detailing the underlying matching method, which is a fair question to put directly to a vendor rather than assume.

Ecosystem gravity

Three of the eleven tools here have their strongest case tied to a specific ecosystem you may already be paying for. Adobe Experience Platform is the shared data layer under Adobe's Real-Time CDP and only pays off fully alongside other Adobe Experience Cloud apps (Target, Journey Optimizer, Campaign). Salesforce Data 360 is native to Salesforce's Sales, Service and Marketing clouds and is now positioned as the data foundation for Salesforce's Agentforce AI agents specifically. Tealium grew out of Tealium iQ tag management, so a company already using Tealium to collect data has a natural, lower-friction path to extending it into a full CDP. If your organization is already committed to one of these ecosystems, that CDP starts with a real integration advantage before you evaluate anything else.

Open source versus commercial, and what "free" really means

RudderStack is the only open-source option in this category, released under the Elastic License 2.0, with a free self-serve tier up to 250,000 events per month and a paid Growth plan beyond that. Being open source here doesn't mean free to run at scale: self-hosting still costs infrastructure and the engineering time to maintain it, and RudderStack's own paid tiers exist because most teams choose the managed cloud version anyway. Every other tool in this category is closed-source and quote-only or usage-based.

How pricing scales

Pricing model is one of the few things that predicts total cost before a sales call. RudderStack is the only tool with published self-serve numbers (by event volume). Salesforce Data 360 is explicitly usage-based across three components — consumption credits, data storage and premium add-ons. Bloomreach combines a module fee per channel with a usage fee tied to customer and message volume. The rest — Adobe, Amperity, BlueConic, mParticle, Simon Data, Tealium, Treasure Data, Zeotap — are quote-only with no public figures at all. For any quote-only vendor, ask what specifically drives the number: event volume, data-source count, number of unified profiles, or a flat enterprise license.

An AI layer is now standard, not a differentiator

Nearly every vendor in this category has added a generative or agentic AI layer in the last cycle: Bloomreach's Loomi, Salesforce's Agent Context Engine feeding Agentforce, Zeotap's ZeoAI, and Treasure Data's rebrand to an "Agentic Experience Platform" with its own AI Studio. Simon Data, now rebranded Simon AI after its 2026 acquisition by personalization vendor Monetate, has gone furthest in this direction — layering AI agents that plan and launch campaigns from a stated goal on top of its original warehouse-native CDP. Treat these as a feature to test on your own data rather than a reason to choose one platform over another; the underlying data-unification job is still the harder, more consequential decision.

A shortlist by situation

• If marketing needs day-to-day, no-code control without depending on engineering, look at BlueConic or Bloomreach Engagement.
• If you want the CDP to read from a warehouse you already control, with reverse ETL in the same pipeline, look at RudderStack.
• If your identity data is especially messy — many overlapping loyalty, POS and e-commerce identifiers — look at Amperity.
• If you're already standardized on Adobe or Salesforce, that platform's own CDP (Adobe Experience Platform or Salesforce Data 360) has a real head start.
• If you need governed, compliant event data fanned out to many downstream analytics and marketing tools, look at mParticle or Tealium.
• If you're European or privacy-sensitive and want flexible deployment architecture, look at Zeotap.

Questions to ask vendors

• Walk us through exactly how identity resolution works on our data — deterministic, probabilistic, or both — and what a false match looks like.
• Does the platform keep its own copy of our data, or can it read directly from our warehouse?
• What triggers a cost increase on our contract — events, profiles, data sources, or seats?
• How is consent enforced before data reaches downstream destinations, and can we audit it?
• If we later want to leave, what does data export actually include, and in what format?

Common mistakes

Buying an engineering-first CDP because it has the strongest brand recognition, then discovering no one on the marketing team can operate it, is the most common and expensive mistake in this category. A close second is underestimating implementation: unifying genuinely fragmented first-party data and building a working segmentation model is a project measured in months, not a configuration step. And treating an open-source label as "free" ignores the hosting and engineering cost of running RudderStack yourself instead of its managed cloud tier.

For two direct comparisons inside this category, see Adobe Experience Platform vs Salesforce Data 360 and mParticle vs Tealium. The full list of tools in this category is at every tool in this category.