Customer Data Platforms (CDP)

These vendors are hired to take customer data that's scattered across a CRM, an ecommerce platform, email tool, app analytics, point-of-sale system and ad accounts, and turn it into one profile per customer that other systems can actually use. The concrete work is identity resolution (matching an anonymous website visitor to an email subscriber to a paying customer), building pipelines that pull data in from those sources continuously, and pushing unified segments back out to the tools that run campaigns — ad platforms, email and SMS senders, personalization engines, call centre software. Once it's working, a marketer can build a segment like 'lapsed high-value customers who opened the last email but didn't click' without filing a ticket with engineering.

People end up looking for this when the customer picture inside their company has fractured. The same person gets emailed as a prospect after they've already bought. Support has no idea what the customer looked at online before they called. Marketing wants to run a win-back campaign but the data needed to define it sits in three databases nobody can join cleanly, so building the list takes weeks and by the time it's ready the moment has passed. Often there's also a compliance angle: consent and preference data living in yet another system, making it hard to prove who opted out of what.

The providers here split roughly into two camps. One group sells a packaged CDP with its own data store, built-in identity resolution, and marketing features layered on top — segmentation, journey building, sometimes predictive scoring for churn or lifetime value. The other group is closer to reverse-ETL: they assume your data already lives in a warehouse and their job is just to sync it out to activation tools, leaving the modelling to you. That split drives most of the real trade-offs: owning your own warehouse gives more control and avoids duplicating data, but means more engineering work; a packaged CDP gets marketers self-sufficient faster but locks more of your customer data inside the vendor. Beyond that, compare them on how identity resolution actually works (deterministic matching on logged-in IDs versus probabilistic matching across devices), whether they're built for high-volume consumer behavioural data or smaller B2B account-based data, how real-time the activation is, and what's genuinely built-in versus what needs a consultant or integration partner to make work.