CDP - Customer Data Platform

Last updated Apr 16, 2024

A customer Data Platform (CDP) is a system that collects large quantities of customer data from a variety of channels and devices, enhancing accessibility for those who need it. CDPs sort, categorize, and cleanse data, removing inaccuracies or outdated information.

From What’s a Customer Data Platform? The Ultimate Guide to CDPs:

The Customer Data Platform, or CDP, provides a unique, unified view into your customers’ minds and needs amid a world of a million customer touchpoints and interactions.

Interesting distinctions include the differences between Reverse ETL and ELT: Differences to (Reverse)ELT. CDPs are often compared with CDI (Customer Data Infrastructure).

Arpit Choudhury writes extensively on this topic.

CDPs initially powered various software types like marketing automation suites, personalization engines, or campaign management tools.

The history of the customer data platform (CDP) begins with the evolution of marketing technologies, particularly with the development of customer relationship management (CRM) systems and data management platforms (DMP).

The first CRM software, TeleMagic, debuted in 1985, offering more than just simple contact database functionalities. This was soon followed by Act! in 1987, and GoldMine in 1990, which integrated contact management with sales and marketing tools.

The customer data management (CDM) market emerged in the 1990s, evolving from on-premises CRMs to cloud-based solutions by the late 1990s, with Salesforce pioneering the software subscription model in 1999. Over 50% of CRM installations were considered failures by 2006, leading to significant dissatisfaction.

To overcome integration challenges, vendors enhanced their databases with APIs, evolving them into what are now known as CDPs. These platforms connected with various martech tools to improve customer experience management.

DMPs, developed in the 2000s, focused on managing anonymous customer profiles for advertising but struggled with long-term data storage and known customer data management. This led to a demand for more versatile data management platforms.

By the early 2010s, the martech landscape was highly fragmented, leading to the need for unified customer data management. The term “customer data platform” was coined by David Raab in 2013. Since then, CDPs have evolved into sophisticated platforms incorporating AI and machine learning for advanced data analysis and management. Today, they are central to martech stacks, offering features for privacy regulation compliance and predictive customer insights. The market for CDPs is expected to see significant growth.

Further reading: The History of the Customer Data Platform (CDP) and CRM

# Legacy CDP according to Hightouch

In a legacy CDP, marketing teams could centralize customer data directly within the CDP as an all-in-one solution. What we initially missed was the rise of the cloud data warehouse and its impact on managing and activating customer data.

# Composable CDP

Hightouch now uses the term “composable CDP” to describe their new approach to customer data platforms RW Friends Don’t Let Friends Buy a CDP Hightouch:

Data Collection:

• Event tracking (often using tools like Snowplow or Segment/Rudderstack): Generate, enhance, and model high-quality behavioral data across all platforms and channels in a uniform format, streaming it into your data warehouse or lake.
• ETL (commonly Fivetran): Replicate data from your SaaS tools and databases across various domains such as marketing, sales, and finance into your data warehouse.

Data Transformation:

• dbt: Post data collection, use SQL to clean up and transform the raw data into usable tables/views within your data warehouse.

Data Activation:

• Hightouch: Sync data from the data warehouse into business-critical tools like Salesforce, Marketo, and Facebook Ads.

To me, CDPs are similar to data warehouses or data engineering platforms, tailored specifically for one domain. There’s no need for a new term or separate tools for what essentially involves data integration followed by a data lake or data warehouse and, ultimately, generating insights.

This image, though depicting a GTM stack and not a CDP, illustrates the similarity, starting with data integration, moving to a data lake or data warehouse, and culminating in insights. Image from [Approaching Go-to-Market as a Data Engineer](

Origin: Unbundling the CDP
References:
Created 2022-03-16