Skip to content

What is a Customer Data Platform (CDP)?

ActionIQ, Smart hub Customer Data Platform

A CDP is an off-the-shelf technology that functions as the brain of an organization’s marketing technology stack. It starts by collecting customer identities and interaction data from all first- and third-party sources (through prebuilt, real-time data connectors), and stitches them together into a single, persistent profile for each customer. It then empowers marketers and other business teams with intelligence by providing a user-friendly interface for analyzing customers, segmenting audiences, and predicting next-best-actions. Finally, the user interface enables automated orchestration of 1:1 personalized journeys across all marketing, CX, and commerce channels, including test design and measurement.

What is a CDP?

As a marketer or business person, you may have heard the term CDP a lot recently; however, not a lot of people actually understand what the acronym means. So what is CDP? Customer data analytics are revolutionizing the business world; A customer data platform (CDP) is a prebuilt technology used by marketers and customer experience professionals to understand their organization’s customers and scale personalized experiences across marketing, sales, and service channels. Essentially by combining multiple sources of data (1st party data and 3rd party data), companies and marketing teams can now leverage machine learning for actionable insights and potentially the capability of certain levels of marketing automation in the near future.

The customer data platform system does this by:

  1. Collecting customer identities and real-time interaction data from all first- and third-party data sources (through prebuilt data connectors)
  2. Stitching them together and removing duplicates to form a single, persistent history for each customer
  3. Empowering marketers and other business professionals through a single view user-friendly interface to gain individualized customer insight by analyzing each customer, segment audience, and predict next-best-actions to foster customer relationships and purchasing.
  4. Enabling automated orchestration of 1:1 personalized campaign journeys across all marketing, customer experience, and commerce channels, including test design and measurement

What isn't a CDP?

Because of its generic name, a true CDP is sometimes confused with other technologies. Here is a sample of common misconceptions:

  • It’s not a data management platform (DMP)— which is designed to plan and manage paid digital advertising marketing platforms that are conducted on 3rd party websites (e.g. espn.com), search platforms (e.g. google.com) and social platforms (e.g. twitter.com) for anonymous individuals
  • It’s not a digital event distribution platform— which is designed for IT organizations to transmit data from one technology to another after a digital event has occurred (e.g. triggering content IDs from your content management system to your workflow management tool once new content has been created, or triggering your ESP that a frustrated customer just complained to the call center)
  • It’s not a digital personalization engine— which is designed to personalize the user experience on an organization’s first-party website.
  • It’s not a master data management platform (MDM)— which is designed for IT organizations within large, multichannel enterprises to resolve customer identities (PII only) into a single golden identity (which can then be used as an input into a CDP).
  • It’s not a multichannel marketing hub (MMH)— which is designed to manage and deliver marketing campaigns within email, social, text, and push channels (note: they are sometimes called marketing clouds).
  • It’s not a customer relationship management tool (CRM)— which is designed as a sales and service tool for logging direct interactions with customers (e.g. customer complaint, price quote on a product, customer questions) and updating basic customer information (e.g. name, phone, email). Check out our blog on CDP vs CRM for additional information on the key differences.
  • It’s not a data lake— which is designed for IT organizations to store any type of enterprise data (e.g. financial, product, store, service, transaction, HR, customer, marketing) to support enterprise analytics and data science. Data is accessible only via programming languages (e.g., SQL, Python, R, Java).
ActionIQ customer data platform competitive market guide
RECOMMENDED CONTENT

Separate the CDP posers
from the players

3 Key Benefits of a CDP

  • Incremental revenue generation—By providing meaningful insight into customer behavior and content preferences, CDPs enable marketers to unearth new opportunities and act on them quickly.
  • Saving marketing costs—By leveraging a CDP’s customer intelligence, marketers can reduce marketing costs through suppression and optimization tactics.
  • Increased operational efficiency—By integrating customer data and empowering self-service operations to business professionals in a unified view, CDPs eliminate operational bottlenecks between technical and business teams.

“With [ActionIQ], we were able to marry the usability of a typical campaign management solution with infrastructure grade architecture and scalability. Meaning, they know how to make a useful product for marketers, and build tech you can actually integrate into a modern infrastructure.
Nick Rockwell, Former CTO of the New York Times

Why a CDP?

“Personalization is impossible if marketers don’t have the means to understand the needs of customers on an ongoing basis. Setting up a centralized customer data platform (CDP) to unify paid and owned data from across channels is essential to these efforts."
McKinsey

Customer Data Platforms are the next level of marketing technology and empower business users and analysts to leverage all available data and insights to make every customer engagement count. They consolidate data silos, create a 360-degree customer view, and democratize access to customer data for purposes of personalizing interactions across marketing, sales, and service channels. Ultimately, a CDP gives marketers and customer experience professionals the ability to provide real-time personalization to customers by consolidating data silos and creating authentic customer interactions.

Why is Customer Data Important?

Customer data is important for three reasons.

First, consumers expect organizations to serve relevant experiences whenever they interact (thanks to companies like Amazon), and customer data is the fuel that enables organizations to identify the types of engagements each consumer values most.

Second, it’s more than just what experiences matter most, it’s also about being able to serve the customer a seamless experience no matter which channel they interact through. So if they were on the website last night but are calling into the call center today, the agent should have access to any frustrations the consumer was experiencing on the website last night to quickly provide assistance for their problem. The consumer should not have to re-explain their past behaviors each time they interact with a new channel/person.

The last reason why customer data is important is that marketing is expensive, and customer data helps you measure which types of mediums deliver the most ROI (e.g. display advertising, email, direct mail, etc.).

What is customer data?

Below are the types of customer data often managed by a customer data platform (CDP):

  • PII info—Personal customer information such as name, address, phone, email, etc. (oppositely for anonymous users, CDPs will store device ID or cookie ID)
  • Campaign and customer engagement info—Data reflecting the campaigns a customer received, the engagement within them, and other interaction data across marketing, advertising, sales, and support channels
  • Transaction info—Examples include products purchased, channels transacted through, purchase prices, discounts used
  • Metrics and scores—Examples include customer health scores, average click rates, customer loyalty status, lifetime value
  • Demographic or firmographic Info—Examples include dwelling type, estimated income, company type, company industry, number of employees
  • CSAT data—Examples include satisfaction scores from surveys and interactions with customer support associates
ActionIQ and retail TouchPoints
RECOMMENDED CONTENT

Customer Data Matters:
Supercharge Brand Loyalty With
Your Own Customer Data

How do CDPs leverage artificial intelligence (AI), machine learning (ML), and other predictive analytics?

CDPs use advanced analytics in four ways:

  1. Offering algorithmic identity resolution methods to deduplicate customer profiles
  2. Offering marketers prebuilt analytic models for improved speed-to-market (e.g. propensity models, churn models, look-alike models, clustering models)
  3. Offering integration of an organization’s homemade model scores for marketers to derive insights, perform segmentation, and scale personalization
  4. Offering algorithmic performance measurement methods that use statistical methodologies to measure campaign lift

CDPs do most or all of the hard work of preparing data for predictive analytics by gathering raw, disparate customer data into a single customer view. Marketers can instantly apply prebuilt predictive algorithms and data scientists can concentrate on building new algorithms, rather than 50+% of their time spent on wrangling data.

How do CDPs enable omnichannel customer experiences?

CDPs support omnichannel experiences in four ways:

  1. Data democracy. The CDP creates operational efficiency by ensuring customer experience stakeholders always have access to clean, up-to-date profiles.
  2. Data-driven insights. The CDP serves as the helpful brain, collecting customer interactions from all marketing, CX, and commerce systems for analyzing customers’ needs and predicting their next-best-actions.
  3. Journey design, orchestration, and measurement. The CDP empowers stakeholders to design omnichannel, fully-automated journeys that dynamically adjust the channel, content, and time that messages are sent based on customer preferences and campaign performance.

Omnichannel execution. Lastly, the CDP brings hundreds of real-time data connectors to activate these journeys, audiences, and offers to final-mile execution tools within marketing, commerce, and CX channels.

How do CDPs support personalization at scale?

CDPs provide marketers with a “personalization brain” that centralizes, democratizes, and automates the personalization process, thanks to the following capabilities:

  1. Automated data unification. Prebuilt data connectors ingest data from any channel and stitch it together into a persistent, single profile for each customer.
  2. Democratized data access. Business users gain self-service access to data for unimpeded customer analytics, predictive intelligence, and journey design for any and all channels.
  3. 1:1 offers and recommendations. CDPs include out-of-the-box analytic models to predict products, content, and offers that maximize business outcomes.
  4. Automated syndication. Prebuilt data connectors are provided (and vendor-maintained) for automated campaign execution across any and all channels.
  5. Automated measurement and reporting. Easy, business-friendly testing and measurement is included for rapid understanding of which offers resonate most with customers.

What are the main use cases of a CDP?

CDP use cases can be divided into three categories based on the value they deliver to the business:

  • Incremental revenue generation. By providing meaningful insight into customer behaviors and preferences, CDPs enable marketers to discover net-new opportunities and act on them quickly. Two common examples include: 1) Detecting at-risk customers and delivering automated messages to prevent churn; and 2) Identifying premier prospects via lookalike models and targeting them via channels like direct mail, digital advertising, or social.
  • Saving marketing costs. By calculating customer propensities via CDP marketing analytics, you can reduce marketing spend by targeting only customers with a high likelihood of conversion. Two common examples include: 1) Measuring discount sensitivity and minimize sending coupons to customers that don't display a need for them in order to buy; and 2) Driving more efficient prospecting by suppressing existing customers from prospecting campaigns across paid media and direct mail channels.
  • Increased operational efficiency. By automating data unification and empowering self-service operations, CDPs eliminate operational bottlenecks and reliance on expensive technical professionals. Two common examples include: 1) Creating a single-source-of-truth for customer data where employees can gain accurate customer insights without fears of data inaccuracies, and 2) Providing a user interface that enables scaled customer analytics, audience discovery, campaign list creation, campaign journey design, and campaign measurement.

9 Common Use Cases of a CDP

  1. Preventing Churn—Detecting at-risk customers and delivering automated messages across customers’ preferred channels to prevent churn
  2. Optimizing Prospecting—Identifying premier prospects via lookalike models and targeting them via channels like direct mail, digital advertising, or social
  3. Enabling Real-time PersonalizationEliminating one-size-fits-all campaigns and replacing them with a 1:1 marketing campaign that leverages an analytic perspective to include customers’ preferred products or preferred content
  4. Suppressing Customers—Suppressing customers in advertising channels who are already customers of their brand
  5. Eliminating Waste—Optimizing the frequency of touches to extinguish communications that don’t spark customers’ interests
  6. Modeling Propensities—Using customers’ propensities to target only those customers with a high likelihood of conversion for each type of content via segmentation (while also minimizing discounts to those who don’t need them in order to convert)
  7. Democratizing Data—Enabling business users to determine the size of an opportunity themselves, instead of requiring technical resources to write SQL against a data warehouse
  8. Eliminating Dependencies—Enabling business analysts to adjust data definitions and data formats without having to submit a change request ticket to IT
  9. Increasing Agility—Increasing productivity and time-to-market by enabling marketers to design tests, measure performance, and iterate campaign setups, all without having to wait for technical assistance.
Reduce Costs & Drive Results Across the Customer Lifecycle
RECOMMENDED CONTENT

Reduce Costs & Drive Results
Across the Customer Lifecycle

How do I build a business case for a CDP?

Besides marketing, CDPs have a significant impact on IT, analytics, and finance organizations. Buy-in from all parties is essential to success. Thankfully, the right CDP can be positioned as a win-win for all parties. 

Here are five key steps to getting buy-in for a CDP: 

  • Prioritize 7-10 specific use cases. Collaborate with stakeholders to identify the specific objectives they want a CDP to help achieve. Prioritize them based on potential business impact.
  • Demonstrate a CDP’s capabilities. Build familiarity and excitement for the solution by having two or three premier vendors demo their product against key use cases. 
  • Estimate ROI. Estimate the ROI of each use case by asking vendors about their typical return on investment for each case.  
  • Create a strong pitch. Build a pitch deck that is financially-focused and conveys how the CDP bridges the gap to achieving stakeholders’ goals. Include an appendix with supplemental info. 

Seek consensus. Present the deck to stakeholders one at a time so you can understand their perspectives, address any concerns, and optimize the pitch deck before presenting to the final decision-makers for sign-off.

What is the difference between a CDP and data infrastructure such as an Enterprise data warehouse (EDW), data lake, and/or big data platform?

In simple terms, an enterprise data warehouse (EDW), data lake, or big data platform is a repository for data, whereas a CDP is a widely-connected hub with a data repository at the center of it.

Many organizations with best-of-breed technology stacks have a data lake (inclusive of a big data platform) for enterprise data storage and analytics, while having a CDP to create a single customer view for marketing and for serving as the ommichannel brain across all touchpoints. As far as data sharing goes, EDWs, data lakes and big data platforms often send transaction data to CDPs. And in the same manner, CDPs often send their comprehensive customer profiles to EDWs/DLs/BDPs to enable greater precision and accuracy when performing enterprise-themed analyses.

How does a CDP fit my martech stack?

At a high level, a good CDP is designed to improve the effectiveness of the martech stack (aka marketing stack). In more technical terms, CDPs often sit in the middle of one’s martech stack, acting as the intelligent command center that receives information from siloed systems (e.g. data warehouses, MDM, loyalty, marketing cloud, website, POS, etc.) and orchestrates a customer journey based on a unified customer profile and next-best-action recommendations across downstream, customer-facing systems that deliver the final experiences to the customer (e.g. CRM, call center, POS, website, ESP, DSP, direct mail, apps, etc.). In a nutshell, it has the potential to be the marketing automation of the future!

For the benefit of the organization, a CDP system can enable a best-of-breed technology stack by being agnostic, i.e. seamlessly connecting to any source of customer data as well as any execution channel. This gives organizations the freedom and flexibility to plug and play innovative, best-in-class point solutions (such as a premier ESP or an innovative DSP) in place of expensive — and often outdated — marketing cloud suites that falsely claim to be great at everything.

six principles of a customer data platform
RECOMMENDED CONTENT

Six Principles to Anchor Your
Martech Strategy

How do you Distinguish an Enterprise Grade vs a Mid-Market CDP?

Enterprise CDPs must cater to a distinct set of requirements, most notably around grander customer scale, increased data flexibility, greater technology interoperation, and more stringent data privacy. Read more to understand the details of each.

Scalability

Fortune 500 organizations often have millions of customers (if not hundreds of millions). These companies often serve customers through multiple online and offline channels, meaning a CDP must be able to process petabytes of incoming data to become an effective business tool and marketing system. If a petabyte of data is hard to grasp, consider that one petabyte is equivalent to twenty million four-drawer filing cabinets full of text. This means your CDP needs enormous horsepower to compute all of that data during ingestion, segmentation, predictive analytics, and activation exercises.

Flexibility

  • Customer data platform software has the capability to ingest data in any format especially compared to a marketing stack utilizing packaged software that is not as highly customizable. This is necessary for two reasons: 1) the wide variety of data external systems and sources enterprises work with; and 2) the limited IT resources available to preconfigure them for insights and activation.
  • Ability to resolve identities, given the probability that customers will represent themselves with small variations any time they interact (different email, different phone)
  • Ability of marketers to configure data within the UI, eliminating dependency on technical resources

Connectivity

  • Ability to connect with any system, since F500s have multiple channels across sales, service, and marketing.
  • Ability to further organizations’ best-of-breed, hub and spoke strategies wherein they want best-in-class technologies at each corner of their tech stack, but with minimal overlap among them.

Privacy

F500s are held to higher levels of scrutiny, given that they are often in the public eye (as publicly-traded companies) and possess greater quantities of customers to protect. This means they must hold certifications in areas beyond the GDPR and CCPA, extending into SOC Type II compliance.

Most Critical Capabilities of a CDP

  1. 360-degree view of customer (including sophisticated identity resolution)
  2. Customer analytics (descriptive, diagnostic, and predictive)
  3. Experience orchestration (including automation of multichannel campaign customer journey)
  4. User interface providing business-friendly access (removes dependency on IT)
customer data platform evaluation guide
RECOMMENDED CONTENT

Evaluate a CDP based on its ability to
unify data, analyze your customers, and activate personalized experiences

How Long Does It Take to Implement a CDP?

While CDPs are foundational technologies in your martech stack, they should not take an enormous effort to implement. Instead, CDPs should take anywhere from 4 to 12 weeks once the ink is dry on the contract.

If a CDP vendor suggests it should take longer, they likely do not have the flexible system architecture that streamlines implementation. By contrast, leading vendors accomplish such rapid implementation via a load-first, define-later architecture, wherein data is ingested in its original source format—saving the need to involve IT resources for data transformation before ingestion—and then modified later by a data analyst through the UI when it’s needed for use.

What’s most important when implementing a CDP is the change management needed to ensure its effective adoption. Organizations are doomed to disappointment if they think they can ignore humans’ tendency to resist change. Strategically, organizations need to invest in process iteration, skill adaptation, and sponsor-led accountability in order to be successful.

CDP Change Management Best Practices
RECOMMENDED CONTENT

Step-by-step guide on Change Management

Should I Build or Buy My CDP?

Here are some of the main questions organizations face when considering to build or buy their CDP:

How will I handle my business’s unique needs and requirements?

Many of us believe our business has unique requirements that won’t be met by a packaged solution. In practice, every enterprise solution employs some mix of buy, configure, and build. To make a concrete assessment, start by mapping out your top priority use cases, and then ask your CDP vendor to demonstrate how they’ve delivered on those use cases in real-world customer implementations.

Which approach has the fastest time to value?

If there is a buy option for your CDP that can deliver on your priority use cases, it will undoubtedly get you to value faster. Your marketing teams will hit the ground running with an efficient UI/UX for business users that lets them quickly launch and iterate campaigns—all supported by a vendor and an existing user community. Your technical teams will be freed up to focus on value added work such as delivering on new ROI-laden business use cases, versus being heads down building foundational tech.

Building an enterprise-grade CDP can take 1-2 years (given the complexity of features). And because of the constantly changing nature of customer data, stakeholder needs, and the increasing quantity of execution channels, an in-house option is likely to result in: 

  • Slower time-to-market (given the complexity of solution)
  • Higher cost of ownership (given needed resources for maintenance and enhancement)
  • Higher risks of latency (given need to build and maintain APIs in-house)
  • Sub-optimal data democracy (unless a full-scale, non-technical UI is created)

By contrast, an organization can install a pre-built CDP in just two to three months, and at significantly lower costs and risks. While organizations cannot customize every specification, enterprise CDPs are engineered with flexibility of data storage, data structure, and data syndication. In addition, premium CDP vendors add value by constantly innovating new analytics and personalization features. These solutions can future-proof your investment, by ensuring that the CDP grows and adapts to changing technological and business challenges.

“Building software is hard—it always takes 2x longer than you expect. Organizations should buy speed where possible."

Nick Rockwell, SVP Engineering at Fastly, Former CTO of The New York Times

Which approach offers the lowest total cost of ownership (TCO)?

If you have deep in-house skills and experience, the costs of building may seem preferable to external vendor or integrator costs. When evaluating TCO, be sure to include these considerations in your assessment:

  • Costs of deployment as well as ongoing maintenance and enhancements
  • Ability to deliver and ongoingly update a high quality business-focused UI/UX
  • Readiness to add & maintain integrations to new channels and martech as needs arise
  • Opportunity costs of IT building and maintaining a CDP versus other initiatives

Can my existing data and analytics investments deliver on CDP requirements?

Your CDP should leverage all your existing data investments into actionable insight for your marketing channels. But your EDW or data lake or BI tool doesn't constitute a CDP. You need to have a data orchestration solution that can be operated by a marketing user and not an IT persona or a data scientist. And you need to be able to build and maintain for that requirement.

Overly focusing on the data and analytics elements, while underestimating the user experience—and the workflows and orchestration it empowers businesspeople to manage and deploy—is a common pitfall of the build approach.

I’m concerned about protecting my strategic data assets.

It’s your data and it must stay that way. Be sure to work with a partner who is a data processor—not a data originator. An enterprise-class CDP provider with a modern solution architecture will follow strict security and privacy standards.

customer data platform evaluation guide
RECOMMENDED CONTENT

Develop your key requirements for your CDP solution, and criteria for evaluating whether a particular vendor—or a homegrown solution—can meet them.

What is the History of CDPs?

Providing a 360-degree view of the customer has become increasingly important for marketers and the modern marketing system, most notably in the last 10 years as consumers engage through a growing number of channels, e.g. website, apps, call center, social, etc. A challenge every modern marketer must now face is the proliferation of channels (for example, the never-ending wave of new social media apps & sites), which means more interactions to keep track of and more customer records to resolve into a golden record. By creating a unified customer database, a marketing team can utilize this data to make a unified individual customer profile that spans all touch points.

CDPs arrived in 2015 to deliver the elusive 360-degree customer view. To appreciate what a CDP solution can offer, let’s review the technologies that attempted such a quest before them—and the ways they would inevitably fall short of providing the single customer view:

1990s

Enterprise Data Warehouse (EDW)

  • Biggest Weakness: Non-technical users could not access data (no user interface)
  • Sample Vendor: Teradata

Customer Relationship Management (CRM) tool

  • Biggest Weakness: Could not store data from all marketing , sales, and customer service channels
  • Sample Vendor: Salesforce Sales Cloud

Multichannel Campaign Management (MCCM) tool

  • Biggest Weakness: Could not store full depth of data, nor reduce reliance on IT for data and query iterations
  • Sample Vendor: IBM Unica

2000s

Email Service Providers (ESPs)

  • Biggest Weakness: Could not store depth or breadth of data
  • Sample Vendor: Salesforce ExactTarget

Master Data Management (MDM)

  • Biggest Weakness: Could not store data from all sources
  • Sample Vendor: IBM Infosphere

Customer Identity & Access Management (CIAM)

  • Biggest Weakness: Only stored ID’s, not engagement or transaction data
  • Sample Vendor: SAP Gigya

Data Management Platform (DMP)

  • Biggest Weakness: Could not store PII data, nor breadth of data sources
  • Sample Vendor: Lotame

2010s

Data Lakes

  • Biggest Weakness: Non-technical users could not access data (no UI)
  • Sample Vendor: Hortonworks

Multichannel Marketing Hubs (MMH)

  • Biggest Weakness: Could not store data from all sources (no scale), nor deduplicate customer records (no resolution)
  • Sample Vendor: Adobe Marketing Cloud

Digital Event Distribution Platforms (DEDP)

  • Biggest Weakness: Could not store data from all data sources nor deduplicate customer records
  • Sample Vendor: Segment
Scroll To Top