Customer Data Platforms Overview
Introduction
Customer Data Platforms(CDP) have recently received a lot of attention as they are becoming an essential part of the digital strategy for an enterprise. This article provides an introduction to CDPs to help demystify their functionality and usage. Basically, a CDP is a marketer-managed which creates a persistent, unified customer database that is accessible to other systems. In general, it is built or bought infrastructure operated by marketers who act on real-time signals that the customers provide to in-turn deliver delightful personalized experiences back to the same customers.
Marketer-managed means that marketers can operate a CDP with minimal assistance from engineering or other external vendors — a high degree of technical know-how is not required to perform common activities. Common activities include discovering customer segments, creating and forwarding segments and running A/B tests across channels.
A persistent, unified customer database is achieved by assimilating data from disparate systems, bringing it together in a single profile store and making these profiles available to all marketing, analysis and product personalization tools. Creating a single profile of a customer created from a collation of data from disparate systems requires the CDP to have advanced data ingestion capabilities that include data transformation and processing to create identity linkages.
The key benefits of a CDP are
- Efficient Operations: They help create a strategic customer data asset that can be managed and leveraged by both non-technical and technical professionals (marketing, analysts, product managers, data engineers).
- Data Governance & Compliance: With greater control over data, and through the creation of a ‘single source of truth’, a CDP mitigates some of the risks associated with adhering to data governance and auditing laws — e.g EU General Data Protection Regulation (GDPR) or CA Consumer Consumer Privacy Act(CCPA).
- Data-Driven Performance: Enables a unified view of customers and segments in all tools, including up-to-date static and dynamic attributes and predictive scores, to optimize spend and communication campaigns enabled through real-time operational KPIs/dashboards.
- Right Content to Right Customer at the Right Time: CDP coordinates campaigns in an omni-channel manner with maximum effectiveness. This enables orchestration of customer journeys as the customer transitions from one touchpoint to another by delivering consistent communications across all channels.
Types of CDP
CDP are often categorized into 3 types based on the capabilities they support.
1. Data Assembly CDP is the most basic form that collects customer data from disparate systems to create a unified persistent database. It collects data from multiple internal — web, mobile, CRM — and external source systems including social & paid media — e.g DMPs, Google, Facebook through 3rd party connectors — in different data formats — structured and unstructured.
Once the data is collected it is processed and transformed to create a clean single customer record. Identity mapping and resolution generally uses different match types like customerId , visitorid, address, phone, email, etc to disambiguate and link identities across different devices and channels.
Some CDPs also provide pre-processing and post-processing capabilities, for example to validate, clean, or transform incoming data or map data in a custom way to match destination formatting needs. In the profile creation process it enriches the profile attributes with the pre and post sign-in activity . Cross channel and device Ids are generally combined together in a process called identity stitching. Additionally it can join the 1st and 3rd party data sets to create a 360 view of the customer profile. The stitched unified profile is made available for consumption to all external systems through batch feeds, real time streaming, or Application Program Interfaces (API).
2. Analytical CDP builds upon the core capability of Data Assembly CDP and adds the ability to provide insight into customer behavior by tracking across channels. Additionally they provide capabilities for audience segmentation, data modeling, and have the ability to activate by sending segmented lists to marketing tools, like paid media platforms, social platforms, email, and DMPs. The disambiguated identity and profile is used to build deterministic and predictive audience segments and generate aggregate measures such as lifetime value and revenue. Reporting capability monitors the inflow and outflow of data and detects anomalies. Modern CDPs generally support a UI for creating segments of users to sync in real-time and/or manually export to other downstream tools (advertising, marketing, sales, campaign management, customer service, etc), and keep those segments up to date in real-time.
3. Customer Experience CDP builds upon the capabilities of Analytical CDP and adds decisioning and optimization. The decisioning capability embeds a decision engine that uses simple rules and/or advanced predictive ML models. The decision engine is generally supported through a UI tool for creating calculated attributes on profiles based on rules applied to incoming events, and keeping those calculated attributes up to date in real-time (e.g. number of emails sent, last sign-in, abandoned cart in last session, high-value customer, LTV). Some CDPs extend into machine learning to score interests and propensities of users. Most provide ways for data science platforms to receive changes, apply machine learning models as needed, and write back scores to the CDP profile.
- Personalization is driven by selecting the next best action which provides tailored messages, content, offers or product recommendations. This is enabled by leveraging the i) customer profile, ii) past behavior iii) transactions data iv) interaction data (impression, click and conversion) and v) current context to coordinate data and messages, instructing the delivery system with a simple message in social media to content in an email template to a view in native mobile app to a fully rendered personalized web page.
- Customer journey orchestration from Awareness all the way to Loyalty is achieved by consistently leveraging the unified decision engine irrespective of which stage the customer is in its lifecycle.
The following table enumerates the functional differences.
CDP vs DMP vs CRM
CDPs are often compared to Data Management Platforms(DMP) and Customer Relationship Management platforms(CRM).
Similarities — CDP and CRM are similar in the sense that they both provide access to customer profile data. CDP and DMP are similar when it comes to access to real time data.
Differences — The primary use case(s) for CDP is for Marketing while DMP is for Advertising and CRM is for Sales & Customer Service.
CDP vs DMP — CDP’s true power is to unify the customer data from all internal and external sources. Additionally, CDP has the ability to activate omni-channel campaigns. DMPs use aggregated (mostly third-party) data to target using anonymous cookies for advertising and retargeting. On the other hand CDPs use first-party data — including Personally Identifiable Information (PII) — enhanced with third-party data, to deliver personalized experiences across many marketing channels.
CDP vs CRM — A CRM stores known customer data and helps optimize a company’s interactions with its known customers — i.e. cannot identify prospects or unknown customers. A CDP, on the other hand, can not only identify customers but also prospects.
DMP, unlike CDP and CRM (row 2 vs row 4 below) primarily works with aggregated anonymous data.
To summarize, CDP consolidates the myriad customer data silos to stitch together a comprehensive view of a user’s journey. This helps drive a consistent and coherent way to activate & target audiences and orchestrate the customer journey across the lifecycle.
To summarize, CDP consolidates the myriad customer data silos to stitch together a comprehensive view of a user’s journey. This helps drive a consistent and coherent way to activate & target audiences and orchestrate the customer journey across the lifecycle.
Thanks Raffi Norian and Kiran Divvela for reviewing the article.Thanks Raffi Norian and Kiran Divvela for reviewing the article.
Originally published at https://www.linkedin.com.
