A CDP Checklist - NO!

Abstract

This article is an all-encompassing look at Enterprise Data Management for Marketing Technologists that need a clear context for the components required to support robust marketing operations in a multi-dept, multi-faceted, and large enterprise. UPDATE: The EDM supports and feeds data to the CDP. The components below should, in many cases, be separated from the CDP. While some features below in the checklist are located in the CDP, many are more predominant as part of an EDM platform. Many vendors are abstracting their UI/UX and mixing metaphors of data management. While that is not wrong, it does cause more complexity for the marketer to understand specific data roles across the workflow. My argument premise is that the CDP provides a more narrow role. It does not house Identity resolution, dataset management, or even other more granular capabilities like AI/ML, Query, or Profile Management tools. My argument is it should supply the enforcement options for the profiles, what destinations they are subscribed to or assigned, and macro-audience management to coordinate omnichannel subscription and movement into the delivery tools that provide the assigned profile with the right content from broad-based to 1:1 narrow-based campaign initiatives. It's a 4-layer, perhaps even a 5-layer architecture from source to destination, horizontally, then it's 5-6 or more layers vertically at each bookend for the sources and the delivery destinations.

In an era driven by data and customer personalization, Marketing Technologists must comprehensively understand the tools and techniques that drive success. Below is what a marketing dept and marketing technologists need to look at to support their enterprise data management strategy, data strategy, and overall Martech stack. The analysis includes components like Customer Data Platforms (CDP) and downstream data apps like Campaign Management, Personalization, and Journey Management. Note that these apps are not an Enterprise Data Platform - they are mostly subscriber components of the data platform.

Many vendors will say their CDP is an Enterprise Data Management Platform, but this is far from the truth. There are only a few vendors in this genre - you will hear Reverse ETL, ELT, Zero-copy, Lifetime CDP, Universal Data Hub, et al...and many other confusing terms that cloud(pun haha) the reality of a real Enterprise Data Platform and strategy and what you must have and need to support your marketers.

Additionally, some companies(primarily pundits, CEO/CTOs of vendors, and consulting firms) must clarify what a CDP is and NOT blend it into an Enterprise Data Platform. A CDP is a small data app at the edge - it should always be distinct from your company's enterprise and Marketing Data Platform. A CDP is a component that complements an EDM, not replaces it, as many of these pundits allude to on their sites and collateral.

Adobe is one vendor that provides a multi-faceted, multi-source data platform like a Data Lake and Lake House without buying a CDP component, but they do have a CDP. Adobe makes the data platform an abstraction of many other technologies, so your company or a marketing technologist team doesn't have to go through all the processes of multi-vendor selection. Salesforce comes close, but they still need some primary capability today and are working on catching up. They have hired several people with these skill sets lately. There are a few others in this genre and many that are not. Buyer beware.

The technology abstractions the more prominent vendors are creating are freeing up massive implementation time. You still have to implement it's just a shorter time frame, and you have to commit deeply to the process but in bite-size steps to succeed. You could build a modern data platform yourself, as Adobe has done. For example, imagine having to hire staff to develop and provision all of the tools to create a self-hosted or your private cloud with a mix of solutions like Databricks or Snowflake, along with the periphery tools like Kafka, Airflow, dBT, Fivetran, and dozens more. I have the list if you are interested. You could still do this, and enterprises do, but there are benefits for the marketing department and the dedicated technologist assigned to them not to be in that business. It is faster for them to use these vendor-level abstractions, and then tie that system or platform of tools into a more significant transformation initiative within the enterprise. It's only an opinion, but it seems easier based on skillsets out there right now, time-to-market, and overall cost.

Do your due diligence on all vendors and map your requirements to what they can provide. The list below should give an excellent guide to what you need to ask and look for in your selection process. Many vendors offer all or some of these capabilities. Adobe and Salesforce have deep abstractions for them, Algonomy(a niche player for retailers and QSR-quick-serve restaurants), Amperity(deep identity resolution and analytics tools but lacks in other areas), Amplitude( approaching the Adobe/Salesforce level), and Segment and Tealium(mainly orchestration tools). There are dozens if not hundreds more(Acquia, RedPoint, TreasureData, Blueconic, SimonData et al...) - be careful and thoughtful based on need.

What does this ultimately look like? It could look like this, and much of the infrastructure to the LEFT of the Data apps could be abstracted to no UI screens or APIs, not seen, or touched at all by the user, or it could be completely exposed for more control by the client owner(you) with deep UI/UX and robust APIs. Both approaches are neither right nor wrong – depending on how much control the client(you) wants. The more control they want, the more staff and processes they must incorporate. Regardless of which approach per vendor you choose, you have opted to have them help you with managing devops or the core infrastructure vs building it out all on your own. Build vs. buy is always a common debate with larger enterprises, which is a huge task and would require even more staff and resources and huge amounts of time. I would opt for depth and breadth, lots of transparency to the infrastructure, and deep API-first approaches, but not so much that I need to manage it 100%. Adobe, Salesforce, and a few others are along this line, whereas other smaller CDP-only vendors only give you enough on the far right for a few capabilities within the CDP that would traditionally be EDM features in most cases. The biggest key is to ensure you have 100% of your historical data stored to take full advantage of your data.

Summary

Here's a summary list of what an Enterprise Data Platform may include, including the periphery or edge data apps that subscribe, interact, and consume the data from the platform. you could argue that these are independent micro-services within a larger platform, now called composable components or services, or they could be additional independent vendor tools integrated through APIs or a comprehensive monolithic single platform. The latter I do not recommend buying.

Data Integration Frameworks: The ability to seamlessly ingest structured and unstructured data, including large volumes of data, batch inputs, real-time data, and more, across various source channels.

Data Management: Encompassing everything from managing cookies and personal information to industry-specific data models, this part emphasizes adaptability, scalability, and long-term data retention.

Identity and Golden Customer Record: Ensuring customer data's cleanliness, verification, and probabilistic match, strongly focusing on maintaining a 'golden record' for individual customers and households.

Analytics: Providing tools for data exploration, predictive modeling, pre-built dashboards for various KPIs, advanced segmentation, and recommendation engines. An essential aspect for extracting actionable insights.

Reporting: Tools for customized data analysis, automated daily reports, and alert-based reports. Essential for a tailored understanding of business conditions or anomalies.

Advanced Segmentation and Predictive Modeling: Includes segmentation of customer personas, look-alike models, market basket analysis, churn prediction, and tools for micro-segmentation to complement macro-segmentation. These are key for targeted marketing strategies.

Personalization and Recommendations: Leveraging user behavior and contextual data to provide personalized content, offers, and recommendations across different channels and platforms. A vital aspect of modern customer engagement.

Customer Journey Management: Encompassing tools for journey mapping, real-time orchestration, cross-channel integration, and event-triggered actions. Essential for unified and optimized customer experiences.

Data Serving: Concentrating on timely and versatile data sharing with external systems, including real-time sharing, API-based connections, and analytical data set conversions.

Marketing Channels: A robust system for managing various marketing channels, including email, SMS, mobile app, web, and API support. The tools provided cater to modern marketing demands.

Data Security: Recognizing the critical importance of security, this section encompasses everything from data center selection to encryption, secure deletion of data, GDPR compliance, and operational security. Deep security controls, labels, published frameworks, and dedicated staffing ensure trust and compliance with legal and regulatory requirements.

By integrating these components, marketing technologists can build a sophisticated and responsive CDP that resonates with today's dynamic marketing environment. It allows a multi-faceted approach to customer engagement, personalization, segmentation, and more, all under a secure and efficient umbrella.

Data Integration Frameworks

Data Management

Identity Resolution and Golden Customer Record

Analytics

Reporting

Advanced Segmentation and Predictive Modeling

Personalization and Recommendations

Customer Journey Management

Data Serving

Marketing Channels

Data Security