Jakub Jirsak, Getty Images/iStockphoto
Every now and then I have the honor and privilege to request and receive a guest post from a true thought leader and industry luminary. David Raab is the most influential thought leader in the world of marketing technology. There is no one more astute or more looked to than David. I took a look at something that he did on what he calls a customer data platform (CDP) and not only immediately saw its value but realized that it needs to be explained to a wide audience. David, graciously, agreed to my request for a post explaining it.
His formal bio is very straightforward. He is a consultant specializing in marketing technology and analytics. He is founder of the CDP Institute, a vendor-neutral organization that educates marketers and marketing technologists about customer data management and Customer Data Platforms.
That’s what he tells you. But he’s so much more than that.
Take it away David….
It wasn’t long ago that business marketers had little technology and even less data to help with their work. Today, marketers have literally thousands of software products vying for their attention and find data wrangling is one of their most important skills.
These changes have given marketers new opportunities. But they’ve also introduced a new challenge: unifying the data that’s spread among the many marketing systems. CRM and marketing automation products compete to be the master data store but neither is really designed to collect and expose data from all internal and external sources.
Enter the Customer Data Platform (CDP).
The CDP is defined as a marketer-controlled system that creates a unified, persistent customer database which is open to external access. In other words, it’s designed specifically to meet marketers’ data unification needs. CDPs started appearing several years ago but have only recently gained wide attention. A recent report from the CDP Institute found more than two dozen vendors totaling over $300 million revenue in 2016 and on track to reach $1 billion revenue by 2019.
What’s So Special About CDPs?
It’s understandable if your hype meter is starting to jiggle. But CDPs do have real advantages over alternative methods of assembling customer data.
- Unlike enterprise-wide data warehouses, CDPs serve only the marketing department. So they can be more nimble and responsive.
- Unlike integration platforms that route data directly between systems, CDPs store the data in a central database so it’s available whenever any system needs it.
- Unlike marketing suites or clouds, CDPs allow easy access to all systems rather than only suite components.
CDPs don’t solve all problems. Technical work is still needed to set them up and format data for specific applications. Incoming data often isn’t the quality you need. Application systems may not be able to use the CDP data. Departmental barriers, limited budgets, and inadequate staffing can foil the best technology.
Still, CDPs do offer some real benefits. Yet there’s one more obstacle to overcome. Marketers must realize that all CDPs are not the same. Products have not yet converged on a standard feature set and perhaps they never will. To make things even more confusing, there are no CDP Police (the CDP PD?) to keep people from using the label even if their products don’t qualify.
So let’s take a two-tiered approach to exploring CDPs. First we’ll define what any system must do to be considered a CDP. Then we’ll look at features that are present in some CDPs but not others.
The CDP Institute’s definition of a CDP has three conditions. These are:
Marketer-controlled system. This means marketers can determine what data the CDP will include, how it will be processed, and how it will be presented for access. “Control” doesn’t mean the marketers do the work themselves. It does mean that marketers can decide and have their decisions executed in hours, days, or weeks, not in months or years and only if someone else approves. Although “control” isn’t a technical feature, it does imply design and technology choices for CDP systems.
Unified, persistent database. This means the CDP must assemble data from multiple sources, associate data that relates to the same customer, and keep a permanent copy of that data. There are many ways to do this but if it doesn’t unify and store customer detailed data, it’s not a CDP.
Accessible by other systems. The CDP data must be available to other systems. The CDP doesn’t necessarily expose every bit of data it holds but it does provide substantial amounts of customer-level detail. Aggregated reports or customer lists without attributes are not enough. This requirement also means any system can access the data, not just systems controlled by the CDP vendor. This implies CDP data can be reached through standard query languages, documented APIs, or, at a minimum, tools to create file extracts.
The core CDP requirements can be met in many different ways. In addition, most CDPs do more than just the basics. To help clarify the options, here are some features you might want but can’t assume will be present in every CDP.
Anonymous audiences. CDPs are primarily designed to assemble data about known individuals, including personal identifiers such as name, phone number, and email address. But marketers might also want to build advertising audiences with profiles for anonymous entities such as cookies and mobile devices. Such profiles are typically managed in a Data Management Platform (DMP). Some CDPs include DMP functions and/or integrate with public DMPs like BlueKai or Krux.
Identity resolution. CDPs combine data from multiple sources into unified customer profiles. Sometimes all inputs carry a common identifier, such as a customer ID. In other cases, the CDP starts with different identifiers and decides which belong to the same person. Some CDPs have extensive built-in features to do this. Others have limited features, integrate with external identity resolution services, or don’t do it at all.
On-demand sources. CDPs usually ingest data from source systems and store it internally. This lets them manipulate the data and present it for access. But it doesn’t make sense to load some data, for example when someone else owns it or it changes very frequently. Some CDPs can easily look up such data on demand, merge it with internally-stored data, and present the combined result.
Real-time access. Nearly all CDPs allow access through file exports, API connections and database queries. But only some CDPs return results quickly enough to support real-time interactions such as responding to Web behavior or bidding on ad impressions. Depending on the channel, this might require response in as little as 30 nanoseconds. A second dimension is how quickly new data can be incorporated into the CDP database. Again, the required speed depends on the application. Some CDPs are engineered to be much faster than others.
Applications. The only job a CDP must perform is to assemble data. But many CDPs offer additional capabilities including segmentation, predictive models, revenue attribution, personalized message selection, or campaign management. These applications are sometimes the main reason a marketer purchases a CDP. So long as the CDP makes its data available other systems, there’s no reason it can’t offer a built-in application as well.
What You Should Do
These descriptions may give you some ideas for how a CDP could benefit your business. Typical applications include cross-channel customer identification, segmentation, offer selection, and marketing performance analysis. The next step is to figure out which of your current problems a CDP might solve and what new opportunities it would let you grasp. You can then work backwards to define what CDP features you need to support your goals. Only then can you start looking at CDPs to see which is the right one for you.
One place to start is the CDP Institute Library, a free resource with dozens of papers that describe what other people have done with CDPs, how CDPs work, and what features are needed to fulfill different requirements. Like any new technology, CDPs take some study before you can use them wisely. Take time to make that investment and you’ll be rewarded.
Video: Building customer loyalty is all about data
(Cross-posted @ ZDNet | Social CRM: The Conversation)