Data is seemingly coming from every interaction, from every direction, from a growing number of channels, while using a growing number of devices… and it’s coming nonstop. This presents a great opportunity for companies to understand what customers and prospects need and what at levels unimaginable only a few short years ago. And while that’s sounds great, trying to pull all that data together in order to make the best use of the information coming from all that data is no easy task. In fact it’s incredibly difficult unless because most of the system businesses have been using over the years weren’t meant to specifically handle that task. Which is why Customer Data Platforms (CDP) were created and have grown of great importance over the last six years or so. But over those years there has been some apprehension and confusion about where CDPs fit and how they compare with other applications organizations depend on like CRM.
But it feels like some of that confusion and apprehension over what CDP is and what it does has been dissipating at an accelerated pace over the last year or so, as more interactions go digital. And to get more insight into the current state of CDP and how it’s being used in businesses today, I recently spoke with David Raab, founder and CEO of The CDP Institute, and to all in the industry the preeminent thought leader on the subject.
Below is an edited transcript of a portion of our conversation. Click on the embedded SoundCloud player.
David Raab: I think that people were very confused and unfortunately many people are still confused about how a CDP differs from a CRM or a DMP (data management platform) or a data warehouse or any of the other things. We already have a customer database, and in fact we have several, right? So what could possibly be different? But it’s now been established I think in most knowledgeable people’s minds, that none of those other systems does the same thing that a CDP does and none of them solve the problem that a CDP has, which you’re about to ask me what it is? So I’ll tell you …
The problem it solves is that all those other systems have customer data, but they have “their” customer data. CRM has its customer data, the web personalization system has its data, the eCommerce system has its data and it doesn’t come together anywhere. And that’s what the CDP does, is it brings it all together in one place. None of those other systems do that natively and architecturally. They’re all set up to do something else and they’re not really suited at the technical level to pull together all those different kinds of data and assemble from a customer view as opposed to some other view, and then share that all out, which is pretty much the definition of a CDP.
So, some of them might assemble a view, but not share it for example, because they just want to use it internally. CRMs are really bad, for example , at sharing data; and CRMs aren’t that great at pulling external data either. So, it took a while I think for people to realize this is a different thing and it’s worthwhile. The pain was there and the reason that CDPs took off and then grew steadily was because there were so many people, even before they ever heard what a CDP was, who were struggling with the fact that my data is all siloed, right?
We’ve been hearing about silos for decades and they thought their data warehouse would solve or the data lake would solve it. Although they pull in data from different places, they don’t pull it all in or they don’t prep it from a customer view. There’s again, just slightly nuances in what they do that a CDP does. Now there’s actually this particular set of requirements that nothing else does, and that’s what a CDP does. I’ll be optimistic, maybe people actually understand that a little better now or more people understand it.
Evolution of CDP Tech
Brent Leary: So, how has CDP technology evolved over the last two years. Has the technology itself changed because of what’s going on over the last two years?
David Raab: Well interestingly, the tech, the actual technology used by the tools hasn’t changed all that much. It’s a little different, but still a lot of this, Hadoop and all the big data structures and the elastic search and those, and those have all been around. The only real interesting technical change is things like Snowflake and Databricks have become way more prominent. And they’re very interesting technically because they can handle the mix of data types that a traditional relational database couldn’t do, but they can still take a SQL query like a traditional relational database. So they made it a little easier to build a CDP. Other than that, what’s changed is not so much the underlying technology that’s used, but the maturity of the applications that are built on that technology. So we have much better tools now for data quality management and governance and all the privacy related features that have become much more important are built increasingly into the CDPs.
Then the CDP vendors who go beyond just building profiles, which all CDPs do, but to also providing campaign management and analytics and sort of functionalities, application functionalities or activation functionalities. On top of the CDP, those guys have all had now five years to mature their products. They just keep getting better are richer and more advanced. So, it depends on where the vendor came from. A lot of these guys started out as say campaign managers, what I call marketing automation. Nobody calls them campaign managers except for me, cause I’m ancient. That’s what they were back in the day, they were campaign managers. You kids today don’t know. Some of them started like that, so they were already very good at journey orchestration. Others added that on and so, you always start simple and you get better.
Importance of Realtime Insights
Brent Leary: So, what are the most important pieces to the CDP puzzle today or components? And even… What are things that are still kind of misunderstood from a brand’s perspective, wanting to leverage what CDPs can bring them in terms of being able to understand their customers and interact and engage more, I guess in a more Real-time, with using the data, being able to interact in the right frame in order to connect? What are the things that are maybe big now that weren’t necessary five years ago?
David Raab: Well actually Real-time, I’m glad you mentioned that, because that has really cropped up. And I would say even in the last six months… I was just on a call yesterday with one of the vendors, I forget who, who they were saying, like most of their CDP applications or Real-time applications, that really surprised me. And we consider Real-time to be part of the six things that you have to do to be a CDP. But we actually added Real-time a year ago, because it was not part of our original list, but it’s now become so critical. Because if you want to manage the actual interactions, as opposed to just do analysis, I maybe push out audience lists to Facebook or to an email system or something. If you want to personalize their website interaction with the CDP data, you got to be able to read that CDP data in Real-time.
Or if you want to update your retargeting list in Real- time, based on what somebody just bought on the eCommerce site. Again, you need to do that in Real-time. So, Real-time has actually been a big change. That’s happened. Other features that have been added or that have been critical… I think there is a lot of interesting, I’m just going to call it ambiguity, about whether identity resolution belongs in the CDP or not, because we don’t think it does. I mean, somehow you got to do the identity resolution, but we think third-party systems can do that. And because so much identity resolution requires third party data, you almost have to use a third-party to do the ID resolution. So we don’t see that as a critical thing, but about half the CDPs have that capability.
And there’s certainly situations where you want more of these activation features. The industry’s almost splitting up into the enterprise systems that are just building a database and are assuming the data is going to be used by some external or other system for the website and for the marketing automation and so on. And then the ones that are used at marketing departments… So if I’m going to sell my CDP to marketing, not to the enterprise level, now the marketers want all the marketing features, right?
You can have a journey builder in there, that’s fabulous. You can have an email engine in there, even great, I can get rid of my email vendor. They like to buy from as few vendors as possible. So, those guys are enriching the product in that way. It’s almost two different industries evolving, serving two fundamentally different sets of requirements and even selling to different people.
Example of Good Use of CDP
David Raab: We are a game company and we need to be able to analyze how people are using our product. When the engineers put in the new feature that lets you, I don’t know, do somersaults or whatever of course they didn’t bother to instrument the system for that because they’re engineers and they just want to get the job done and go on to adding the next feature for the reverse somersault. So, then product [team] was like, “Well, what’d the somersault thing do?” But the CDP can just pull all that data in, if it’s a good CDP, it was the right CDP and extract that somersault feature – it’s an event – and do that analysis.
So, when you have something that specific, you need a system that’s going to let you do that. Those things come in and literally in days we hear about people getting specific value out of the system. So the more concrete you are in your requirements and more narrow, that’s a very narrow requirement. The bit more likely you are to get a good, quick result.
Can SMBs benefit from CDP
Brent Leary: So, generally speaking, would you say you have to be a certain size business? Is it mid-market, enterprise up? Or are there use cases where smaller businesses would benefit from having a CDP at their disposal?
David Raab: There are a couple of CDP vendors who focus on small business. What they’re usually going to do is they’re going to have prebuilt integrations with a platform like a Shopify, some common system that a lot of small businesses are going to use and usually in retail. But by and large, small businesses don’t have that much data. They tend to use one or two main systems. “I use Shopify plus, I use the marketing automation that integrates Shopify”, something like that. You don’t have a dozen data sources or 150 data sources like a big enterprise. So, there’s a little less value to the CDP and they are expensive.
The ones that are built for small business are priced appropriately, but the ones that are more generic and more for mid-market and enterprise level companies, those are the ones who have the complexity and the people and the budget and the resources to take advantage of the findings, right? Because, it’s fun to look at the findings, but unless you can do something with it, doesn’t do you any good. So, there’s a certain scale. A $100 million in revenue, something like that would probably be the lower end to use a more of a general purpose CDP.
Future of CDPs
Brent Leary: What role will CDPs be playing two years out, five years out?
David Raab: I think my achievement in life is that every enterprise architecture will have a box that says CDP and that’s like my legacy to the world. It’s, that little box, is going to be there kind of permanently. People now realize that yeah, you do need to pull all your customer data into some system that’s just designed for that purpose and nothing else. So, I don’t think that goes away. Now, is that box embedded inside some gigantic marketing suite or enterprise suite, an SAP type thing? Sure. Is it embedded in a industry specific system, a hospital management system or education management?
All those guys are actually adding CDPs and as long as they meet our criteria, as long as they will pull in external data and share the data out, that’s okay. If it’s in applications, we still consider that a CDP. There’s going to be a lot different ways of getting at that requirement for that unified customer. But I think the understanding that there is a need for unified customer data that comes from all sources and is shared out to all sources. I think that does not change and that’s the most important thing that we’ve learned in the last five years, and that hopefully we continue to do in the future.
This is part of the One-on-One Interview series with thought leaders. The transcript has been edited for publication. If it's an audio or video interview, click on the embedded player above, or subscribe via iTunes or via Stitcher.