In this episode of Customer Data Perspectives, host Isaac Sacolick, of StarCIO, sits down with Kim Davis, editorial director, MarTech, to explore how to get data-driven marketing activations right, attribution challenges, and what companies need to consider before implementing a CDP.
- What CDP Use Cases Should You Focus On?
- What is an Attribution Model?
- How to Integrate a CDP Into Your MarTech Stack
Read the Transcript
Sacolick: Hello, everyone and welcome to this episode of Customer Data Perspectives. Joining me today is Kim Davis. He’s the Editorial Director at MarTech. Hello Kim.
Davis: Hi Isaac, thank you for the invitation. Pleasure to be here.
Sacolick: Great, I’m so happy you’re here, because I love talking to people who’ve been immersed in marketing technology, pretty much your whole career. Do I have that right?
Davis: I’ve certainly been in marketing technology for a few years now. Before that, for a while I wrote generally about enterprise software. And if you go back far enough, I have all kinds of other journalistic odds and ends on my resume, from music journalism, to philosophy.
Sacolick: Excellent. Now, when we last chatted, you talked about the challenges marketers face in activating their customer data. And I thought it was interesting, because you talked a lot about centralizing, cleansing, normalizing, joining – all this stuff that as techies, know we have to deal with, and put technology and process in place around, what I call the “not fun stuff.” But let’s talk about what the marketers are really trying to get at at the other end of that funnel. What does it mean, when we say activations? What are they actually trying to accomplish there?
Davis: Well, I think the outcome they’re looking for is what is commonly called a single source of truth. It’s also sometimes called 360-degree view of the customer. But all that means in simple terms, is that you start from a position where you have customer data in all kinds of different locations, and silos throughout the company – maybe there’s sales data, marketing automation platform, there’s transaction data, customer success, call centers – it’s all over the place. And where you want to get to is that for one customer, say it’s Kim Davis, you know about his interactions with all those different parts of the business, and you want to try and get it all into one place so you can activate messaging that’s really relevant and personalized and shows that you know who the person is.
Understanding Marketing Activations
Sacolick: So Kim, you used the word activations, and that word really interests me. Can we talk about some examples of activations? What are marketers trying to do, and what’s similar – and maybe dissimilar – around activations in a B2B and a B2C world?
Davis: Very good. Well the kinds of things you want to activate. You could just use the word messaging, that’s a very simple word, which covers it all. But obviously, email campaigns, you want to deliver relevant email messages and preferably develop, deliver them at the right time when people are open to reading them. You want to participate in the great social media conversation, the worldwide conversation that’s going on as appropriate for your brand. You want to deliver these days you want to deliver push notifications, text messages.
The question which each brand has to figure out is, where are their customers? What channels are they in because the customer is going to be insistent that you meet them where they are. And that brings me on to your second point about B2B and B2C. I think the lines are blurring. We can all remember, it doesn’t seem very long ago when most of the B2B buyer’s journey involved direct contact between sales reps, meetings, you know, sitting across the desk from someone, as well as webinars, that kind of thing. But increasingly, the B2B buyer expects a frictionless and largely self-serve digital experience the same as they get from B2C. So they’re arranging now across all kinds of channels, they’re doing their own research, they’re talking to their peers. And you’re going to have to look at B2C techniques to reach B2B buyers these days.
How CDPs Can Vary By Capability
Sacolick: Yeah, I think that’s really smart. And I’ll just use a personal example there. I’m a B2B buyer, I’m a technologist, I’m an influencer on social media. On some channels, I will get called out from a marketer thinking that I’m a CIO for a large enterprise company, that’s not the case. And other cases, when I call up customer support on a technology issue, they’ll look at me just like any other person that they’re servicing, and give me the same rundown of clearing my browser and restarting my machine, not realizing that I’ve probably done all those things. And all that translates to what I amount to as a poor experience. So now we’re trying to bring all this data in. So if I’m a marketer, in support, I can look at this notion of a customer 360 as much of that data, knowing that it’s all over the place, I think, firms 10, 20, even up to several 100 different platforms that were collecting data in. And, and so you think CDPs are part of the answer on how to bring all this data in, but you also mentioned there are different kinds of CDPs. Can you elaborate on the kinds of CDP’s that are available?
Davis: Yeah, there are different kinds. There are the standalone CDPs, independent – the advantage there is that they’re agnostic as to what other systems you plug them into. They try and integrate with as many of the big and well-known other kinds of solutions, whether it be marketing automation, or CRM, that are out there. Then you also have the very big marketing clouds, I don’t need to name them, we know who they are. And in each case, there I witnessed this over a period of a couple of years, they each built their own CDPs internally. And there was a lot of debate about whether to call them CDPs, but the category was hot so then they went with that. And then interestingly, over the last couple of years, I’ve seen slightly smaller marketing suites, like your digital experience platforms, things which used to just call themselves commerce plus web content management, but then the DXPs, they’ve been acquiring CDPs. Some very well known and quite well-established CDPs have been picked up by the DXPs and integrated with their systems. So obviously, with the case with the big marketing clouds and the DXPs, you’re not buying an independent CDP, you’re buying part of a whole package, so it’s going to work. It’s going to have other integrations, sure. But it’s going to work primarily with the surrounding technology within that suite.
Determining Whether Your Business Is Ready For A CDP
Sacolick: Yeah. And so I think about this, because I’m an architect by nature, you know, I might have an e-commerce platform, a content management system, and maybe several CRMs. And all three of them are coming at me now and saying we have CDP capabilities. And my architect is coming to me and saying, maybe we should get a single pane of glass type of CDP, a horizontal that lives across all those. And maybe my data analyst is coming back and saying, well, I was looking at master data management technologies. And maybe I should be looking at this. Is that sort of the set of questions organizations have to sift through to figure out what’s their best play here?
Davis: Well, there’s really a whole list of questions you need to ask yourself, so it gets quite complex. But, uh, certainly one thing I’d say is that, I remember I came across an example of a very big brand, and they had ended up with more than one CDP, or the CDPs had slightly different capabilities, and they decided they needed this, they needed that. But then suddenly, your data is not all in one place anymore. You’ve got duplicate records in different places. You don’t want to go that way. I mean, if you’re thinking about Are you ready for a CDP? You have to ask yourself questions like, what’s the current state of our data? What’s the purpose, what do we want to be able to do with the customer data? And you’ve got to think about whether you’re really ready for it. Have you got the staffing? Have you got people who can operate it? And it may be for some businesses, if you’re working successfully with a combination of marketing, automation, and CRM, even though you’ve got duplicate records, you can kind of make that work and maybe you don’t yet need this, as you describe it, the one big layer which covers everything. But at some point If you really want to do personalized messaging, if you want to do it in real time at scale, you’re going to have to think about the CDP.
How CDPs Are Evolving Alongside Industry Needs
Sacolick: Yeah, I like what you said just now about figuring out what do you want to do with this thing, because we get so caught up with getting the data in one place, getting everybody to buy in on the technology, you know, figuring out where all the activation technologies need to come in. And then, you know, I started thinking about, am I going to do workflow out of my CDP? Am I going to do workflow out of my out of my marketing system? And somewhere in here, you know, where am I going to do all my data and analytics around this, which we’ll talk about soon. But you know, if you take a step back, and you say, we’re going to really look at using a mix of technologies, including a CDP, to solve the following problems, maybe we want to advertise better, maybe we want to improve messaging, maybe we want to do a full lead gen funnel for a B2B company. And these are the actual functions that we’re trying to use within the CDP, then it’s going to make your selection a lot easier. For your company and multiple CDPs. It really begs the question, is CDP overly generalized? And are we going to start seeing industry specific CDPs? Or are we going to more likely see a horizontal CDP get into industry specific specializations?
Davis: I would guess the latter if I had to bet money on it. But we are, I’m certainly seeing some of the enterprise CDPs out there developing specific offerings for specific verticals. And the two obvious ones which come to mind are healthcare, and financial services. And I think the reason is obvious with healthcare, you’ve got very high levels of compliance that’s required, you’re going to need some kind of HIPAA certification – I’m not an expert on that. And similarly, with financial services, compliance, regulation, confidentiality. So you can understand CDPs developing very specific capabilities for those kinds of verticals. But I’m talking about primarily, general CDPs, which can operate in a number of verticals. I’m sure there are some CDPs out there who only do this one vertical, but I haven’t seen so many of those.
Understanding CDP Use Cases
Sacolick: Yeah, I’m not gonna weigh in on that. But I understand the financial services, challenges around this, because, you know, they have historically a lot of data. I mean, financial services companies have been around for a long time, they generate their own data, they use a lot of third-party data. You know, they’re all using Bloomberg, they’re you somebody’s using Reuters, they have all these different data sets that come in. And you know, what comes out, isn’t going to necessarily be a one-size-fits-all. I might all be bringing Bloomberg data, but I’m going to use it in different ways. I’m going to normalize it in different ways. I’m going to say which parts are my customer profiles, and which are my segments a little bit different. Again, going back to what you said – what are you trying to do with the thing? What are you doing to get to the outcome? So even though the data is the same, maybe even elements of the data model is the same, but what I’m trying to do with it is a little bit different. So, go ahead.
Davis: Now, obviously, the thing to start off with is use cases, and probably not a huge number of them. Some specific use cases, which are well defined, and when you can also define what will count as success. And, you know, start small, try them out.
Data Ingestion, Implementation, and Prioritization
Sacolick: So let’s get into a little bit of the selection and practices around enabling a CDP to be successful. You know, I’m a marketer, I have a CRM, I’ve got a customer support system, I’ve got some operational and fulfillment systems. How do I think about whether or not, or how these CDPs can integrate with my systems in my data so I’m not spending a year or two getting all that work done upfront?
Davis: Yeah, and I have to say, you’re not exaggerating. I’ve spoken to brands who have spent at least a couple of years getting these things up and running. And even brands who are satisfied with that, thinking it could have been worse. But yeah, it comes back to the use case, as I was just talking about. You’ve got to decide what kinds of practical business outcomes you’re looking for, first of all. And then that’s going to guide you in terms of which sets of data you need to be focused on. You’re gonna have to ingest data into the CDP, especially if it’s one of these independent neutral standalones that we were talking about. So what do you need? Is it the data from your marketing automation system? Is it the data from your CRM? it depends on what kind of use cases you’re going to try, first of all. For a lot of people, they’ll eventually want data from their social media management platforms, but then that may not be high-priority. So I’m suggesting you don’t try and do it all at once. Don’t try and grab all your data from everywhere and push it into the CDP. Apart from anything else, going back to the not-fun task you refer to, the data you’re putting into the CDP, you do want it to be clean, and deduplicated, and all those other things. And maybe some of that’s going to be manual. And some of it’s going to take some time. But be guided by the use cases you’re planning to undertake. And that will give you some idea of the prioritization of the data which you need to be focused on.
Sacolick: I think that’s right. I mean, you need that inventory. But more importantly in what you said, is getting one level deeper and saying what data within this system is really important here. You know, the knee jerk reaction is, bring everything over, right? And that’s just going to pollute your CDP with a lot of noise, a lot of data to figure out. I think your question about priority is really important. A good CDP shouldn’t take a year or two to load in. But if I don’t need social media data to start off with for what I’m trying to accomplish, maybe I put that in tier two, and I load in second all.
So how do we get time-to-value because we’ve done a good job of prioritizing what we’re bringing in? And then ultimately, just to put on my technology hat on…there’s low-code to bring data in, there’s, I need a developer to bring data in, and there’s professional services that I need to bring data in, right? And CDPs are going to have one, two, or three of those options. And, you know, sometimes, professional services is what you want, you just don’t want to get your hands dirty at all. And other times, you want to be really flexible – your teams and marketers using spreadsheets and other data sources that they want to be able to load in themselves, and join that and create segments themselves.It’s not a one-size-fits-all to do all this. And so that’s why we have multiple platforms to select from. That’s why, go through the analysis to figure out what’s going to be most successful.
Davis: I want to point out there, and something which people should certainly look at when they’re considering CDP selection. And that is that some of the CDP vendors do offer professional services, and you want to know how extensive they are and what kind of support they’re going to give you. And then consider whether in addition to that, you need some kind of external implementation partner, which is obviously an additional cost, but it may be the way to go.
The Challenges of Marketing Attribution
Sacolick: Well look, it’s a trade off, right? Do you want to do it yourself? You’re paying for it. Do you want to take two years to do it? Then you’re paying for it. And you know, hopefully, the professional services that are being offered are going to give you some kind of acceleration and also some kind of support model for what you’re doing.
I want to switch gears here. Now, you have, obviously, a lot of writing on this, and some of your writing really caught my attention. You have an article about attribution, right? And which, of course, I want you to define, I’ll supplement it. But what is really realistic around attribution and what’s still hard to do? So first define it. And let’s talk about – what are the easy stuff everybody should be able to do, and what are some of the things that everybody’s saying we should be doing, but it’s still truly hard to do?
Davis: There’s a big debate going on about this, it’s been going on for some time now. Attribution I simply take as being able to identify the action that was taken by the marketing and/or sales organization which caused the conversion. Now, back in the old days, like only a couple of channels, not so hard. And obviously, the the aim of attribution is to be able to do some kind of predictive modeling: showing this ad on this TV show, at this time of day, causes this uptick in conversions. We predict if we do the same again, we’ll see the same result. Unfortunately, the world got really complicated over the last five to 10 years. The number of touch points on a customer journey now, is bewildering. Online, offline. Apps, web advertisements, social media, peer reviews, I mean, customers are getting information from all those different places. And how on earth can you tell which of those touch points was really the tipping point for conversion?
There are all kinds of models out there. There are first-touch models, which really look at how you originally grabbed the customer’s attention. There are last-touch models, which focus on the final actions the customer took. There are multi-touch models. And I think a lot of marketers would say you have to adopt some kind of modeling of this kind, otherwise, you’re not going to be able to support your budgeting and planning for the future. You need some to be able to show some kind of understanding of what’s working. But there’s another more radical view, which says that in today’s exceptionally complex world, that predictive modeling doesn’t work anymore. The best you can do is to have a real-time understanding of what customers are doing at any moment – where they are, what touch points are touching them or affecting them, and be able in real time to optimize. So in other words, you’re starting off with an outcome you desire, but you’re not predicting the whole journey to get there. You’re not predicting the whole campaign. You’re adjusting as you go along in response to what the customers are doing. Now, that’s a great model, in theory. You’re going to need some really good software and some really good processes to be able to do it.
Sacolick: I think the point of trying to be completely causal in your attribution model is a fallacy. You know, even in the completely digital world, if you just think about a subscription website trying to figure out what’s driving churn, or what’s driving the subscription rates, or an e-commerce site – there’s a completely digital experience going on there. You have this notion that everything should be measurable and easy to do. And then you find out, there’s one link in your platforms that doesn’t collect the data in a way that you can connect the dots on very easily. And then of course, you start getting into real-world situations, right?
So I’m walking into the store, I’m walking into my bank, I walked into the hospital, you know, how realistic is it that you’re going to be able to connect the data, my identity with what’s happening in the digital world? So you got the whole real world scenario on top of that. And then one layer on top of that is, you know, I’m sitting at a bar, and somebody recommends something to me. You don’t have any visibility around something around that. But at the same time, I feel the pain point of the marketer – there’s a bunch of dollars coming in, they have a certain amount of money to spend on their marketing activities, they’re charged with bringing leads in. They can’t just look at macro conditions, revenues up, number of customers or market shares up. They need some way of connecting the dots and telling a story around that. And that’s what attribution is trying to do. Do I have that right?
Davis: I think that’s absolutely right. And it’s that ancient comment, isn’t it? I know, half of my marketing dollars are wasted. I just don’t know which half. And it’s not that bad anymore. We have better understanding now, especially in the digital world. There is a lot of data we can collect, even though as you say, it’s never going to be complete.
Connecting Data Across the Ecosystem
Sacolick: Yeah. And I think one of the things that’s worked in our favor, if you get good at it, is connecting the support functions backward. Right? Everybody is trying to get the subscription models. Everybody is trying to get loyalty and repeat customers. Somewhere in this experience, I’m probably having an interface with a brand after I bought the product in some way. And this is a big story of the CDP right? Connect the forward picture with the backward picture and say, how do we improve the experience? Do I have that right as well?
Davis: That’s an absolutely critical point, you’ve touched on there. Because if you’re just looking at marketing and sales, you’re really not getting a rich picture of what’s going on. You absolutely want to look at customer success if it’s B2B customer service and call centers. you need that data, and you need to be able to connect it back. And additionally, something I’ve seen is some CDPs are now selling into customer success organizations. And they’re not just selling to the CMO anymore.
Sacolick: Yeah. And I think that’s ultimately what this comes down to is when we talk about a CDP, in this notion of 360, we’re not just talking about the data and trying to bring it in one place, we’re trying to bring the entire ecosystem to have access to that data to be able to use that data in some place. So when I do call up customer support, they know something about me, they know the products that I bought, they know that I’m an influencer, they know I’m a technologist, and they’re going to maybe even route my support call to somebody who can identify with me a little bit better, and respond a little bit better than somebody else coming through. Do I have that right too?
Davis: Well, the word challenges the word I’d emphasize there, because we all as individual consumers know how brands really struggle to do this. For every seamless, perfect customer experience I have, I have goodness knows how many where it’s clear that a brand I’ve been dealing with for years has no idea who I am or what I’ve done, or what my journey has been. It’s tough to do. But there’s not much choice because customers really expect you to make an effort.
Data Privacy, Security and Governance
Sacolick: Let’s just cover one more nuance to this, obviously not a nuance, I want to talk about compliance and security around all of this data. What are you seeing? What are you writing about in terms of how companies can do this well? Bring all this data in this one place, and not end up with a security gap that exposes all that data?
Davis: Yeah, I think most of the CDPs I look at, certainly the ones in the big marketing clouds, they have this built in, or at least, they will tell you they have it built in. Compliance, security comes along with the package. In a sense, it’s something to be concerned about, but the fact is, your data already sits on a whole bunch of different platforms, in a whole bunch of different places. And there’s no reason to think that the CDP is going to be more vulnerable. It takes me back to the early days of the cloud, when people were saying, I’d like to move my data into the cloud, but I’m worried about security, I’m worried about it leaking out. It turns out the cloud was pretty much safe among a lot of the on premise solutions they were using.
Sacolick: Yeah, I agree with you there. And I think the vendors coming back and saying we have itt baked in is maybe 20 percent of the journey. I think the other 80 percent is creating your policies around it. It’s instrumenting it. I mean, you have an advantage when you tell your organization that we’ve centralized 30, 50, 70, 90 percent of the data, now you have a central place where you can manage it, you can decide who gets access to it, you can define what’s an acceptable use of that data, what’s an unacceptable use of that data. If it’s centrally located, it also means I am less likely needing to pull data myself from three or four places and put it into spreadsheets. Chances are, it’s already done there for me, and I don’t need to pull it into separate tools. So you have, between the vendors’ implementation and your centralization of data, the opportunity to finally get this right. But it’s not going to come for free, right? You need to actually take this step and say, okay, as I bring my data source in, how do I define policies around this? Who gets access to it? What’s acceptable use? And then you start getting to, I have a a more compliant system. Do I have that right too?
Davis: Yeah, and of course, the vendor is going to tell you, though, don’t worry, it’s baked in, we’re handling it, You want to have a whole list of questions to ask them so you can understand how they’re handling it, exactly what they’re doing, what they’re compliant with, how they’re ensuring that compliance. You’re gonna have to test them.
Sacolick: Yeah, I don’t think the CEO is going to be able to take the stand and say, Oh, it’s vendor X’s problem unless they can really prove that. Kim, this has been a great conversation. We’ve covered a lot of ground around CDPs, around activations, around attribution. I’m going to ask a question, I asked every panelist who’s come on and been on the show for Customer Data Perspectives. What’s your wish list, Easier button for gaining competitive advantage with customer data? I mean, we’ve been talking customer-360 for at least a decade. So what needs to get easier for everybody?
If You Had an ‘Easy Button’ for Customer Data, What Would it Be?
Davis: Well, I’ll tell you something I thought of, I’m trying to surprise you. I think people need to think about zero-party data. Now, let’s, that’s a buzzword which is around. But I think there’s some genuinely interesting things about it. It’s a bit of a misnomer because it’s really first-party data of a certain kind. In other words, this is data which you’re going to be collecting from your customers, your users, your visitors, which they’re giving you, consensually. Now, obviously, if that’s an email address, or a phone number, name, something like that, that’s classic first-party data. What zero-party data is, is data offered in that way, but which tells you something about their interests, their activities, what they’re doing – gives you something actionable.
And I saw some good examples of this. So there was one e-commerce site, which when a new customer came there, a new visitor there unknown, they don’t know anything about them. And rather than trying to find out who they are, collect the demographic data, what they did instead was have a pop-up quiz, and you can incentivize it. But a fun, quick quiz, three or four questions about what the customer is interested in. And suddenly, even though you don’t yet know who they are, you start to know what they’re looking for, what they’re interested in, and you can start to personalize their experience. I think, especially with third-party cookies, famously being deprecated, we’re going to see a lot more interest in zero-party data. I think that what that is going to make easier is starting to deliver a personalized experience very early in the customer journey. In other words, don’t be obsessed with who they are, find out what they’re interested in.
Sacolick: You know, I love that because we go back to, you know, talk about financial services and trying to figure out, you know, I’m a small business owner, I’m interested in managing my small business. Or I walk into the hospital, I’m a runner, I’m a biker, right? I’m interested in proactive ways of taking care of my health. You don’t need to know much else about me, you know how to serve me there. And that’s ultimately why I’m coming into your establishment. I’m trying to research, I’m trying to learn. I have a screen this big and this amount of time. And all this technology is there to say – what’s the best thing I could put in front of them that gains loyalty that gains buyers that gains interest? And ultimately, that’s what customer data perspectives and platforms are all about. Kim, any last words? I think this has been great.
Davis:I think just to repeat what I said, know what your customers are interested in. And don’t be obsessed with finding their postal address and their email address and their phone number.
Sacolick: This has been a great talk with Kim Davis of MarTech, and thank you for joining Customer Data Perspectives today. Look forward to seeing you in our next episodes. Have a great day.