#053: Customer Data Platforms with Todd Belcher

Do you care about acquiring customers? Do you care about data? Do you like wearing shoes that have soles that are 2-3″ thick? Put those three things together and it means you care — or should care — about customer data platforms. On this episode, Todd Belcher from BlueConic joins us to explain what CDPs are and what they’re good for. Tune in to hear Todd masterfully steer clear of a sales pitch for his company…while Michael transitions on the fly from getting a basic understanding of CDPs…to installing BlueConic on this site…to pitching BlueConic himself!

Miscellany Mentioned on This Episode


Episode Transcript

00:00 Michael Helbling: Hi, everyone. Before we start the show, we have a couple of quick announcements. The first is that we’ll be doing a live recording of the podcast at Superweek in Hungary at the end of January. We would love to meet you there, so head on over to superweek.hu and give that a look. Next, we’ve got a couple of big announcements. As we kick off our third year of the podcast, we’ve got some exciting updates. We now have a website. You can find us at analyticshour.io. Tim informs that the “.io” means that we’re cool. And on that site we now have an official way for you to submit a topic you would like to have us cover on the show. Second, starting with episode 51, the one with Moe Kiss, we’re now posting transcriptions of each episode on our site. We don’t think you wanna read the show, but maybe it will help if there’s something you heard that you wanna go back and find. Anyways, we’re excited about what 2017 will bring. Check out superweek.hu, analyticshour.io, and now on with the show.


01:18 Announcer: Welcome to The Digital Analytics Power Hour. Tim, Michael and the occasional guest discussing digital analytics issues of the day. Find them on Facebook at facebook.com/analyticshour and their website, analyticshour.io. And now, The Digital Analytics Power Hour.

01:42 MH: Hi, everyone, welcome to The Digital Analytics Power Hour. This is episode 53, the first episode of 2017, but we won’t let that get us stressed out. It’s a brand new year, and we’re gonna start on something that probably not a lot of us are sure of. Have you ever not been sure about what something is but been afraid to admit it? Well, welcome to this episode of the show where we are going to talk about…

02:12 Tim Wilson: We’re gonna talk about JavaScript.

02:13 MH: Customer data platforms. [chuckle] Three words that sound like they go together, but do they really mean what we think they mean? This is shrouded in mystery and luckily we have a guest who can help us. Are you ready for this adventure Tim Wilson?

02:29 TW: I am ready.

02:31 MH: Okay. And are you ready for 2017?

02:33 TW: Easy there, let’s not push it.

02:35 MH: Okay. Well, let’s get Todd introduced. Todd Belcher is our guest, he is the Director of Customer Success at BlueConic, and prior to that he did digital analytics at Putnam Investments, and back in the mists of time, he was part of the Sane Solutions/UNICA/IBM tribe of analytics warriors. No, that’s probably not… Anyways, he is also a Certified Web Analyst and a certified Bostonian. Welcome to the show, Todd.

03:05 Todd Belcher: Thank you, Michael. Thank you, Tim. It is great to be here. And yes, a warrior.


03:10 MH: Which test was harder? The certification for being a Bostonian, or the certification for being a web analyst? Just curious.

03:20 TB: They’re just different. I’m going to call apples and oranges here.


03:27 MH: Once you park your car at Harvard yard, you are good to go, and that’s enormous… And I’m talking about the web analyst test, of course.

03:34 TB: Let’s just say in both situation, I look over my shoulder everytime I say I’m one of them.

03:38 MH: That’s right. [chuckle] You are one of us. Welcome to the show.

03:43 TB: Thank you, sirs.

03:44 MH: So to kick things off, how would you tell someone who’s never heard, of or doesn’t really understand what a customer data platform? How would you describe it to them? There’s a lot of different technologies that I think touch on it, but I’d love to start with your definition, we can go from there.

04:03 TB: Sure, and now, I’d like to start with activation of data. Basically, I think it’s a lot more than that. I think that David Rob, among some others, has done a really good job of trying to define what a customer data platform is. The word “Marketing” is often used, so being a marketer-controlled system where data can be activated across channels is probably the simplest and shortest way to try to define it.

04:32 TW: No buzz words in there. [laughter]

04:33 TB: Yeah, marketer.


04:37 MH: So, what does that mean? It’s a cross channel. And activating the data meaning you’re getting the cross-channel data with the customer-centric orientation of the data and you’re then making it available to other systems that actually do stuff with it?

05:00 TB: I think so, but at the same time, [chuckle] it’s about anonymous people, if you will, just as much as it is about known entities. It’s about associating those identities across devices when possible and trying to use the information that you either have or that you can collect to deliver a more relevant experience so that we hopefully make that whatever the experience is more effortless for the consumer, the user, whoever that might be.

05:31 MH: It sounds a lot like the promise of CRM since time in memorial, and what CRM’s claim… I guess except for the anonymous piece. Is that fair?

05:44 TB: I think in a way, the CRM, at least what it has become, is maybe a little bit more tied to offline channels and less capable of collecting data or playing nice with other platforms.

06:03 MH: Well, CRM doesn’t typically try to record your impressions or click-throughs on media. Would a Customer Data Platform attempt to track that? Would it go that far upstream? Is that, like a CRM in my mind generally, and maybe I’m thinking a little bit in Salesforce speak, but it starts with kind of a lead. They’re a known entity who has come to some owned property and has therefore been loaded into your system so you’ve got them in the space where you can mark it to them as a known entity whereas it sounds like a CDP as much, much broader. Or maybe the other, but the other extreme is a DMP, Data Measurement Platform, is kind of all about the media and audiences and they wanna stitch as much stuff together as they can but they’re not there to actually say, “This is what your sales team is gonna work.” Does the CDP fall somewhere in this weak spot that brings both of those together or no? Not so much?

07:15 TB: Yeah, yeah, and anyways, I guess what I’m laughing about at the moment is that we’re starting off the conversation trying to talk about DMP versus CDP, and not what the differences are. I think we could talk about that, but what instantly comes to my mind the way that you’re talking about this is at what point is it even not okay to have all this data in the same place, because with the Customer Data Platform, what we’re trying to do is we’re trying to associate an actual identity to the people who are anonymous users of your websites and mobile apps. So to take what we’re trying to do, and there are many different approaches to trying to do that on top of identity association. I think customer data platforms are… They’re typically geared towards the acquisition, for try to get someone who is anonymous to convert into lead.

08:12 TB: There are ways that you can help do that with data, whether it’s collected by your system, analyzed in another system and brought back, or just the anonymized data is used within your system to come up with recommendations, or to otherwise steer the messaging towards your persona so that you do become an identified entity. And so, yeah, I guess going back to why, I thought the question was funny, is I hadn’t really spent any time thinking about what happens when you take a person who’s been associated and identified as an anonymous user out on the web, and what ads are seeing, and what sites is all part of their anonymous profile that the DMPs hold. What happens when you have that and you mix that data with the CDP that is there to try to drive a real name and associate PII with that profile, I think, it’s just a really interesting conversation that I wasn’t planning on having.


09:14 TW: Tim stumbled right over that? What? No, I don’t know if that was on purpose or not, Tim, but that is a very interesting question and frankly it’s alarming. So we won’t talk about it. No, I’m just kidding. [laughter] Yeah, no, it was not intentional. I remain somewhat befuddled and that can probably rattle this off because BlueConic’s not one of the companies in play, but, with a client that’s got… It seems like every time they turn around they’re getting a new platform. They’re gonna load everybody into a CRM, and then next time you turn around, they’ve got a DMP and they bought the DMP, and then they’ve bought, what I realized I had to do another prep for this, was a CDP. And in that case sitting and looking at the CDP and the DMP I’m like, “You guys sound like you’re promising the same thing,” which really was, by the time it got translated to the business owners, they would say, the DMP would say, “No, no, no, we’re different from what the CDP is doing.” The CDP would say, “No, no, no, we’re different from what the DMP is doing.”

10:21 TW: But then, as soon as the vendors had left the room, and they were never in the room at the same time, the business owner would literally describe them exactly the same. “Oh, it’s so we can get audiences, and we can look at their behavior, and we can kinda track them across different things, and then we can target them with ads and other messaging,” and there was a total lack of clarity. And I fall into that a little bit too when I look at what people have, what organizations have implemented now in adding another layer. Which one do they wanna pursue? And I’m sure there’s somebody who has the full marked tech landscape totally internalized could say, “Ah, I see how all of these pieces fit together, but I am not one of the people who sees how all the pieces fit together and don’t wind up doing some form of land grab, or maybe at some form of prioritization, and just debating the problem that they’re solving is being a higher order problem and the problem that a different type of platform is solving.”

11:29 TB: It’s hard, you can’t really help it when you get on these subjects, unfortunately, especially when you pull out the acronyms. But the DMP, you can look at it as a specialization in identities for advertising where I think a CDP is more specializing for identities for your known customers. And, yes, maybe that includes a lot of anonymous data, but the point is to use that anonymous data to build an anonymous profile, that could one day be stitched with a known profile at that point that that user authenticates, and I think…

12:13 MH: But wait, known customers but anonymous data, and so it’s potentially known customers, everything that you’re collecting, you’re saying, “It may be anonymous now, but I have a pathway to where you would become known?”

12:27 TB: That’s right, and it’s an important distinction just because the CRM, I think you lose that information in a CRM. So, that’s one of the distinctions between those two platforms, is a CRM starts when that person becomes identified and has to start speaking with you based on whatever information they’re able to gather at that point. Where if connected with a CDP, the CRM has all the anonymous data that occurred prior to that point, already loaded up because there was a CDP helping with the anonymous data collection leading up to the lead generation.


13:03 MH: I think back to my Eloqua days of marketing automation, and they were touting the unknown to known. And in a sense, and even in Google Analytics saying that maybe that’s it’s a subset. They’re saying, “Sure, if you assign them a user ID and as long as you have built your system to give them a user ID, even if they are anonymous to you, once they become known, you’ve still got them tied to their pre-known state.” But is that inherently… Well you’re solving this one little one-off thing. If your CDP was what was providing that user ID that you were hooking in, then you’re kind of setting yourself up for a pull that from other different places and get a fuller picture when they become known. Is that it?

13:52 TB: No, I think that there are so many different things that CDPs are there to do that if you take any one of them there are going to be platforms that do them. And similarly, that’s how a lot of these things became part of a CDP. It starts with creating a massive user profile with data from multiple sources, and then because you have the profile, you’re trying to do more things with it, you just need to be able to segment those profiles and ship off. In real-time, some group of users out to other systems and then, next thing you know, you have essentially online testing and optimization that goes along with personalized messaging on sites. So, every platform, I think, has taken a different path and has different features, and I think it’s kind of the same thing as the land grab that you mentioned with other platforms that have existed for a long time, whether they’re CRM, or tag management, or other systems that are starting to think about persistent data or being able to target audiences, in the case of some of our optimization tools. There is a lot of bleeding over the lines in marketing technology but it’s driven by what customers need and sometimes there are customers who can just take the features that they get from one platform and replace a whole other system with that until they become… They reach a certain level of maturity, and then maybe they need a pure player in whatever that technology stack is.

15:26 TW: Alright. Do you have to be a customer to be in a CDP or no? You just have to maybe, gonna be a customer?


15:33 TB: Yeah, [chuckle] I think that’s the way that we look at it, right? It’s yeah, the anonymous users are maybe they’re our customers already and you just haven’t been able to identify that or…

15:45 TW: In a lot of sense it’s like, as marketers, we’ve been doing this kind of work for a long time in terms of trying to integrate, back-end customer data with sort of digital behavioral data and building logic, and stuff like that. We’ve just been doing it in different platforms of what’s now called the CDP. Is that fair to say?

16:03 TB: I guess, except that the CDP is activated by nature. So, whatever data is in the CDP, you have it on your websites…

16:13 TW: Gotcha, so, it’s basically now accessible. So, is this, I can now hook into this in all the different ways.

16:20 TB: Right, and I think there’s lots of different approaches, lots of different connections to different platforms as is always the case. But, I think, one of the things is just the personalization is bolted right on top of your website or your mobile apps and being able to use whatever data is flowing through the customer data platform for that.

16:36 TW: And are all of these considered to be a CDP? Do you need to be real-time or is some real-time and some not real-time? What’s the standard there?

16:46 TB: [chuckle] I would say, yes, you need to be real-time. I would say there probably are plenty of them out there but that there maybe aren’t. But yeah, I think real-time is a big part of it. As soon as you get information about a customer, whether that comes from an online or offline channel, if the information is hitting systems then other systems need to know about it. And so, yeah, the real-time is, in my mind, a necessity.

17:11 TW: Okay. Well, now I can see how to be pretty handy ’cause if you’re doing stuff on a website and that’s triggering “Mailers,” or whatever else I know about you or whatever, that’s all good ideas.

17:23 TB: Yeah, and a lot of the time, it’s about helping people take the next step and it’s not always about marketing, even though, yes, the customer data platform is mainly geared towards marketers. I think, what’s nice is that that also means it should be easy to use. Another thing I like to say is, even though we’re developed for marketers, we’re developed by developers, so, they are very smart people, data scientists, and programmers who think about theses things and so it’s not just that marketing is the only application, it just so happens we are very useful one, where you’re trying to get people from one step to the next and there’s really clear definition around what that is.

18:00 TW: I feel like we’re still being kinda abstract what’s a…

18:02 TB: Can’t help it.


18:04 TW: Well, I think we can build a scenario, probably.

18:10 MH: Well, yes. So build a scenario of what somebody has today and this is almost gonna sound like it’s asking you to give a BlueConic sales pitch which it’s not really… I don’t wanna get that specific but, what is the “These are the things I have today. This is the pain that I’m feeling or the opportunity that I’m missing?”

18:28 TB: Yeah. I think there are a few probably pretty decent ones. One is actually just folks who have marketing automation. They have campaign management systems and marketing automation around all kinds of [chuckle] offline data. And these things have been working for, say, 10 or 20 years, but somehow synchronizing the same campaign messaging online is still a problem for a lot of people. So that’s one simple one is just, you’ve got propensity modeling, you have segmented messaging that happens offline, whether it’s on call centers or it’s in an actual direct mail. But that same ability to target messaging doesn’t exist in online channels, so a customer data platform is there to make that easier. That’s one simple one, it’s not that’s every use case but…

19:24 TW: So you could say, “Okay, well I’ve got a campaign rule that’s gonna go and do something because I learned something about you in, I don’t know, wherever the data is coming from.” And then you have maybe a site targeting rule or something like that. So your tool, all it does is just make the profile or the data that I’m looking for available and I’m hooking into that with my other things like my campaign management tool, my site targeting tool, my bad words customizations, or segment building, or whatever?

19:58 TB: Yeah, I think on top of that though comes the real-time first party data collection, the actual collection of what’s happening right then and there. So whatever the propensity model is, or whatever the next best step for the customer is supposed to be, there are a couple of other things that go along with that. All that campaign management, that segmentation is all for known users, so it’s about taking the same segmentation, and mirroring what you did in the offline world in the online world with anonymous people so that you can still put relevant messages in front of people even though you don’t know who they are, and carry on similar campaigns until the point that they become known and then everything ties together on that end.

20:44 TW: In essence, instead of bringing in a customer ID, you’re building a customer ID that may or may not have a customer attached to it. So it could be like, “Hey, cookie 123 has visited us five times and never purchase so all I have is just their behavioral information but they match all of these characteristics or attributes that I defined at this segment. So now I’m gonna just keep putting them in this segment the next time they come to the website.”

21:10 TB: Exactly right. And so then they see different carousel sliders or they get different messaging about signing up for an account with some message that really hits home for that particular user group.

21:24 TW: But that being said until they go to a certain level of authentication, I can’t do things like trigger an email or send the mailer, or whatever, or I don’t know if I’m doing that right ’cause they haven’t identified all those behaviors as a customer. So you probably have a lot of logic that looks at people and says, “Oh it’s these two records, now I’m gonna merge them into one record ’cause all this anonymous data is now associated with this customer ID that I’ve got from before.”

21:50 TB: Exactly right. That’s where our buzz words, they’re our identities association. So taking some identifier that’s been acquired and merging in any other data that we have based on that identifier, or attached to that identifier in real-time. So within that same page view, or click, or whatever it is, you’re able to have a unified record of information.

22:14 MH: So is that… I have three. So one is, Michael, as you were describing it, I swear it sounded like you were describing the Adobe marketing cloud and the reason for the marketing cloud ID, that what it’s called whatever the…

22:29 TW: Yeah. Adobe has a marketing cloud visitor ID, so it’s trying to be their bridge over top of everything. And in a certain sense, it could be thought of as a customer data platform. However, there is something really key there which is worth noting, I’m sure Adobe loves that I’m gonna bring this up right now, but it takes about 24 hours for attributes or segments you create to be available on the other tools. So this is where that real-time piece that I asked about earlier makes a big difference potentially for marketers ’cause if I find out something new about you as a visitor today, the first time I can target content at you… So like I learned something about you via Adobe Analytics, then tomorrow, I can use Adobe Target to content target you but I couldn’t do it on your next visit 30 minutes from now.

23:19 MH: Got it. That was awesome. That was helpful. So then, next question is, how does the anonymous tie into the offline-online cross-channel linking? Aren’t those somewhat mutually exclusive that you can track them in multiple channels but until they have done something to provide the key… And that can take a, “Hey, you clicked through on this email and the email had a… ” Digitally, it’s easier to say, “You clicked through on the same from two different devices. I can stitch that together.” But is there some mechanism or concept that it is probabilistically trying to stitch them together, or when you’re making the links and saying this is the same person, offline activity, online activity is that inherently waiting until they’ve sort of somehow self-identified “I’m the same person?”

24:19 TB: I think by nature customer data platforms are deterministic. I couldn’t speak to the probabilistic capabilities of any given platform, but I could say that it’s only a matter of time and if it’s not happening now then, the probabilistic modeling will happen. I think it’s all about control, at that point, and being able to make sure that the marketer or who ever is controlling the customer data platform can really dictate what happens with that information. Does it stick with the customer profile forever? Is it a short term thing? I think there are some important rules that you want to associate with probabilistic association. For the most part, I deal with deterministic, so we get some identifier, like you said. It could be that somebody’s clicking through an email and that’s another decision point. Do you assume every time that somebody clicks on an email that’s the same person, or is it only that you assume that up until the point that they actually log in?

25:15 MH: And that was kind of the third thing that came up, as we were talking is that, I sort of had a flashback to my… And I think this is actually gonna be a similar answer to what, Michael your point was, I spent a period where we were digging into customer data integration and customer data management, and that did have a probabilistic component, but that was heavily, heavily not, not real-time, and this was big vendors, and I’m blanking on their names, trying to say, “Oh, wow, you have these seven contacts in your database.” But, I think, that was the challenge, these were all… We were dealing with our database, our customer database of known users, and like a lot of companies had multiple systems, and even within one system, because they had multiple phone numbers, they had multiple email addresses. We could look, we could manually look at a record and say, “Holy crap, we have seven records. This is clearly the same person. How do we kinda compress that?”

26:13 TW: This was probably ten years ago and it was messy as hell then. It was kind of a sub-genre of master data management. It was the customer’s side of master data management. It was absolutely not gonna be real-time. It was really just trying to get us to clean up our transactional systems, ultimately getting into the data warehouse. Now, what we’re talking about is like, “Oh, by the way, cast the net a hell of a lot broader, go to your anonymous stuff, and do it in a real-time,” but we’re gonna draw the barricade and say, “You know what, we’re gonna draw a line and say we’re gonna stick with deterministic.” We’re not gonna try to say, “Wow, you sure look like you’re probably the same person.” What level of false positive or false negative are we willing to take on? I’m not sure where that was going. I just realized that was another thing that… CDI, there was another acronym, CDM, customer data management, early precursors of this, but that was not at the activation level. Certainly wasn’t at the real-time or near real-time level, but it’s starting to get a little clear in my head where this fits.

27:19 MH: That’s good.


27:25 S?: Great show, thanks everybody.


27:29 MH: So, we talked about why Adobe would be different, and I think Google isn’t even casting as broad a number of… Their suite is not as broad, but Michael, you brought up Tealium, Insight has kind of said the same thing. Signal is kind of moving in that direction, as well.

27:51 TW: Correct me if I’m wrong, Todd, but there are some other pure play vendors in this space, like, yeah Tim, you mentioned Litix. Is treasured data considered in this space or TMM data? Or as I like to call them “Too much data?”

28:07 TB: No, there are several others in this space. Obviously, Scott Brinker’s marketing technology landscape is always a good reference point, but if you really look hard a it, some of the companies that you mentioned, like Insight and Tealium, I don’t believe are in our little neck of the woods there. So, I think it’s still fuzzy from a category perspective, because each platform has taken a different approach. Some are very mobile, some are very not mobile. Some work with both platforms just fine and then the features that have been bolted on top, as far as how do you execute on the data, is what I think is most important. It’s having the data and being able to activate that on the web, in your email program, especially for a lot of my peeps the digital marketers in the world, who have had an immense amount of data out there, but haven’t necessarily been able to create a simple segment of users, based on information that they have, and market to them differently. Whether that’s on the website or in an email. From a vendor’s standpoint, each one stands alone, in a way that the core, hopefully, the vendor that folks are choosing solves the problems that they need to solve. The customer data platform itself, I don’t think that you could point it out and say that there’s a very clear set of features other than that profile and being able to take data in and out in real-time from other systems.

29:40 TW: So, you collect data from, say websites and things like that, do you allow people to just bring over digital data? So, could I send you like all my Adobe Analytics or Google Analytics data?

29:52 TB: Right, and I guess where we see it probably more often, is with customer data, so you just don’t have if you’re retailer or publisher, never mind that you might not have people coming to your site out of the hundreds of thousands or millions of people you know. But even when they do come to the site, you don’t know it’s them. So there’s a lot of data that folks have that can be activated. The customer data platform can help build the segments. So the segments are built in real time. You’re pointing and clicking, defining a segment that you want to be able to market to an email or maybe in Facebook ads or Google ads that accept some a known identifier like an email address or a phone number. So you can actually use a customer data platform with data that’s coming out of CRM exclusively or point of sale system, or some other marketing database, customer database. And use the customer data platform with those profiles just like you’d use it with the anonymous profiles or profiles of users who have been to your digital properties.

31:06 TW: But is there in your case, analytics has an analytics tag. And I know every DMP says, “Tag the stuff.” Are you guys the same way? Do you have a… That is probably the preferred or the clean way is to say everything that can be tracked where your tracking technology can be there, that gives a direct pipe from action taken, information recorded, stored in the database. And then it’s kinda how broadly will that cover and then what are you supplementing that with that you can actually stitch together? Is that…

31:40 TB: I think that’s right. Yeah, if you’re purely online business, then you put your JavaScript on the websites, and your apps, and tie data together that way. Otherwise, you’ve got other systems that have customer data in them and you can bring that onboard and synchronize that with everything that’s happening online.

32:00 TW: Excellent. Well, I should probably just go ahead and let you know Tim I have installed the free version of BlueConic on our website, analyticshour.io while we’ve been talking? So they have a handy-dandy tag?


32:13 MH: Nicely done.

32:14 TB: Hey, I’m trying so hard here not to be the sales guy and you’re in the background installing. I appreciate it.

32:20 MH: You’re doing a great job, Todd. I figured, you know what, what’s the better way for us to get a little bit smarter about this? And then I went to your website and it’s like, “Oh, you can get a thousand profiles and look at them for free.” I was like, “That’s a great way for us to figure out…

32:34 S?: What are we gonna do about the people who come in the second hour of the day that you know you got this?

32:38 TW: There’s only one power hour. So I think what we’ll do is we’ll give a discount on downloads during that hour.


32:52 MH: So good.

32:54 TW: So it automatically already connected to our Google tag manager and universal analytics? I don’t really know how it did that, but it just said, “Yup.” Or it’s selecting those connections. Start at dimension one.


33:09 MH: This is not gonna be a real-time presentation.

33:13 TW: Start using BlueConic. Yeah, it took five minutes to set that up. And again, that’s not how an enterprise would do it, but I mean, it’s cool, Todd to me that you guys leave a version that people can actually go and start to use for free because that tells me two things. One is it tells me you guys are not afraid of people kicking the tires on your technology, which means that you’re confident in what you’ve built and that makes me excited ’cause it means maybe you’ve built something really cool. The second thing is it lets jokers like me pretend like I know what I’m talking about if I get in there and ‘futz’ around with it.

33:50 TW: But actually, that’s really good because a lot of people like junior people, analytics people who don’t have as much as many years in, a lot of times they want or need to access, or would like to learn about things, but not have access to those tools. So we all can get access to Google Analytics, but unless you have some way to get a hold of maybe Adobe Analytics or another paid tool, how do you get experience with it? So that’s pretty neat. So thanks to you, guys. I don’t know if other vendors do that. So maybe you’re just the same as everybody else. Let us know on our Facebook page or on our new website, analyticshour.io. [laughter] If you do, do that ’cause obviously, we’re not trying to prefer one vendor over the other, but it just so happens Todd is with BlueConic, so we’re talking a little bit about that product today.

34:45 TB: Yeah, I’m glad Michael signed up there. Hopefully you’re setting it up, Michael, you’re adding the Java script to the site. Now are we going to launch some user interactions? That’s what I’m looking for. We got to go live tonight.

34:56 TW: Well, yeah, so we’ve already got one visitor and now I’m looking at your listeners ’cause I was like, “Oh listeners, that’s what we have.” But I think you mean event listeners.

35:10 TB: Data collectors, yes.

35:12 TW: Yeah. So looking at this, it looks like you have another group of tags that I can add to specific events or activities that are… Or are these pre-configured? ’cause it…

35:25 TB: That’s right, there’s no change to the tag. I think that’s another CDP thing. Is the marketers supposed to be able to do it. I won’t say that all the listeners are 100%…

35:36 TW: Oh boy, we ran into this on a TMS space, don’t sell directly… Do sell directly to marketers, but be careful, we all stumble over ourselves saying we’ll never need to talk to IT again and just…

35:49 TB: No, yeah, exactly. I was just about to connect you… From an implementation perspective, our customers lean on us quite a bit to help set up the data collection because the problem is you just bought new software to do something, and you don’t need to spend a lot of time trying to learn how to collect data with it. It was built for marketers, but it’s still very technical, that’s why you have onboarding services or implementation services or whatever it is. So our customers can learn by us doing something, and then them seeing what we did, and us teaching them why we did it that way, as opposed to them just kind of having to [chuckle] learn everything before they can get data activated which is really the point, and what folks wanna do within weeks of signing up.

36:40 TW: No, we’ve got the tag on the site, and so now it is just up Tim and I to put together some strategies for what our segments are, and kind of divide…

36:50 MH: It’s great. We got the tool, and now we gotta go figure out what problem we’re trying to solve.

36:54 TW: Hey, sometimes things happen in that order, Tim, you just have to go with it.

37:00 MH: It’s been well over 45 episodes as we were accused of having a sales pitch, we need to be very clear that [chuckle] Todd is desperately trying to not [laughter]..

37:08 TW: This sales pitch is coming from Michael Helbling he’d like to not getting any money for this, but would like you to consider getting a free BlueConic account. And no one’s asked me to say that.


37:21 MH: We can change the topic anytime, but I’m also happy to keep using BlueConic, let’s just keep it up. Let’s get my face on your website, right now, if you’re logging in from North Attleboro, Massachusetts. [chuckle]

37:34 TW: Perfect. Oh, let me make sure I’m at my dashboard. Alright. I only see one visitor right now, but it is pretty cool Tim. Alright, we might have to cut out some of this.

37:47 MH: Can you add multiple users? The question: Am I gonna have access to this?

37:52 TB: Oh yeah, I think so. Oh yeah, up to two users on the free version.

37:55 MH: MH: Okay, perfect.

37:57 TW: Perfect. Yeah.


37:58 TW: I researched this ahead of time, Tim, no worries.

38:00 MH: MH: That’s good.

38:01 TW: Yeah, anyways, alright, I think we still had some other ideas or questions or… So we should probably get back to that. Sorry to distract.

38:12 MH: Well, I think for me this has been pretty clarifying, I feel like a lot of… I was walking and ready to play the three-letter record and bingo, and it’s a lot clearer to me where these fit. I’m sufficiently down on media that to me I’m like, “Why don’t you get a little closer to the people you can acquire that a CDP [38:34] ____?” If I’m gonna to use a CDP or a DMP, and I think Todd, you kind of called it out, it’s really different constituencies in an organization, that are gonna chase them. The challenge is when… ‘Cause DMPs want you to throw their tags on every interaction and transaction, and the TMSs that are moving into that world are saying, “Hey, drop our thing on it,” so there’s still a challenge. It seems like there’s got to still be an internal organizational challenge to say, “This is our way forward, we know this stuff is out there, we could bring it in, and then this is what we would specifically do with it,” not just we would have it or we’d be able to activate. We could say, “We know the volume is there. We know this is what we could do with it.” We know right now, we’re lookin’ at the data. We’re just lookin’ at the data 36 hours after it would’ve been really useful.

39:27 TB: I was gonna say, coming back to software for me wasn’t something I expected to do, I was a digital analytics person, I liked crunching data, I liked spending all the time transforming the data, [chuckle] and then coming out with something in the end. That was a process that I enjoyed, but at the same time there was and still is a lot of frustration amongst digital analytics people that may be we know things about the way that things happen and feel very strongly about some of the insights that have been produced. And, it’s very hard for an organization to take action on the information, so I think this is another avenue to help with it. It doesn’t solve the problem by itself, but I think it does give data another channel in the organization aside from reporting and presentations.

40:22 TW: While I am looking into this new tool that I got access to, which I always liked to learn new technologies, we should start to talk about our last call. So this is something we like to do on the show where we talk about things that are interesting. We think our audience will find interesting as well. Technically, BlueConic was not my last call, but I am finding interesting, at this moment, but Tim why don’t we start with you? What’s your last call this week?

40:50 MH: Sure, so this is one, I tweeted about it and posted it on Slack a month or two ago about it, but it’s so cool that it still… It will not leave me. It is a blog post from blog.data.gov.sg, which is basically the Singaporean Government I think called, “How the Circle Line Rogue Train Was Caught With Data,” and it’s basically some analysts or data scientists. There was an issue with the trains in Singapore where just randomly every day there were these random, different trains would have electrical issues caused by signal interference, and they could not figure out what the root cause was. It was happening to different trains, in different places, at different times of the day, but it was consistently happening every day. And this blog post actually just goes through where these analysts got the data from the transit organization, the Land Transport Authority, and crunched it and it walks through their iterations of how they ultimately got to what the root cause was, and it’s got some cool visualization, it really exposes their process. They supposedly did it in one day. Then they had to go out and actually validate it. It’s just a really cool article about how analysis is done well… They head down some sort of dead ends, but the dead ends kinda spark other ideas. It’s just a really nifty little read, so I recommend it: “How the Circle Line Rogue Train Was Caught With Data.”

42:27 TW: Nice. What about you Todd?

42:30 TB: From a reading standpoint on analytics, I’ve been enjoying where Stephan Hammill and Gary Angel have been lately. I’m just going to stick two thumbs up for two classics. I think that the radical analytics concept there is a good one of Mr. Hammill. And I think that customer data platforms and that idea of being radical and finding ways to solve problems differently is… I think they go together well. I think Gary also has, at least for me, I just caught him at eMetrics couple months back, and I think he had some things to say there that really spoke to some of the problems that customer data platforms are meant to solve. Those have been top of mind for me tonight.

43:25 TW: What’s funny is my last call has a Stephan Hammill connection too ’cause I believe it was something he posted or tweeted about, or recommended on Medium, I don’t know, but there’s an article called “The Data Science Delusion” on Medium, and the guy who wrote it Anand Ramanathan, I don’t know how to pronounce his last name, and I’m sorry, I don’t. Anyways, I read that recently, and really loved it because not only did he do a great job talking about data science, generally, he did a really great job breaking down why companies struggle to incorporate data science effectively, and some of the challenges on both sides that can occur. So I felt like it was reading that I wanted to go and send to every single person I’d ever talk to about data science to be like, “This is what I’ve been trying to say”, but he said it in such a good, solid way that I really was a fan of it. So, that was my last call.

44:24 MH: That’s funny he’s got a Venn diagram, I think I’ve seen the Venn diagram he’s got in another context that rift off of it.

44:31 TW: Oh really?

44:31 MH: Venn diagrams.

[background conversation]

44:37 TW: Alright, as you’ve been listening…

44:40 MH: Let’s not fuck Tim Wilson, oh sorry.

44:41 TW: Hey, [chuckle] I guess, that had to happened. [chuckle] You’ve been listening, and you have been interested, and you’ve been learning about what a customer data platform is. Or, you know so much more about it than what we’ve been able to cover because of Tim and I’s lack of knowing what it is. We would love to hear from you, and now there’s new and exciting ways that we can hear from you. We have a website, and you can comment on it should you want to.

45:13 MH: But we also have our Facebook page.

45:15 S?: Well they need to have what’s called an internet connection and a web browser, because now that we have a website.

45:23 TW: Yeah, or if you’d like us to post your comment for you, please send a self-addressed stamped envelope: [chuckle] Carson City, to PO Box 5424, Carson City, California, 94127. Okay that’s not a real address, please do not send things to that address. But it would be hilarious if you could send some kind of postcard and people would post it. Actually, there was a really popular website in the early part of the web, Post Secret, that literally did that. You could send things and they would post it for you. Anyways, we’d love to hear from you, both about customer data platforms or just about the show generally, and the best places to do that are our new website, the Facebook page, and the Measure Slack, and we’ll throw Twitter in there for good measure, but the real cool kids are hanging out in the Measure Slack, including our guest Todd Belcher. So Todd, thank you very much for coming on the show. Sorry that we could not provide a more intelligent platform, but I like the foundation we’re able to set. And hey, at least you have one more free user in me, and potentially Tim Wilson, so maybe not a complete loss in terms of your time.


46:42 TB: I totally appreciate it. I’m a big fan of you guys, and, anytime. We can have customer data platforms part two once you’ve had time to use BlueConic and you’ve got a better sense of it, be happy to do it.

46:58 TW: Well, now we just need to do some crossover validation, looking at different tools and stuff like that. But still, major props to BlueConic for having a free version out there that people can dig into and use, I love that. Lets people see what they’re getting themselves into before they cough up whatever this costs. What should a company ballpark figure be looking to spend to put a customer data platform together based on…

47:30 TB: I’d like to say a lot less than a DMP. That’s the way I like to think of it.

47:36 MH: There you go.

47:37 TB: And I’d also like to think that 2017 is the year of the customer data platform. You heard it here first.

47:49 TW: Usually, our predictions show is in February, Todd. But, thank you for putting this one out there right on the new year, great job.

47:58 TB: Couldn’t much help that…

48:00 TW: No, that’s great. You heard it here first. 2017, year of the customer data platform. In mobile, of course. Tim Wilson. [chuckle] Anyways, thanks so much for listening. For Tim, my co-host and I. Keep analyzing.


48:20 Announcer: Thanks for listening and don’t forget to join the conversation on Facebook, Twitter or Measure Slack Group. We welcome your comments and questions. Visit us on the web at analyticshour.io, facebook.com/analyticshour, or at analyticshour on Twitter.

48:40 Charles Barkley: Smart guys want to fit in so they made up a term called analytics. Analytics don’t work.

48:50 TB: Do I sound alright, by the way? I’ve had a cold. This is one of the first days I’m coming back. I feel like I have a voice at least.

48:57 MH: No, you still have a little bit of a lingering cold. Can you get that cleared up in the next two, three minutes?


49:02 TB: Yeah, yeah. I’ll see what I can do.

49:06 MH: Hi, everyone. Seriously? [chuckle] Your hand was up. I felt like I had to [49:13] ____. [chuckle] I also… [chuckle]

49:20 TB: Yes, I think I just needed to make sure all of my bodily functions were in order for you, Michael. We’re ready to go here.

49:27 MH: Totally okay. Don’t worry. You’re the star. Let’s do this. [chuckle]

49:34 S?: What’s up with guys named Todd?

49:39 TW: Are you a thought leader? Be careful how you answer that.

49:42 TB: No, I used to be. [chuckle] I’m a recovering.

49:48 TW: Spell my name with a lower case T and then an upper case… The third… The first M would be upper case. It would be like a lower case TI. Upper case M.

49:57 MH: Spelling your name [49:58] ____, okay?

49:58 TW: Oh yeah.

50:00 MH: ‘Cause you’re so good at R now.

50:04 MH: Ramble and ramble and ramble. Here’s a little thought leader tip for you. You don’t have to spend time, thinking about or working [50:14] ____ technologies to just hold forth on big opinion. That’s part of that. [laughter]

50:23 TW: Don’t you realize you’re fucking with the marketers when you guys do that?

50:29 MH: Yeah. That was sloppy on us but you did good, Todd.

50:35 TW: Rock, flag, and customer data!

5 Responses

  1. Todd Belcher says:

    Since I can… thought I would copy in some of my comments taken from a post I just put up on LinkedIn thanking Michael and Tim for the opportunity to be on the podcast, and for running the podcast in the first place… the whole piece is here: https://www.linkedin.com/pulse/2017-year-customer-data-platform-todd-belcher but the relevant stuff is here:

    Regarding collecting data for a CDP, around 31:45 I neglected to say, of course, that you implement in mobile apps with an SDK. This is not something you do with JavaScript. I know that. If you know that, I’m sorry.

    Regarding digital analytics being for reporting and presentations, around 40:00 I mentioned that CDP gives data another avenue inside your organization aside from presentations and reporting. I did not mean to oversimplify the great work that analysts do every day!! This perspective is from the digital marketers, who have reports and presentations, understand that there are things they could do to improve… but simply do not have the tools or ability.

    On data visualization, around 44:35 I missed a great opportunity to talk up our venn diagrams 😉 but really, talking up visualizations in a CDP is a great example of why getting into specifics about a CDP is tough. It is easy for the mind to get carried away and pigeonhole the entire platform into that one solution your mind is settling on. There’s always more.

  2. Ted says:

    Love the mix of product specific AND theory for this episode – too many digital analytics resources are too much product or too much theory – this was a great utilization of both! For me it becomes part of an actionable resource in my role. My only feedback – I’m in the industry and the acronyms were coming fast and furious in this episode – I had to go back and listen to it twice. The transcript was a great resource. Thank you for the great show – my dog Ferb has mailed his comments to the Carson City address – he’ll be excited to see it posted. Two paws up!

    • Tim Wilson says:

      Thanks for taking the time to comment! We do try to catch the acronyms and other buzzwords that get tossed out during the show when we’re editing and then provide links at the top of the show notes. But, point taken — that’s fine after the fact, but I can see how that might be a bit confusing in the moment. We’ll try to keep a more real-time ear out for that in the future and pause for a definition in the moment. Thanks for listening!

  3. […] almost 2019. The first “year of the CDP” was nearly two years ago. Every organization needs at least some combination of features […]

  4. […] customer data platforms (CDPs) — a topic we actually covered way back in January 2017 on episode #053 with Todd Belcher, who, at the time, was with CDP provider BlueConic. Since then, Todd left […]

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Have an Idea for an Upcoming Episode?

Recent Episodes

#243: Being Data-Driven: a Statistical Process Control Perspective with Cedric Chin

#243: Being Data-Driven: a Statistical Process Control Perspective with Cedric Chin

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Episode_243_-_Being_Data-Driven__a_Statistical_Process_Control_Perspective_with_Cedric_Chin.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares