#041: The Productization of Customer Intelligence with Blair Reeves

The following is a straight-up machine translation. It has not been human-reviewed or human-corrected. However, we did replace the original transcription, produced in 2017, with an updated one produced using OpenAI’s WhisperX in 2025, which, trust us, is much, much better than the original. Still, we apologize on behalf of the machines for any text that winds up being incorrect, nonsensical, or offensive. We have asked the machine to do better, but it simply responds with, “I’m sorry, Dave. I’m afraid I can’t do that.”

00:00:04.00 [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 forward slash analytics hour. And now, the Digital Analytics Power Hour.

00:00:25.03 [Michael Helbling]: Hi everyone, welcome to the Digital Analytics Power Hour. This is Episode 41. Hey, Tim Wilson. Hey, Michael Helbling. You know, I’ve been thinking about the customer lately. And, you know, to really understand them, you’ve got to break it down into the two words that make that up, cuss and tumor. But no, seriously. you know, in the world of customer analytics and customer intelligence, lately everybody’s moving into this space and trying to productize it, but who’s really getting it right? You know, customer analytics is kind of a big bear, and it’s got a lot of moving parts. So we wanted to talk about that on this episode. It was something we thought we might need an expert opinion on. So with us is our guest, Blair Reeves. He is currently the principal product manager for SAS Customer Intelligence, and you probably remember him from his time at Demandware and IBM as well. He’s got a lot of great opinions, and we’re excited to have him. Welcome, Blair.

00:01:27.92 [Blair Reeves]: Thanks a lot. Thanks a lot. And I’m sorry your expert didn’t work out for you, but I’m happy to fill in at a pinch.

00:01:33.63 [Michael Helbling]: See? That’s the kind of can-do attitude that’s going to help us figure this out today. All right. So maybe first. Maybe we should define what we’re talking about in terms of customer intelligence or customer analytics.

00:01:48.15 [Blair Reeves]: Yeah. So I mean, I sort of think of it kind of conceptually, right? So you think about who the customer is. So who the customer used to be, used to involve way back in the battle days. It was just, you know, your analog days is just the guy who came into your store or whatever. Then it became kind of, you know, your online visitor and then different companies start splintering and having different kind of views of that, right? So you have your, you know, I’m sure Adobe’s of the world is sort of thinking this in a web analytics sort of way, which, you know, the person who’s on the internet, whose behavior you’re going to analyze, you know, that’s really, you know, who your customer is. And then you can tie that back to whatever data you have about him or her offline. Then Salesforce is kind of the different view of this and the whole CRM sort of view of that around who they are in terms of their information, transactions, that sort of way. All the e-commerce companies, and this is one thing that I got to, you hear a lot about when you work in e-commerce, like for the platform companies, to Manware, Hybris, even WebSphere, ATG, those folks, they think about who that customer is. pretty differently than the Martech people do. How so? They think about customers. They don’t really talk about visitors. If you’re using an e-commerce platform, you care about your visitors, your unconverted visitors, of course, but you care a lot more about making transactions run. I tell you that the e-commerce platforms see the marketing technology platforms as areas to colonize, to subsume. And the marketing platforms think about the e-commerce platforms the same way. They think of it as an add to cart widget, essentially. There’s a whole kind of battle going on there, but again, whole other kind of discussion.

00:03:31.28 [Tim Wilson]: But it’s part of that where the data actually lives. I mean, is that similar, you know, web analytics platforms that say, sure, we integrate with lots of stuff, we’ll pull other people’s data in, but not necessarily as easy to get our data out? Or is that not how the battle manifests itself?

00:03:45.65 [Blair Reeves]: Yeah, you think about it. And, you know, if you’re an e-commerce platform, you’ll naturally expand to capture all parts of that customer’s journey, customer’s, you know, interactions on the site. And that kind of very quickly gets into where your marketing clouds are going. And the marketing clouds would like to say, well, we already have on-site, off-site, email transaction, all your different kind of channel interaction histories as well. We’d like to make e-commerce a one more component of that. So there’s all of that. And then you have the kind of traditional Data science kind of you know systems, right? So you have the people who’ve been working particularly in like telco or financial services with customer information customer data warehouses doing What we now call big data used to just be data using any number of you know bi tools for that and those are all customers into customer intelligence But they’re all kind of have their origins and their strengths in different areas and what we’re I think we’re seeing now is a convergence of all those things and no one’s really there yet and what we’re trying to figure out is how we productize a lot of the kind of dirty work that goes on there.

00:04:59.36 [Tim Wilson]: So is that but every one of all of those are it’s not really moving up market or down market you’re saying is everybody’s expanding everybody generally has a sense that they could be the central spot for the entire journey from pre-awareness all the way through acquisition and loyalty and tracking that person.

00:05:20.18 [Blair Reeves]: Yeah, so I think some certainly do. So you have the consolidators out there, right? You have your IBM, Oracle, Adobe, Salesforce. A lot of them, to some extent or another, are offering you a consolidated package. So you take this platform and we will do all parts of your marketing stack. And that works for you, that’s great. For most companies, that is not the case, right? I mean, Scott Brinker talks about this all the time, but the future of most marketing stacks, even in the enterprise kind of tier, is using VestaBread, modular technologies, right? In reality, either approach necessitates, you’re going to have a whole lot of data from different channels, much of which does not actually integrate together either at all, if it’s VestaBread, or very easily, even if it’s a consolidator, because all the consolidated clouds are made up of modular technologies that are mostly acquired and have integrated post facto, right? So there’s a real tension there. And then, of course, you have all the historical data as well. So right now, there’s a whole breed of data scientists out there. There’s a whole constellation of great consulting firms. And they’re working with very large companies to say, we have a enormous sums of data and we have to put it together and derive some sort of value out of it. And to some extent, they used to use the traditional BI tools for this, but it’s becoming with more digital data and more hosted data and data that belongs to different proprietary vendors. They’re finding it harder to do that. So that’s kind of what I mean by, when I talk about customer intelligence, I really mean taking all that channel data, historical data, online data, putting it and doing something productive with it, which is still really hard for a lot of companies and practically out of reach for a lot of companies, both in terms of finding the right people to do it and finding the right tools to do it. A lot of what Domo says, I think, is where a lot of this is headed. I don’t think they’re very yet, but I think a lot of what they’re talking about, where they’re talking about going is being the vendor that puts all any data you have from any channel in one place so that anybody in your organization can use it and derive some sort of insight out of it.

00:07:37.85 [Tim Wilson]: Except Domo’s story is kind of a little more almost generic, it seems. They’re saying we will provide the connector and any data anybody wants, they can use it. That to me is very different from we will help you kind of get the right data from the different sources and actually think about the customer. I mean, where does customer intelligence meet up with marketers? So coming from the marketing side, Oh, we’ve got to be customer centric. We got to be customer centric to the point that it’s, it’s almost a meaningless term. And then you’ll hear people rail about what truly being customer centric means. And to me, there, there’s a piece of that that says, Hey, when you’re saying customer intelligence or when you’re, when you’re saying customer, There are sort of two things that I think can get lost in that. One, when you say customer, do you really mean somebody who has already purchased from you or do you mean a prospect or a customer? And I’m not asking you right now, though, if you have thoughts, let me know. And then the second thing is when we say customer, the immediate thing comes to mind as a person, whereas where does customer intelligence fall when you’re talking about your B2B and your customer is an organization or your B2C. But the household matters as much as the individual person. And so even when I hear something like customer intelligence, I think, crap, they’re already definitionally right out of the gate. Two things that I bet you say customer intelligence to 10 different people. Well, you won’t be able to get 10 different answers, but there will be kind of assumptions even like what that is. Yeah, definitely.

00:09:04.97 [Blair Reeves]: Yeah, that’s a good point. So I guess put the two and a half, right? So intelligence, a lot of the people doing this right now, there are people who kind of approach from different sides. Intelligence ideally should mean taking your data and driving some sort of insider value out of it, right? So that by definition, that would mean more than reporting, more than simply reporting what data is there and then visualizing it in some sort of way. Data visualization and reporting is really important, right? And it’s really hard, and it’s really, you know, it’s a really big area for, really big and important area for any kind of company. That’s why, you know, Data Studio from Google, they must have spent a lot of time working on that, and it’s a really, really cool tool. Tableau’s out there, and that’s an area that you’re gonna see a lot of growth in. It’s not really what we mean by intelligence, you know, and I think that when we talk about customer intelligence in the market, I think that people generally mean taking your data and doing something useful like prediction, discovery of, you know, either patterns or segments or, you know, affinities, for example, the sort of things we classically use kind of analytics to do.

00:10:10.06 [Michael Helbling]: Yeah, because I mean, when I think about it, I basically my mind immediately goes to customer acquisition as one major category, customer value as the second one, and then customer retention as the third. And sort of you spend time in your analysis hitting one of those three usually all the time. So, you know, just in the ones that examples you gave, those are all fit in one of those three categories. How do I get new customers? How do I make those customers more valuable, whether that be through affinities or stuff like that, and then how do I keep those customers around so they don’t churn out of my customer file?

00:10:48.51 [Blair Reeves]: Yeah, all things that on some level require some interrogation of the data, as well as some sort of reporting to do that, right, to make it known to different people in the company.

00:10:58.66 [Michael Helbling]: Right, so could we even stack it up? Like there’s sort of data visualization, data exploration, and data querying. I don’t know if that’s the right word for it, like the deepest level. and then various tools let you do more or less of that.

00:11:13.24 [Blair Reeves]: Yeah, it seems to me that there are a lot of, so there are any number of companies right now that have a channel strategy and say, if your data comes through this channel, we can take it, we can do all kinds of, we can make it stand up and do tricks. When you want to combine that data with another channel or with another set of data or what have you, that’s when it gets really tricky. Yes, that is correct. And there are, like I said, there’s any number of people trying to do this in different ways. And right now, it’s frankly a really kind of manual process for a lot of companies. And like I said, it’s giving way to a whole industry of people who do this on a contractual basis. So when I say productization, I think that this is an area where I think every vendor in our industry is working as hard as they can to productize a lot of that work because that work is where I think the future lives for any company that wishes to be competitive.

00:12:09.77 [Tim Wilson]: So what’s your sense on getting to that? And you started to talk to it earlier of sort of get all the data and centralize it into a warehouse or into a, what’d you call it? And not an aggregator. Maybe it was an aggregator. Get us a platform. I don’t even know if they’re a thing anymore. Let’s say, let’s say netiza.

00:12:31.62 [Michael Helbling]: You got to put it all in netiza data.

00:12:34.26 [Tim Wilson]: Yeah, I put it all. Okay, they’re throwing it all into BigQuery. I mean, there’s sort of one way to go is to say, we have a powerful, maybe expensive, maybe not expensive, but can handle large sets of a range of data types, structured and unstructured. And we’re just going to go and do an assessment of all of the data sources, all the systems that have something that might be useful. And we’re going to pump it all into some repository that we feel like we can then query. Yeah. I guess on the other extreme is that I’ve got one system. What would be the most valuable second set of data that would complement that? I’ve got my web analytics and stitching that together with my CRM is going to give me a step function, increased level of insight or intelligence into my customer. And I’ll just figure out the best way to do that with some eye to the fact that I’m then going to come wrong and throw my fulfillment system or ERP, into it and kind of taking an incremental approach. Or is that even a fair distinction to say that’s kind of the range of the spectrum of its bite this whole thing off and say, we have got to solve customer intelligence. And there may or may not be a product out there. There’ll be lots of companies saying they have a product and that product may be technology. It may be people, you know, it may be a service to solve customer intelligence versus a more incremental approach to say, We need to slowly be moving to where we’re bringing more and more stuff in. We don’t want to paint ourselves into a corner. So we don’t want to pick something that’s Uber proprietary and restricted. And that seems like more of kind of an incremental approach, probably more driven internally as opposed to using outside.

00:14:11.71 [Blair Reeves]: So, I mean, I think that traditionally there’s been a couple of approaches, right? So, one approach that’s been around for a while is, well, hell, you take all your data, throw it in a Pied Piper box, right? And then get someone to, you know, give them six months. You get a PhD statistician and a couple of developers and give them a whole lot of time and let them run some models over it and drive some sort of value out of, you know, all the all the stuff in your warehouse. Traditionally, that’s kind of, I don’t know, 10, 20, 30 years ago. That’s why banks started buying data, they wouldn’t call it data science then, but you know what I mean? Data solutions, right? That’s why SaaS is a company. That’s why IBM had a big business in that Microsoft had big business in that lots of us still do. I think that’s a architecture that might work for some people by this thing is it’s not necessary anymore. And it’s probably not the best approach anymore. And then if you look at like the attribution people what they’re doing today. There are also kind of two different approaches. One is we’ll sell you a whole lot of hours of these PhDs who are real whiz bangs with putting together disparate sources of data, put it all together and model it to develop some sort of value from it. And there are companies that do that today. The other approach is we’re going to apply analytics to it in an open sort of way so that you give it different sums of data and through the wizardry of technology, we will deliver some sort of value from it. And of course, it’s not really that simple. It’s usually a lot messier than that, but that is a message. Between those two, I think you see the direction that the general market is headed as well. As we go forward, you’re going to have a lot of sources of data. You have more than we did on just a couple of years ago. You’re going to have more in the future. I think we’re running out of runway for the approach in which any vendor can build a special solution to integrate every data source into it. different kinds of companies are going to have to find ways of feeding data in some sort of way that makes sense to them into a system that applies analytics to it that doesn’t necessarily require specially designing models for any particular data source. That’s kind of what I mean by productization. If it was just a manual process, we were doing that already. By productization, I sort of see this kind of going the same way that web analytics is going, which common tasks in web analytics around reporting, around segment discovery, customer value, things like this have been largely productized and that sets us down a path of ever more sophistication. So it’s very skeptical.

00:16:49.98 [Tim Wilson]: Well, I mean, I wonder, so a challenge with web analytics is that, you know, web analytics is data being 100% data being collected using internet protocols that were not designed or ever really intended for data collection. So major challenge of digital analytics is that it’s. Taking technology, taking a standard, taking the internet standards and figuring out how to get data out of that using mechanisms. Hey, we can use JavaScript to pass stuff.

00:17:20.70 [Michael Helbling]: Well, and our understanding of the customer is basically the browser cookie.

00:17:24.71 [Tim Wilson]: Right. That’s actually the essence of the big challenge there is it wasn’t designed for that. I’m wondering, and I’m thinking kind of XML and my limited knowledge of XML, you can start defining schemas that are a customer definition or I’m not an XML guy, I was going back like 10 years from where I was learning about it, you could You could have a schema, you could have a language or a structure that is generally accepted and you could have systems start to speak in whatever that language is. I don’t know if it go that far. I think it’s one of the challenges now with any of these solutions is you’re taking these point solution systems that are collecting data and everything about them is structured to what works well for their specific process.

00:18:16.29 [Blair Reeves]: Well, yeah, so I think I kind of track with what you’re saying here a little bit. So given that, you know, I guess it came from, you know, Coremetrics and the web analytics, you know, kind of world as well. You know, one issue that you have in web analytics is a little bit similar to what you have in like CRM or actually e-commerce. And that is that every implementation is different. Every implementation between different vendors and also with this, you know, with the same vendor, one implementation is going to be quite different from another. We’ll have certain similarities as well, but each one is different. So how do you take the data from one implementation and apply it to another? And what that gets back to is apply it to a data mile that runs in some sort of analytic process over that data that delivers, like I said, those kind of use cases that we talked about. One way you can think about this is by opening your schema by saying, doing some sort of mapping. So just in the same way you do it in e-commerce or web analytics implementation, you say, well, this event means this. Or this variable or property might be calculated, might be custom, or might be standard. And it means this thing that we can open up within a data schema. And that part is what I like to call the connected tissue. The connective tissue of doing this between, just for one data source, but then for many other ones, is a big challenge. And I think that some people have begun to get really sophisticated with this. I think Salesforce and Adobe do that real well. And the rest of us, we’re still working on it. The other part of it that is really hard is the actual analytics behind it. Because analytics has kind of become this buzzword where people call reporting analytics. And it is. And that’s like everyone knows that. And it isn’t. But if you can actually apply actual analytical models to that, actual analytical technology, then it can begin to derive a lot of value. So both parts of that. So some companies in our industry are really good at the connective tissue stuff. And I think that for them, then applying true analytics to the data that they’re able to plug into that sort of schema is kind of the next place that they’re going to be headed. For other ones, who are still kind of working on the connective tissue part of this and collecting that data together and then applying kind of analytical assets they already have, that’s going to be the next challenge. And one issue that, the other issue that I kind of see here is that All that data isn’t going to live in the same place, right? So if you’re a very large company, you probably don’t want to recreate or replace your entire data warehouse in the cloud, right? Like if you have sensitive customer information, you probably don’t want to like upload that to some sort of proprietary cloud. Domo is kind of getting around this in some interesting ways and you know every vendor in our industry you know kind of has ways of you upload you know certain files certain segments or whatever and that works kind of right it doesn’t work terribly well but at the same time we know we’re kind of we live in a cloud world we don’t really necessarily want to rely on an on-premise technology anymore. So really what you’re starting to talk about is a hybrid solution where some companies are going to be perfectly fine using all hosted customer information. That’s great. That makes things a lot easier, right? For a lot of other companies, especially larger companies, especially older companies, you’re a bank, you’re an airline, I don’t know, telco, whatever. Public sector, obviously, health care, all these kind of companies, you probably don’t want to or can’t do those things. So to make these kinds of solutions scale, you need a way to make those two work together in a way that doesn’t force you to buy a whole kind of data management solution.

00:22:01.39 [Tim Wilson]: Well, as I was thinking as you were talking, that’s the other thing we haven’t talked about is that customer data is notoriously messy and short lived. So even as you’re pulling from these different systems, you can say, well, I’ve got Blair Reeves is, you know, I’ve got from when I interacted with him a year ago, I know that Blair Reeves is in Raleigh, North Carolina, and that has pumped into this system. It’s outward of whatever the source system. I find Blair Reeves somewhere else, I stitch them together, and I’ve got something that’s factually inaccurate because you’re not in Raleigh right now. There’s the quality of underlying data and customer data management master data management with customer data management specifically has been around forever with much simpler systems when it’s all one Oracle platform and you’re still screwed with the decay of the quality of the customer so. I just realized we haven’t really hit on that. We’re saying getting the data from these different places. I’m claiming that a lot of these, you’re kind of getting what data you can, but it wasn’t really generated and it’s not being provided to help this larger, greater good of intelligence. So it’s got that stitching that you have to do, that connective tissue, you know, is having to do that transformation, hopefully without losing information along the way, just, you know, by the nature of it. And then you’ve got the time factor that really when you start talking about customer intelligence and the lifetime value, if you’re talking about lifetime value, you’re trying to figure out things over periods of time and over periods of time. A mobile society or many people are remote workers. I mean, we’ve got much squishier and more fluid data. Yeah. The signals that we’re generating are more This is depressing. I’m just going to go report on some traffic to my website.

00:23:49.06 [Michael Helbling]: It’s so funny because the principles for marketing and things like that are exactly the same as they used to be, but we are constantly having to figure out new ways to handle all the different ways that we are connected.

00:24:02.54 [Tim Wilson]: I am very skeptical of the approach, which I think larger, older companies are more prone to take, which is we’re going to go solve this. We are at point X now, and it’s laughable that you’d think it would be six months like large companies. They’ll kick off a two and a half year project to do to get to customer insights. And maybe they bring in a couple of PhDs and the PhDs start crunching on the data in the warehouse and quickly turn up that the data is complete and utter garbage. And they’ve got to go fix their underlying systems. And I mean, I’m not, I’m not arguing against this being where we want to head. I just, I feel like there’s something missing that makes it a practice to get to a practical reality. It is an enormous lift to get to that.

00:24:49.84 [Blair Reeves]: And you go to like a marketing technology conference today and you say master data management and people will just, you know, they’ll clear the room. You know, it’s like a go over like a fart in church. That’s kind of where we are because a lot of the marketing technology field is cluttered by, gosh, how do I do this? I’m going to be diplomatic about this. Great companies, right? But they kind of do one thing and they do one thing real, real well, assuming you use their solution and keep all your data in your account.

00:25:22.17 [Tim Wilson]: They have the classic, you know, that’s their hammer and all the world of them is a nail. So no matter what you throw at them, they’re going to frame it in terms of the scope of what their solution does.

00:25:30.58 [Blair Reeves]: And look, there are some big companies have been doing massive data management for and just data management generally, real well for a long time. And by a lot of those solutions just aren’t They’re not really appropriate for the kind of world of customer data we live in today. And like I said, some of that data is going to have to live at rest. Some of that data is going to live in the cloud. Some is going to live maybe in cold storage somewhere, you know? Either on a tape or even a glacier somewhere, whatever. So when you start talking about customer intelligence and scale, I think that the one issue you run into real fast is how you work with on-premise data and cloud data Or whatever data that a client feels comfortable keeping in the cloud for how long and how you do it. And the other thing you mentioned there is identity. Identity is a real tough nut to crack because you need some way of identifying who customers are. So for a long time we talked about kind of cookie ID mapping and that’s just kind of something everyone can do. But it gets to a lot more than that. Then you start talking about probabilistic matching or fuzzy matching with phone numbers and emails and names and addresses and all this stuff, which again is really hard to do unless you have a data management solution that can match whatever customer data you have in your warehouse with whatever is in your web analytics data or your e-commerce data somewhere. And those are things that I don’t think that the marketing clouds and the e-commerce platforms can really do. I don’t think that they have the reach to really do that very well. And they don’t necessarily have the incentive to really reach into those different sources to do those things. Because that’s fundamentally not really what they’re, it’s not a first order of business for those kinds of platforms. At least not today. Maybe in the future it will be, but for what they’re doing right now, I don’t see it.

00:27:24.25 [Michael Helbling]: Yeah, I think I agree with you because a lot of the companies that we’re talking about, Adobe and some of the other ones, they don’t, in my mind, don’t really touch customer analytics as much as they’re doing. some of the things around it and then you take that data and stick it into where you’re looking at customers.

00:27:42.56 [Blair Reeves]: Yeah well you know if you look at it so I know that it’s not really cool to talk about analysts but you know they do a lot of like you’ve seen a lot of this kind of surge in interest in like customer analytics, digital marketing analytics and those sorts of things as distinct from kind of the marketing hubs or the marketing clouds or whatever That’s sort of research. So there’s research coming out now more about kind of customer analytics and what that means Google has been talking about this for quite a while Yeah, they’ve been trying to move to the concept of a user or a concept of a person. Well, yeah I think they do it really well And I think they do there’s a lot of really they have people there who are like no tasks specifically doing that, which I think is really interesting. Google, of course, comes at it from an advertising perspective, which, you know, makes a lot of sense for where they are, and that means they’re kind of a different thing for a customer identity and their kind of use case for it. And there aren’t a whole lot of, you know, there aren’t a whole lot of people out there who have that kind of reach. But there’s a lot of research coming out. Gardner did a thing, Forster did a thing, of course. Some other ones have done other research around what customer analytics means, what it’s going to mean technologically for vendors to meet, and then for what you have to do. And I think that kind of getting back to the purpose of the podcast here, I mean, I think that we’re in the very early days of how that substance product dies, because right now it is a lot of elbow grease of analysts knee deep in a data warehouse.

00:29:08.61 [Tim Wilson]: I guess I’m still fuzzy on what the productisable aspects of the whole challenge are. You know, if we’ve got that, we’re not gonna necessarily make the data cleaner. So you’re gonna need that connective tissue piece. I mean, you could argue that, well, that’s kind of, you know, DOMO and other, you know, DOMO trying to build exillion connectors in a sense that’s maybe connective tissue. They’re trying to productize that piece of it. And I think is what you’re saying that if you get it to where you’ve got some common language where you can have repeatable across different industries, across different customers, ways of going at generating that analysis to make a super simplistic analogy, like a market basket analysis from 30 years ago. The concept of a market basket analysis was something that, okay, this is what it is. This is what it gives you. And here’s how you do it. And once that was kind of outlined, people who were trained in that could go with anything where there was more than one product put in the basket and they could do a market basket analysis across industries. Is that part of what the productization of customer intelligence is? Is what are the different ways that this gets approached?

00:30:20.01 [Blair Reeves]: So I see privatization as happening on two ends. One of those is the connected tissue like you just talked about. So instead of doing manual joins of n numbers of tables in your data warehouse, or instead of dumping all those into some sort of solution where you have a bunch of people attack it manually, you prioritize joining different customers to get data together. Domo has taken one direction to doing that, which is build a billion connectors. From a product perspective, I see a lot of difficulty in doing that. God bless them. They have a lot of smart people over there, and they got a lot of money. They had Alec Baldwin on the job. Yeah, they could do that. The other way you go about this, though, I think it’s more to think about it as a web analytics or even a CRM implementation, right? So for implementations of those kinds of systems, you kind of are, to some extent, kind of given a menu of, you know, here are the kinds of events we can pick up and here’s what you can do with them. But then there are also custom events, calculated fields, calculated metrics and this sort of thing. that you can kind of build yourself. And if you can take some of those and you have an analytical solution that whose schema is open, right? So there’s not a black box solution, but it’s kind of an open schema solution. So you say, well, you know, here’s the kind of field I want to fit in, you know, in this slide. And here’s kind of, I don’t want to fit in this slide. Here’s this, these sets of fields and sets of information that also want to fit into it, you know, then that it’s probably more scalable. In that kind of way, it’s hard to build, but it’s probably more scalable than building a connection for every possible data source you could ever possibly have.

00:32:09.73 [Tim Wilson]: Waving a magic wand, would that be a, and I don’t think consortiums and collective efforts tend to work for that when it comes to, but would there be kind of a body of an evolving schema that kind of gets to this consistent, I think this is where I was trying to go earlier, that is there this master definition think the the w3c but instead it’s just around the customer to say these are the definitions this is the relationships this is how we’re going to talk about the relationships between the data elements here’s the standard and now hey Pick your technology, if you want to play nice in this ecosystem, you need to be able to provide data in that mechanism. And that could be latency, quality, certainly the structure, the model. And it’s tough to generically define something like that across industries, across business models. But is that, would there be like, standardized just just definition like a bunch of bunch of web pages saying this is this is that schema. And now you’ve kind of disintermediated between the people where the data is being collected and the ones trying to crunch and get analysis or store the data. Well, maybe it’ll never happen.

00:33:28.08 [Blair Reeves]: I don’t see vendors wanting to cooperate to do that.

00:33:31.42 [Michael Helbling]: Tim, you just invented the new world order.

00:33:35.81 [Tim Wilson]: That’s what I’m saying. Wave a magic one. Like you can’t get here. You can’t get there from here.

00:33:40.82 [Blair Reeves]: Yeah, but so you think about if you’re using SPSS or SAS or there are other BI tools that exist, but why would you talk about them? So using a BI solution today, what do you do? You feed it data, and you have to map that data to a certain schema for that tool, and you do that differently for each tool. I don’t see why it wouldn’t work the same way here.

00:34:07.34 [Tim Wilson]: And I guess maybe that is to productize it means that I want to have the product that does all this. So it does wind up with this super high stakes with a lot of money behind it for eat a lot of different companies that are trying to come up with like the best solution. It kind of. It’s good for the analyst or the organization that wants to use the solution because competition is good and that means innovation and they’re going to be chasing each other. What sucks is it’s still going to be a pretty high, a high dollar investment to say, this is the horse that we’re going to ride because we think their schema and their approach and the product they’ve put out is the way to go. And it’s just, it’s tough.

00:34:48.71 [Blair Reeves]: It might be. I sort of think about it in the same way that we did. We went through this process many times before. This is literally like the web analytics market 10 years ago or five years ago, whatever it was. This is literally like what the attribution market has been like for a number of years. It’s a specialized solution. But if you want to derive more value on the data you already have, this is one way of doing that. And the added benefit of doing that, I think, is that it also, I don’t want to say it disintermediates the marketing stack, but it kind of sits right on top of that stack. So in addition to your e-commerce layer and your CRM later, your web analytics later, whatever, if you have a customer intelligence layer, that is able to consume those different data sources and do so repeatably and scalably, then you don’t have to rely on any of those lower layers in the stack to provide kind of a channel specific view. If you can get an omni channel, I almost said omni channel. I’m at a whiskey over here, so I have to get another one. If you can get a channel like Nasibia is kind of what I’m saying, from the connected standpoint, and then also apply some sort of real analytic value to that to provide actual business value to the whole purpose of doing that, then that gets as closer to what I would call customer intelligence or kind of what we’re moving towards.

00:36:25.19 [Michael Helbling]: Okay, so you definitely are advocating for this layer of this customer intelligence application of some kind.

00:36:31.86 [Blair Reeves]: Well, I think it’s a thing. I think this is a number of vendors kind of moving there, right? So you think about, I mean, IBM’s talking about Watson in these kinds of terms, right? IBM says, well, give us anything and wants to know where you kind of wave a magic magical on do it domo is It’s sort of saying that I mean mostly on the reporting side But I think that they have you know much larger ambitions beyond that as well Google I gave they’ve got to be you know, not a very short you know far ways off I’m sure they’ve got an AI solution in the works there, and it’s going to be in Google Analytics Premium, and you’re going to give it all kinds of data, and it’s going to come back with value in it, right? I mean, that’s kind of inevitable, I see. Adobe, I mean, there are a lot of smart people over there. I’m sure they’re working on something like this, and they already have, of course, a lot of analytical capabilities that they already kind of bring to bear in that sector as well.

00:37:26.24 [Michael Helbling]: So do you mostly see then sort of companies drilling down from the top layer, the digital analytics or sort of transactional layer down into customer intelligence as opposed to drilling up from the customer data into it? in terms of where you see movement then?

00:37:44.71 [Blair Reeves]: I think it kind of depends on where they are, right? Like, if you are a, you know, a dominant digital analytics layer, then yeah, then I’d move into those, you know, downward into the stack. If you are a CRM player, you know, then that’s probably, you know, so Salesforce has done this, right? So they built their own web analytics tool. They bought a marketing kind of cloud and they’ve done a lot of work in tying the two together. Salesforce has an analytics like a digital analytics tool. Yeah. Yeah, I mean, it’s not very well known It’s it’s it’s it’s it’s there. Yeah. Yeah, I have no idea what it’s called Do I I just I’ve seen you know, I’ve seen PowerPoints But then they of course have that they’re what they’re wave, you know analytics cloud now They have a now they have a commerce cloud so they have a different you know kind of modular parts of that and to what extent they’ve Know they’ve worked to kind of know tie the connective tissue there. They have the assets I should say To work something out of the connective tissue, I know it has to be seen yet. So again, it depends where you sit and where you sit in the sack today, if you’re in the sack at all. If you’re in there, then you either drill down to disintermediate the ones below you. If you’re below, then you kind of work up market, I don’t know.

00:38:57.47 [Michael Helbling]: Yeah, I think that’s interesting because I think as I have observed these kinds of things and I certainly I sit and look at this all from the perspective of a digital analyst. I don’t see a lot of connectivity most of the time between customer analytics and digital analytics unless you do a lot of hard work like we’ve been talking about. And most of those companies don’t really collide at that intersection or do a good job of doing one or the other, like the oracles and the sales forces do a great job. And down here in the CRM and the customer, but do a real crappy job up here in the digital place, whereas Adobe and Google are doing a great job up here and giving us great tools, but don’t live down here where all this other data is. And so it’s interesting to kind of see, you know, to your point where this is all going to go.

00:39:47.97 [Tim Wilson]: For those who don’t have the video version of the podcast, there were hand gestures. You can just imagine when he said up here that his hands were over his head and when he said down here his hands were much lower.

00:40:00.98 [Michael Helbling]: There’s about a 30% chance of digital analytics data coming in from the East.

00:40:08.96 [Tim Wilson]: He’s forecasting. Here’s what he’s doing.

00:40:10.46 [Blair Reeves]: I kind of got like an unfrozen caveman sort of looked at that. You know, if you’ve ever seen that skin missing.

00:40:17.17 [Michael Helbling]: That’s right. I am just, what is it, caveman lawyer? There we go, the pop culture references. Dude, that was Phil Hartman and it was like probably in the early 90s or something.

00:40:29.18 [Blair Reeves]: Yeah, that’s been a long time since it’s been pop culture, man.

00:40:31.60 [Michael Helbling]: Yeah, it still counts though for Tim.

00:40:34.73 [Tim Wilson]: Historical pop culture. I only got to watch eight is enough once a week was my lone TV.

00:40:41.32 [Michael Helbling]: All right. Are we reaching a point at which we’re going to throw our hands up in despair and realize that we just don’t care? I mean, we do care. I think we care deeply. I think probably the three of us should start that company that’s going to do this. I think we’ll be rich.

00:40:55.99 [Tim Wilson]: Rich, I tell you. I think a hundred people have started this company now and they’re going to get squashed like bugs by companies that are much larger that will do a piss poor job of actually pulling it off of integrating it. present company accepted and really not really. I mean, I’m thinking more of what the oracles and the adobes and the sales forces have done from the boardroom drawing of why this acquisition is going to give us this next thing. Like it just, it doesn’t seem like they’re ripe for the innovation that requires to get there.

00:41:29.84 [Michael Helbling]: I have nothing but positive things to say about my adobe acquisition experience.

00:41:37.22 [Tim Wilson]: If anything, I would say Salesforce is the one where it’s just laughable. I mean, Radeon 6, I don’t think they have spent more than 27 minutes of development since they bought Radeon 6 four years ago. It is exactly the same product. Which in a sense, you’d say, if I’m my CRM, there’s all the promise of all this social data, and I’ve got things that are connected into these platforms, and I should be able to bring in some of that. And I’ve never heard of anyone actually doing any sort of integration between Radiance 6 and Salesforce. But that was the, that was the press release was, Hey, we’re bringing in social. Okay, sorry. It took some box, right?

00:42:15.40 [Blair Reeves]: I mean,

00:42:16.04 [Tim Wilson]: Yeah. It’s hard. I mean, no matter what. But that’s my fear for the productization of customer intelligence is that they’re going to be checking a lot of boxes that say, in theory, if we bring X and Y together, and then if we just define this schema, and then we just go out and tell the world that we’ve got it figured out, then everyone will pivot and they will, you know, race to us and they’re going to ignore and skip all the realities of all the messiness around it, which I guess is where I’m finding myself on this discussion going more and more saying, God, I don’t want to be, I don’t want to be within a hundred miles of somebody who’s trying to, to bite it off that way. I would much rather say, Here’s the data you got that’s in good shape. Here’s that other data you got that’s in good shape. Let’s stitch them together. And the sum of the parts is greater than the whole. And now let’s add a third thing. And for the next decade, solve it incrementally and not buy into, you know, we’re going to pull this in from myriad sources and get glorious intelligence. But maybe that’s just the second beer talking.

00:43:20.77 [Blair Reeves]: Yeah, I mean, part of the, you know, from the product management perspective in the large company, I mean, the trick is always that you want to stay as focused as possible, right? Because you want to, you know, you have a use case, you have a very clearly defined use case that you want to satisfy and you want to meet and totally, you know, crush. And every product manager in the world is going to say, yeah, let’s do that. And then, of course, you have, you know, a dozen other customers out there say, well, I got this other thing that I need you to do too. I just need you to tweak that model a little bit. And you have another dozen customers over here who say, well, I got something kind of like that. And I need to go do something like it as well. And everyone’s going to want to do something a little bit different. So you get enough of those kinds of use cases getting together. And suddenly you have a much more general approach where first you are trumping one use case for one very focused application of analytics or application of customer data, all of a sudden you need to consume much more, you know, different types of customer data and many more types of analytics. So that’s a balance that no one has figured out how to, and there’s no, there’s, there is no, you know, solution for it. It’s just that you do the best you can. So we’re working on that. IBM’s working on that. Oracle’s working on that. Adobe Google are definitely working on that and I don’t think there’s gonna be a lot of different approaches that different vendors will have And we don’t we no one really knows who’s right yet.

00:44:49.56 [Michael Helbling]: So we’ll see no, this is good And I think now all of us are working on that.

00:44:55.07 [Michael Helbling]: So we got you get the whole weekend here. So I mean, yeah It’s not that complicated really by Moenday. We’ll at least have a mock-up

00:45:04.95 [Tim Wilson]: Well, it’s interesting. I don’t think we designed it this way, but our last episode having really been kind of exploring and digging into BigQuery, Google BigQuery, and really just trying to get an understanding of what it was. And if anything, given the nature of this podcast and the background of the host, we were coming at it from a Web analytics perspective and as I recall towards the end of that episode there was a little bit of a pivot where Michael Healy the guest was sort of making the comment that that that is one of those places that you can bring in disparate data and you start to get kind of magnified. value, but that’s absolutely kind of where you’ve been talking player. That is like take a sharp data scientist or a few sharp data scientists and get some technical chops and wire this stuff up. That is not productized. That is using emerging technology that is very powerful that can store this stuff and query it very quickly, but it is not productized customer intelligence. It’s just one of the avenues to try to get to the intelligence part of that.

00:46:05.50 [Blair Reeves]: Yeah, we used to talk a lot about digital intelligence. Back in the day, I pounded the table, I don’t know how many times, that digital analytics is the control point for all your digital intelligence a company has. Because it is the one place where you have true raw customer business information that then leads to everything else. Leads to your email, leads to your ads, leads to social, leads to everything else. Being in control of that place is an incredibly valuable place to be. That’s a model that is still true. Moebile changes some of that. channel fragmentation changes some of that, but it’s a really, really big place to be. And that’s why, I mean, that’s why Google and Alexa and Adobe are so dominant, right? Because they control those big control points there. And, you know, where I think customer intelligence changes that just a little bit is, you know, where where that lives on the stack. So whether you apply customer intelligence downward into the marketing stack or whether this is actually a new layer on top of that stack. And both of those approaches kind of mean different things from a technology perspective.

00:47:17.76 [Michael Helbling]: Well, this is good stuff. All right. Well, as much as I think we still haven’t solved this problem, we do have to wrap up the show. Pretty sure that wasn’t that. We weren’t that aspirational. We were so close, guys. We were so close. Five more minutes and we would have had it. Sorry. As you’ve been listening, if you have figured this out, please do drop. Let us know. Please do let us know. We’ll cut you in. I swear. No, but it is a really interesting conversation and it’s one that pretty much every single company is grappling with on some level. So I know it’s applicable. So as you are thinking about it and you’re listening, we’d love to hear from you on Slack or on our Facebook page or whatever. Also, you know, check out Blair Reeves on Twitter. He also writes a number of articles over on Medium, and frankly they are a great reading, so I would encourage you to do that. Player, is there anything you want to plug or talk about that you’re excited about that is your one moment to sell something to everybody listening?

00:48:18.91 [Blair Reeves]: Oh, to sell stuff?

00:48:20.94 [Tim Wilson]: Yeah. Well, no, no. You can sell. It can be a free Slack role-playing game. It also has a web-based version if you’d like.

00:48:30.83 [Blair Reeves]: I was telling these guys before we started recording that I made a mock-up of this old BVS role-playing game that I made, which is called Legend of the Plaid Dragon, which is on the Slack app directory. So you get Legend of the Plaid Dragons on there, but also go look at Salesforce Intelligence, where it’s a lot of fun. So check that out.

00:48:51.18 [Michael Helbling]: So we’ll be adding that to the measure Slack, I’m sure.

00:48:54.39 [Tim Wilson]: And don’t worry, he’s not collecting data on you to make you a future customer because he’s going to have intelligence about you.

00:49:01.11 [Michael Helbling]: So much intelligence. All right. Well, as usual, we go around at the end of the show. We do what’s called last call, something we think is cool that we’ve seen recently. So let’s go around. Who wants to start? You want to start? I love to start. You know, this is totally in another topical area, but there’s something that we’ve been using at Search Discovery that I’m growing in my appreciation of in terms of employee feedback and satisfaction. So we use this new, this tool and it’s probably been around a while, but it’s new to me, new ish called office vibe. And it goes and surveys everybody in the company, like kind of on a weekly basis and asks a few random questions and then collates all the results and gives them to you. And it’s anonymous and it’s super, super helpful to get feedback like that and start to know what you can do to kind of make your employees lives better. So anyways, I’ve been loving that and as hard as it is for me to see where we’re not doing things amazing, it’s the most valuable kind of feedback because it allows us as a company to consider and get better. And I’m like, I literally spend like time every day looking at all the new feedback and being like, interesting, okay, yeah, okay, what do I do with this? So if you are running a big team or thinking about running a big team, I highly recommend Office Vibe. Nice.

00:50:25.50 [Tim Wilson]: Should I go next? Timothy. So I suspect this will not be the last time that a last call ties to 538.com, but based on some of the stuff I’ve been doing with R, sort of reading on statistics, good old p-values, and it’s kind of a multi-part plug. 538.com has written some kind of hilarious articles, getting statisticians to actually define what a p-value is. There’s some kind of totally statistician nerd humor, but they also have the the p hacking site where you go and you uh you get to basically plug in different different values and uh see if your result is significant and it’s kind of making the point that boy the p value can be abused but it’s kind of fun to play around with and I think actually is somewhat sort of elucidating and informative around looking at things and relationships of data. So it’s kind of a little bit more refined than various correlations, not quite as broad-reaching. But 538’s various takes on p-values is my last call. Why are you smirking at me, Mr. Helbling?

00:51:32.69 [Michael Helbling]: I think that’s an amazing last call, and I can’t wait to go read some of those articles and figure out or refresh my memory of what a p-value is.

00:51:42.38 [Tim Wilson]: Well, I’ll say one of my favorite parts of it is when it says the American Statistical Association got 26 statisticians members to sit down and try to say, look, we’ve just got to come up with one simple definition of a p-value. And when they came up with, starts with the word, it says informally, and then it goes on for like a modest length paragraph defining p-value. And even the author of the article is like, yep, that didn’t help. I kind of look at it saying, yeah, you can’t just say, you can’t start by saying informally and make it all of a sudden be like simple and clear. So, fun times. Blair.

00:52:15.81 [Blair Reeves]: All right, so my thing is, I have to apologize, but it is from Anderson Horowitz. It is a video, AI Deep Learning and Machine Learning, a primer by a student named Frank Chen. It’s on their website, but basically it’s like the 30 minute, 35 minute or so, a little primer video, a presentation that this dude, Frank Chen, gives about the history of AI, kind of where it comes from. It’s roots in computer science. you know, how different approaches to AI and machine learning have worked, what has changed for, you know, to kind of like lead to the breakthroughs we’re seeing today in the new applications of AI and what deep learning is and how it differs from machine learning in different kinds of applications. And I don’t know, I learned a lot from it and thought it was really a lot of fun. So, highly recommend.

00:53:07.51 [Michael Helbling]: Nice. Wow. These are some high value last calls this show. So, nice work if anybody I got some real stinkers lined up for the next episode don’t worry the five people that made it this far into the episode now I’m just so thanks for listening Blair thank you so much for coming on the show it’s been great to discuss this topic with you it’s comforting on a certain level maybe that it’s the others haven’t totally figured this out and the rest of us are being left behind and it’s alarming on another level that we’ve all got to solve this big hairy problem so we can really do a good job for our customers. But let’s keep it truckin’ and for all of you listening and my co-host Tim Wilson, keep analyzing.

00:53:56.31 [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. Facebook.com forward slash analytics hour or at analytics hour on Twitter. So smart guys want to fit in. So they made up a term called analytics. Analytics don’t work.

00:54:19.27 [Michael Helbling]: Yeah, I’m almost ready.

00:54:20.71 [Blair Reeves]: Just keep talking amongst yourselves. Okay, you’re not gonna make me a defined customer intelligence, are you?

00:54:26.40 [Michael Helbling]: Got to.

00:54:27.46 [Michael Helbling]: Man, come on. Okay, come on. Tim, let’s focus.

00:54:33.51 [Michael Helbling]: I do not want to drag you kicking and screaming into this podcast. All right, if you don’t want to do it, let’s talk about it. I’ll just tune out. I don’t remember what we did with what intros when. Oh my gosh. I’m just a talent. Like, seriously. this whole section.

00:54:53.29 [Michael Helbling]: Alright.

00:54:55.69 [Michael Helbling]: You might as well be mid-afternoon on the baron.

00:55:01.01 [Tim Wilson]: Look at that, you got a god bless you out of it too. That’s like a southern view of the southern.

00:55:06.53 [Blair Reeves]: So SAS is private. Yeah, it’s entirely owned by Dr. Goodman. We’re also privately held.

00:55:12.26 [Michael Helbling]: We’re a private company. It’s just kidding, town. Fuck you. Hello. I also analytics.

00:55:25.22 [Michael Helbling]: Matt, talk to you.

00:55:28.22 [Blair Reeves]: I analytics real good.

00:55:30.81 [Tim Wilson]: They’re like, we’re gonna have 150 analysts inside of a year. I was like, where are you going to find those people? Because I can tell you 10 companies that want just one.

00:55:39.62 [Michael Helbling]: I don’t think we need to stoke people’s egos, Tim. I mean, he’s on the show. He gets it. He’s part of this now. All right, we’ll fix it in post. Listen.

00:55:56.82 [Michael Helbling]: Don’t want to hear your sass.

00:55:58.60 [Michael Helbling]: Exactly. Does that joke work a lot? Where you, where you’re from?

00:56:03.13 [Tim Wilson]: Oh, I’m just being an ass. You are. Rock flag and cussing tumors.

00:56:15.05 [Michael Helbling]: Cussing tumors. I like it.

Podcast: Download | Embed

Subscribe: RSS