>> Live from Boston, Massachusetts, it's theCUBE, Covering IBM Chief Data Officer Summit. Brought to you by IBM. >> Welcome back to theCUBE's live coverage of IBM Chief Data Strategy Summit. My name is Rebecca Knight, and I'm here with my co-host Dave Vellante, we are joined by Christopher Penn, the VP of Marketing Technology at SHIFT Communications, here in Boston. >> Yes. >> Thanks so much for joining us. >> Thank you for having me. >> So we're going to talk about cognitive marketing. Tell our viewers: what is cognitive marketing, and what your approach to it is. >> Sure, so cognitive marketing essentially is applying machine learning and artificial intelligence strategies, tactics and technologies to the discipline of marketing. For a really long time marketing has been kind of known as the arts and crafts department, which was fine, and there's certainly, creativity is an essential part of the discipline, that's never going away. But we have been tasked with proving our value. What's the ROI of things, is a common question. Where's the data live? The chief data officer would be asking, like, who's responsible for this? And if we don't have good answers to those things, we kind of get shown the door. >> Well it sort of gets back to that old adage in advertising, I know half my marketing budget is wasted, I just don't know which half. >> Exactly. >> So now we're really able to know which half is working. >> Yeah, so I mean, one of the more interesting things that I've been working on recently is using what's called Markov chains, which is a type of very primitive machine learning, to do attribution analysis, to say what actually caused someone to become a new viewer of theCUBE, for example. And you would take all this data that you have from your analytics. Most of it that we have, we don't really do anything with. You might pull up your Google Analytics console, and go, "Okay, I got more visitors today than yesterday." but you don't really get a lot of insights from the stock software. But using a lot of tools, many of which are open source and free of financial cost, if you have technical skills you can get much deeper insights into your marketing. >> So I wonder, just if we can for our audience... When we talk about machine learning, and deep learning, and A.I., we're talking about math, right, largely? >> Well so let's actually go through this, because this is important. A.I. is a bucket category. It means teaching a machine to behave as though it had human intelligence. So if your viewers can see me, and disambiguate me from the background, they're using vision, right? If you're hearing sounds coming out of my mouth and interpreting them into words, that's natural language processing. Humans do this naturally. It is now trying to teach machines to do these things, and we've been trying to do this for centuries, in a lot of ways, right? You have the old Mechanical Turks and stuff like that. Machine learning is based on algorithms, and it is mostly math. And there's two broad categories, supervised and unsupervised. Supervised is you put a bunch of blocks on the table, kids blocks, and you hold the red one, and you show the machine over and over again this is red, this is red, and eventually you train it, that's red. Unsupervised is- >> Not a hot dog. (Laughter) >> This is an apple, not a banana. Sorry CNN. >> Silicon Valley fans. >> Unsupervised is there's a whole bunch of blocks on the table, "Machine, make as many different sequences as possible," some are big, some are small, some are red, some are blue, and so on, and so forth. You can sort, and then you figure out what's in there, and that's a lot of what we do. So if you were to take, for example, all of the comments on every episode of theCUBE, that's a lot, right? No humans going to be able to get through that, but you can take a machine and digest through, just say, what's in the bag? And then there's another category, beyond machine learning, called deep learning, and that's where you hear a lot of talk today. Deep learning, if you think of machine learning as a pancake, now deep learnings like a stack of pancakes, where the data gets passed from one layer to the next, until what you get at the bottom is a much better, more tuned out answer than any human can deliver, because it's like having a hundred humans all at once coming up with the answer. >> So when you hear about, like, rich neural networks, and deep neural networks, that's what we're talking about. >> Exactly, generative adversarial networks. All those things are ... Any kind of a lot of the neural network stuff is deep learning. It's tying all these piece together, so that in concert, they're greater than the sum of any one. >> And the math, I presume, is not new math, right? >> No. >> SVM and, it's stuff that's been around forever, it's just the application of that math. And why now? Cause there's so much data? Cause there's so much processing power? What are the factors that enable this? >> The main factor's cloud. There's a great shirt that says: "There's no cloud, it's just somebody else's computer." Well it's absolutely true, it's all somebody else's computer but because of the scale of this, all these tech companies have massive server farms that are kind of just waiting for something to do. And so they offer this as a service, so now you have computational power that is significantly greater than we've ever had in human history. You have the internet, which is a major contributor, the ability to connect machines and people. And you have all these devices. I mean, this little laptop right here, would have been a supercomputer twenty years ago, right? And the fact that you can go to a service like GitHub or Stack Exchange, and copy and paste some code that someone else has written that's open source, you can run machine learning stuff right on this machine, and get some incredible answers. So that's why now, because you've got this confluence of networks, and cloud, and technology, and processing power that we've never had before. >> Well with this emphasis on math and science in marketing, how does this change the composition of the marketing department at companies around the world? >> So, that's a really interesting question because it means very different skill sets for people. And a lot of people like to say, well there's the left brain and then there's a right brain. The right brains the creative, the left brains the quant, and you can't really do that anymore. You actually have to be both brained. You have to be just as creative as you've always been, but now you have to at least have an understanding of this technology and what to do with it. You may not necessarily have to write code, but you'd better know how to think like a coder, and say, how can I approach this problem systematically? This is kind of a popular culture joke: Is there an app for that, right? Well, think about that with every business problem you face. Is there an app for that? Is there an algorithm for that? Can I automate this? And once you go down that path of thinking, you're on the path towards being a true marketing technologist. >> Can you talk about earned, paid, and owned media? How those lines are blurring, or not, and the relationship between sort of those different forms of media, and results in PR or advertising. >> Yeah, there is no difference, media is media, because you can take a piece of content that this media, this interview that we're doing here on theCUBE is technically earned media. If I go and embed this on my website, is that owned media? Well it's still the same thing, and if I run some ads to it, is it technically now paid media? It's the thing, it's content that has value, and then what we do with it, how we distribute it, is up to us, and who our audience is. One of the things that a lot of veteran marketing and PR practitioners have to overcome is this idea that the PR folks sit over there, and they just smile and dial and get hits, go get another hit. And then the ad folks are over here... No, it's all the same thing. And if we don't, as an industry realize that those silos are artificially imposed, basically to keep people in certain jobs, we will eventually end up turning over all of it to the machines, because the machines will be able to cross those organizational barriers much faster. When you have the data, and whatever the data says that's what you do. So if the data says this channels going to be more effective, yes it's a CUBE interview, but actually it's better off as a paid YouTube video. So the machine will just go do that for us. >> I want to go back to something you were talking about at the very beginning of the conversation, which is really understanding, companies understanding, how their marketing campaigns and approaches are effectively working or not working. So without naming names of clients, can you talk about some specific examples of what you've seen, and how it's really changed the way companies are reaching customers? >> The number one thing that does not work, is for any business executive to have a pre-conceived idea of the way things should be, right? "Well we're the industry leader in this, we should have all the market share." Well no, the world doesn't work like that anymore. This lovely device that we all carry around in our pockets is literally a slot-machine for your attention. >> I like it, you've got to copyright that. A slot machine for your attention. >> And there's a million and a half different options, cause that's how many apps there are in the app store. There's a million and half different options that are more exciting than your white paper. (Laughter) Right, so for companies that are successful, they realize this, they realize they can't boil the ocean, that you are competing every single day with the Pope, the president, with Netflix, you know, all these things. So it's understanding: When is my audience interested in something? Then, what are they interested in? And then, how do I reach those people? There was a story on the news relatively recently, Facebook is saying, "Oh brand pages, we're not going to show "your stuff in the regular news feed anymore, "there will be a special feed over here "that no one will ever look at, unless you pay up." So understanding that if we don't understand our audiences, and recruit these influencers, these people who have the ability to reach these crowds, our ability to do so through the "free" social media continues to dwindle, and that's a major change. >> So the smart companies get this, where are we though, in terms of the journey? >> We're in still very early days. I was at major Fortune 50, not too long ago, who just installed Google Analytics on their website, and this is a company that if I named the name you would know it immediately. They make billions of dollars- >> It would embarrass them. >> They make billions of dollars, and it's like, "Yeah, we're just figuring out this whole internet thing." And I'm like, "Cool, we'd be happy to help you, but why, what took so long?" And it's a lot of organizational inertia. Like, "Well, this is the way we've always done it, and it's gotten us this far." But what they don't realize is the incredible amount of danger they're in, because their more agile competitors are going to eat them for lunch. >> Talking about organizational inertia, and this is a very big problem, we're here at a CDO summit to share best practices, and what to learn from each other, what's your advice for a viewer there who's part of an organization that isn't working fast enough on this topic? >> Update your LinkedIn profile. (Laughter) >> Move on, it's a lost cause. >> One of the things that you have to do an honest assessment of, is whether the organization you're in is capable of pivoting quickly enough to outrun its competition. And in some cases, you may be that laboratory inside, but if you don't have that executive buy in, you're going to be stymied, and your nearest competitor that does have that willingness to pivot, and bet big on a relatively proven change, like hey data is important, yeah, you make want to look for greener pastures. >> Great, well Chris thanks so much for joining us. >> Thank you for having me. >> I'm Rebecca Knight, for Dave Vellante, we will have more of theCUBE's coverage of the IBM Chief Data Strategy Officer Summit, after this.

>> Live from Boston, Massachusetts, it's The Cube! Covering IBM Chief Data Officer Summit, brought to you by IBM. >> Welcome back to The Cube's live coverage of IBM's Chief Data Strategy Summit, I'm your host Rebecca Knight, along with my cohost Dave Vellante. We have a big panel today, these are our social influencers. Starting at the top, we have Christopher Penn, VP Marketing of Shift Communications, then Tripp Braden, Executive Coach and Growth Strategist at Strategic Performance Partners, Mike Tamir, Chief Data Science Officer at TACT, Bob Hayes, President of Business Over Broadway. Thanks so much for joining us. >> Thank you. >> So we're talking about data as a way to engage customers, a way to engage employees. What business functions would you say stand to benefit the most from using data? >> I'll take a whack at that. I don't know if it's the biggest function, but I think the customer experience and customer success. How do you use data to help predict what customers will do, and how do you then use that information to kind of personalize that experience for them and drive up recommendations, retention, upselling, things like that. >> So it's really the customer experience that you're focusing on? >> Yes, and I just released a study. I found that analytical-leading companies tend to use analytics to understand their customers more than say analytical laggards. So those kind of companies who can actually get value from data, they focus their efforts around improving customer loyalty by just gaining a deeper understanding about their customers. >> Chris, you want to jump in here with- >> I was just going to say, as many of us said, we have three things we really care about as business people, right? We want to save money, save time, or make money. So any function that meets those qualifications, is a functional benefit from data. >> I think there's also another interesting dimension to this, when you start to look at the leadership team in the company, now having the ability to anticipate the future. I mean now, we are no longer just looking at static data. We are now looking at anticipatory capability and seeing around corners, so that the person comes to the team, they're bringing something completely different than the team has had in the past. This whole competency of being able to anticipate the future and then take from that, where you take your organization in the future. >> So follow up on that, Tripp, does data now finally trump gut feel? Remember the HBR article of 10, 15 years ago, can't beat gut feel? Is that, we hit a new era now? >> Well, I think we're moving into an era where we have both. I think it's no longer an either or, we have intuition or we have data. Now we have both. The organizations who can leverage both at the same time and develop that capability and earn the trust of the other members by doing that. I see the Chief Data Officer really being a catalyst for organizational change. >> So Dr. Tamir I wonder if I could ask you a question? Maybe the whole panel, but so we've all followed the big data trend and the meme, AI, deep learning, machine learning, same wine, new bottle, or is there something substantive behind it? >> So certainly our capabilities are growing, our capabilities in machine learning, and I think that's part of why now there's this new branding of AI. AI is not what your mother might have thought AI is. It's not robots and cylons and that sort of thing that are going to be able to think intelligently. They just did intelligence tests on the different, like Siri and Alexa, quote AIs from different companies, and they scored horribly. They scored much worse than my, much worse than my very intelligent seven-year old. And that's not a comment on the deficiencies in Alexa or in Siri. It's a comment on these are not actually artificial intelligences. These are just tools that apply machine learning strategically. >> So you are all thinking about data and how it is going to change the future and one of the things you said, Tripp, is that we can now see the future. Talk to me about some of the most exciting things that you're seeing that companies do that are anticipating what customers want. >> Okay, so for example, in the customer success space, a lot of Sass businesses have a monthly subscription, so they're very worried about customer churn. So companies are now leveraging all the user behavior to understand which customers are likely to leave next month, and if they know that, they can reach out to them with maybe some retention campaigns, or even use that data to find out who's most likely to buy more from you in the next month, and then market to those in effective ways. So don't just do a blast for everybody, focus on particular customers, their needs, and try to service them or market to them in a way that resonates with them that increases retention, upselling, and recommendations. >> So they've already seen certain behaviors that show a customer is maybe not going to re-up? >> Exactly, so you just, you throw this data in a machine learning, right. You find the predictors of your outcome that interest you, and then using that information, you say oh, maybe predictors A, B, and C, are the ones that actually drive loyalty behaviors, then you can use that information to segment your customers and market to them appropriately. It's pretty cool stuff. >> February 18th, 2018. >> Okay. >> So we did a study recently just for fun of when people search for the term "Outlook, out of office." Yeah, and you really only search for that term for one reason, you're going on vacation, and you want to figure out how to turn the feature on. So we did a five-year data poll of people, of the search times for that and then inverted it, so when do people search least for that term. That's when they're in the office, and it's the week of February 18th, 2018, will be that time when people like, yep, I'm at the office, I got to work. And knowing that, prediction and data give us specificity, like yeah, we know the first quarter is busy, we know between memorial Day and Labor Day is not as busy in the B to B world. But as a marketer, we need to put specificity, data and predictive analytics gives us specificity. We know what week to send our email campaigns, what week to turn our ad budgets all the way to full, and so on and so forth. If someone's looking for The Cube, when will they be doing that, you know, going forward? That's the power of this stuff, is that specificity. >> They know what we're going to search for before we search for it. (laughter) >> I'd like to know where I'm going to be next week. Why that date? >> That's the date that people least search for the term, "Outlook, out of office." >> Okay. >> So, they're not looking for that feature, which logically means they're in the office. >> Or they're on vacation. (laughter) Right, I'm just saying. >> That brings up a good point on not just, what you're predicting for interactions right now, but also anticipating the trends. So Bob brought up a good point about figuring out when people are churning. There's a flip side to that, which is how do you get your customers to be more engaged? And now we have really an explosion in reinforcement learning in particular, which is a tool for figuring out, not just how to interact with you right now as a one off, statically. But how do I interact with you over time, this week, next week, the week after that? And using reinforcement learning, you can actually do that. This is the the sort-of technique that they used to beat Alpha-Go or to beat humans with Alpha-Go. Machine-learning algorithms, supervised learning, works well when you get that immediate feedback, but if you're playing a game, you don't get that feedback that you're going to win 300 turns from now, right now. You have to create more advanced value functions and ways of anticipating where things are going, this move, so that you see things are on track for winning in 20, 30, 40 moves, down the road. And it's the same thing when you're dealing with customer engagement. You want to, you can make a decision, I'm going to give this customer a coupon that's going to make them spend 50 cents more today, or you can make decisions algorithmically that are going to give them a 50 cent discount this week, next week, and the week after that, that are going to make them become a coffee drinker for life, or customer for life. >> It's about finding those customers for life. >> IBM uses the term cognitive business. We go to these conferences, everybody talks about digital transformation. At the end of the day it's all about how you use data. So my question is, if you think about the bell curve of organizations that you work with, how do they, what's the shape of that curve, part one. And then part two is, where do you see IBM on that curve? >> Well I think a lot of my clients make a living predicting the future, they're insurance companies and financial services. That's where the CDO currently resides and they get a lot of benefit. But one of things we're all talking about, but talking around, is that human element. So now, how do we take the human element and incorporate this into the structure of how we make our decisions? And how do we take this information, and how do we learn to trust that? The one thing I hear from most of the executives I talk to, when they talk about how data is being used in their organizations is the lack of trust. Now, when you have that, and you start to look at the trends that we're dealing with, and we call them data points verses calling them people, now you have a problem, because people become very, almost analytically challenged, right? So how do we get people to start saying, okay, let's look at this from the point of view of, it's not an either or solution in the world we live in today. Cognitive organizations are not going to happen tomorrow morning, even the most progressive organizations are probably five years away from really deploying them completely. But the organizations who take a little bit of an edge, so five, ten percent edge out of there, they now have a really, a different advantage in their markets. And that's what we're talking about, hyper-critical thinking skills. I mean, when you start to say, how do I think like Warren Buffet, how do I start to look and make these kinds of decisions analytically? How do I recreate an artificial intelligence when machine-learning practice, and program that's going to provide that solution for people. And that's where I think organizations that are forward-leaning now are looking and saying, how do I get my people to use these capabilities and ultimately trust the data that they're told. >> So I forget who said it, but it was early on in the big data movement, somebody said that we're further away from a single version of the truth than ever, and it's just going to get worse. So as a data scientist, what say you? >> I'm not familiar with the truth quote, but I think it's very relevant, well very relevant to where we are today. There's almost an arms race of, you hear all the time about automating, putting out fake news, putting out misinformation, and how that can be done using all the technology that we have at our disposal for disbursing that information. The only way that that's going to get solved is also with algorithmic solutions with creating algorithms that are going to be able to detect, is this news, is this something that is trying to attack my emotions and convince me just based on fear, or is this an article that's trying to present actual facts to me and you can do that with machine-learning algorithms. Now we have the technology to do that, algorithmically. >> Better algos than like and share. >> From a technological perspective, to your question about where IBM is, IBM has a ton of stuff that I call AI as a service, essentially where if you're a developer on Bluemix, for example, you can plug in to the different components of Watson at literally pennies per usage, to say I want to do sentiment analysis, I want to do tone analysis, I want personality insights, about this piece, who wrote this piece of content. And to Dr. Tamir's point, this is stuff that, we need these tools to do things like, fingerprint this piece of text. Did the supposed author actually write this? You can tell that, so of all the four magi, we call it, the Microsoft, Amazon, Google, IBM, getting on board, and adding that five or ten percent edge that Tripp was talking about, is easiest with IBM Bluemix. >> Great. >> Well, one of the other parts of this is you start to talk about what we're doing and you start to look at the players that are doing this. They are all organizations that I would not call classical technology organizations. They were 10 years ago, look at a Microsoft. But you look at the leadership of Microsoft today, and they're much more about figuring out what the formula is for success for business, and that's the other place I think we're seeing a transformation occurring, and the early adopters, is they have gone through the first generation, and the pain, you know, of having to have these kinds of things, and now they're moving to that second generation, where they're looking for the gain. And they're looking for people who can bring them capability and have the conversation, and discuss them in ways that they can see the landscape. I mean part of this is if you get caught in the bits and bites, you miss the landscape that you should be seeing in the market, and that's why I think there's a tremendous opportunity for us to really look at multiple markets of the same data. I mean, imagine looking and here's what I see, everyone in this group would have a different opinion in what they're seeing, but now we have the ability to see it five different ways and share that with our executive team and what we're seeing, so we can make better decisions. >> I wonder if we could have a frank conversation, an honest conversation about the data and the data ownership. You heard IBM this morning, saying hey we're going to protect your data, but I'd love you guys, as independents to weigh in. You got this data, you guys are involved with your clients, building models, the data trains the model. I got to believe that that model gets used at a lot of different places, within an industry, like insurance or across retail, whatever it is. So I'm afraid that my data is, my IP is going to seep across the industry. Should I not be worried about that? I wonder if you guys could weigh in. >> Well if you work with a particular vendor, sometimes vendors have a stipulation that we will not share your models with other clients, so you just got to stick to that. But in terms of science, I mean you build a model, right? You want to generalize that to other businesses. >> Right! >> (drowned out by others talking) So maybe if you could work somehow with your existing clients, say here, this is what we want to do, we just want to elevate the waters for everybody, right? So everybody wins when all boats rise, right? So if you can kind of convince your clients that we just want to help the world be better, and function better, make employees happier, customers happier, let's take that approach and just use models in a, that may be generalized to other situations and use them. If if you don't, then you just don't. >> Right, that's your choice. >> It's a choice, it's a choice you have to make. >> As long as you're transparent about it. >> I'm not super worried, I mean, you, Dave, Tripp, and I are all dressed similarly, right? We have the model of shirt and tie so, if I put on your clothes, we wouldn't, but if I were to put on your clothes, it would not be, even though it's the same model, it's just not going to be the same outcome. It's going to look really bad, right, so. Yes, companies can share the models and the general flows and stuff, but there's so much, if a company's doing machine learning well, there's so much feature engineering that's unique to that company that trying to apply that somewhere else, is just going to blow up. >> Yeah, but we could switch ties, like Tripp has got a really cool tie, I'd be using that tie on July 4th. >> This is turning into a different kind of panel (laughter) Chris, Tripp, Mike, and Bob, thanks so much for joining us. This has been a really fun and interesting panel. >> Thank you very much. Thank you. >> Thanks you guys. >> We will have more from the IBM Summit in Boston just after this. (techno music)