Blog Career Nation Show Career strategy

Episode 16: Career Nation show with Eric Siegel

“AI is an extremely powerful technology that helps us automate mass scale decisions. It’s sort of the implementation of private and public policies and automation and mechanization of those societal functions. It brings up a lot of ethical issues around social justice and how this affects all the people about whom decisions are being made.”

Eric Seigel, Ph. D, and founder of Predictive Analytics World joins us in episode 16 of the Career Nation Show.

Here are some of the highlights from this episode:

Why the term AI is misleading?

What is predictive analytics? How does it work?

How it generates a predictive model?

The operationalization of the model.

The opportunities with machine learning and predictive analytics

You can get the copy of Eric Seigel’s best selling book, Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, from here:


Career Nation: Career Nation, welcome to another episode. And today we have a fascinating guest, Eric Siegel. He’s the founder of Predictive Analytics World and he’s also the author of the award-winning book ‘Predictive Analytics’, the power to predict, who will click, buy, lie or die. He’s a former Columbia University professor and he now he helps companies and individuals understand the power of predictive analytics, artificial intelligence, and machine learning. Please welcome Eric Siegel. Eric, welcome to the show.

Eric Siegel: Thanks Abhijeet. Thanks for including me.

Career Nation: Oh, it’s an absolute pleasure. Now this is such a fascinating topic and subject for so many of our audience. before we get into that, can you tell us, a little bit about yourself? Give us your background, give us your intro, if you will.

Eric Siegel: Sure. I’m a former computer science professor at Columbia university where I focused on machine learning and I’m a consultant, author, speaker at, in, in the field and, the founder of, Predictive Analytics World, which is the leading conference series focused on the commercial deployment. That is, it’s not a research conference, or a research and development or academic. It’s focused on the real world usage of machine learning. and it’s a vendor neutral event. So it involves and includes, information that pertains to whatever software tool or solution you may be using. so Predictive Analytics World, has been running since 2009 in our main large North American event. This year is May 31st to June 4th in Las Vegas.

Career Nation: Oh, that sounds fun man. So you think about all this wonderful technology in Vegas. It’s, it’s going to be a lot of fun there. thanks for the intro. There’s so many nuggets there that I would love to unpack. But let’s start here. Maybe, maybe, for, for a lot of our viewers, if you can help us understand a little bit better about what is predictive analytics versus machine learning versus artificial intelligence. A lot of times, people do use them interchangeably, in some cases. So, what in your opinion, in your perspective would help people to understand those things a little bit better? Yes.

Eric Siegel: Let me define machine learning and predictive analytics first. Artificial intelligence, it’s pretty subjective word, used many different ways. So is data science and big data by the way. So, machine learning is technology that learns from experience to make predictions or to classify individuals. So, in business applications that would be per customer, right? And all sorts of medical imaging and online applications. It’s images and that’s also referred to as deep learning and many, and in many cases that’s a type of neural network. So, the idea is that you have a whole bunch of labeled data. So these are, these are also referred to as supervised machine learning. Supervised because you have the data that’s already labeled to learn from, that’s called the training data and it already has a whole bunch of examples where either you knew what turned out to happen.

Eric Siegel: So let’s say you’re trying to predict which customer is going to cancel or churn. It’s called churn modeling. you already have a whole bunch of examples just because of the history of who did actually leave as a customer and you know the outcome. Or you have a bunch of labeled images that show pictures of cats and dogs and you know which are which because humans have already labeled them. Either way from history or human labeling, you have the labels and that’s supervised in the sense that it directs the learning process. It’s a way to measure how well the model labels or classifies or makes predictions per individual. So you can imagine this concept of individual applies very generally, can be per image per individual customer, corporate client, product, product line, satellite that might run out of a battery. There’s, it’s such a universally applicable technology.

Eric Siegel: The idea of machine learning to learn from those examples to generate a predictive model. So, it’s also called predictive modeling and then the model itself, everything that you’ve learned from those labeled data is now applied over data where it’s not labeled. You don’t know how it’s going to turn out, you don’t know which customer is going to cancel. You have images you need to classify, you don’t already have them labeled. Right? So, that’s the whole point. You’ve learned from the data. And it’s so generally applicable across business, applications, applications in government, and political campaigning, basically across all industries.

Career Nation: Oh, that’s wonderful. And so if, if you, if I were to sort of understand some of that, then there is some human form of this. Which is we figure out and apply some labels to these different items and whether we are trying to, you know, research about our customers or research about, you know, products or what have you. And then we, based on this, we can make certain predictions using predictive analytics. And so, this is the sort of the promise of taking sort of machine learning, which is understanding more about these labels and applying them into a predictive way to do predictive analytics. And is artificial intelligence and extension of that? Is that, is that something that’s sort of forward looking from that or it sort of starts from a sort of these predictive analytic type of models?

Eric Siegel: Artificial intelligence is a buzzword and subjectively defined word that can mean pretty much whatever you want because, and typically it’s defined in a circular way. It means making computers intelligent. But the word intelligence was in the word you’re trying to define. So, it doesn’t mean anything. It means whatever you want. It’s a, it typically either is an exact synonym to machine learning or deep learning, which is a specific kind of machine learning or it’s sort of a broad subjective category that includes machine learning as well as like chatbox, chatbots or anything else that a human feels is humanlike from a computer or it’s sort of a vague word that kind of over promises the direction of how quickly machine learning is going to grow and, and computers are going to become more human like in some way. So it literally serves only to hype.

Eric Siegel: There is absolutely no intrinsic meaning to the term Artificial Intelligence. And by the way, I taught the artificial intelligence graduate level course at Columbia university a couple of decades ago. And my opinion hasn’t really changed since then. The hype about it right now is, is good in the sense that yes, machine learning is extremely powerful, but there’s also some over promising going on out there. By the way, I didn’t really differentiate between machine learning and predictive analytics. So, predictive analytics is basically a, a major subset of machine learning. So, if you’re using machine learning for business problems like customer churn prediction, targeting, marketing by predicting which customers can respond, helping inform credit assignment. So, whether a application for credit card should be approved, that’s called credit scoring. Also a use of predictive modeling or machine learning. Those types of business applications are generally also referred to as predictive analytics.

Eric Siegel: So the word prediction, by the way, obviously it means predicting the future, the outcome or behavior per let’s say customer. But sometimes it’s used like predicting, is this transaction going to turn out to be fraud? You’re not predicting the future. It’s just a classification. Predict whether this is a picture of a cat. You’re kind of warping the use of the word predict. But people do use it that way. In any event, when you go to that kind of image classification, people don’t usually call it predictive analytics, but either way it’s definitely machine learning.

Career Nation: Oh, thank you for that. That, that helps to clarify quite a bit and you know, it seems that, you know, Eric, based on what you just shared, it, there could be so many applications of this across industries, across companies. Where have you seen, sort of, people take this and apply it practically commercially to, you know, improve their business, improve their customer experience or help their employees?

Eric Siegel: Like what are some of the examples you’ve come across? Well, since we’ve been running the conference series predictive analytics world since 2009, we’re now in our 12th year. The bread and butter of that conference are real-world case studies from fortune-500 companies. And that’s sort of the whole point. So, I’ve seen a million of them. my book predictive analytics as you mentioned it. And by the way, I’ll just say the subtitle of the book again, because it’s an informal definition of the field. Predictive analytics, the power to predict who will click, buy, lie or die. That book includes a 181 mini case studies in a central color table, across all industry sectors. So, the kind of things I’ve been mentioning, targeting marketing, credit scoring, fraud detection, these are very common practices. All large companies or virtually all large companies use use machine learning. It’s value is also prevalent amongst mid-sized companies and many small companies also can benefit since the requirements, not so much about the size of the company, but about basically the size of the data. So if you are a small company sending direct mail to a large contact list, it’s the size of that contact list that matters with regard to whether there’s a value proposition to be gained by learning from this historical mailings, that is the training data and making predictions in order to target future mailings.

Career Nation: Got it. And let’s say I’m a, you know, a business analyst in a company or I’m a manager or take any title, right? And if I wanted to sort of get started and familiarize myself with machine learning and predictive, and if I wanted to sort of start figuring out how do I apply this, are there some tools that I can just sort of run on my laptop or is this some, you know, something I’d acquire like a cloud subscription? What are, like are there things that would help a professional to sort of get started in this area?

Eric Siegel: So yes, there’s many existing tools. Many of these companies exist as sponsors of our conference since we’re, we’re very much vendor neutral and the majority of the content, but we have these sponsored sessions, a few times a day. And, the trick here though is that it’s actually a little bit bigger scope, project to make use of and, and implement, integrate machine learning at least for a first time within your enterprise. Then, your question I think sort of implied. Because it’s not a matter of like, well I’ve got to get familiar with the tool and use it. there needs to be a large number of people, at least several on the team making predictive analytics work. The actual use of the core tool, the fun part, the scientifically interesting part, the rocket science part. That those tools of which there are many is actually a secondary to the first decision, which is how are we going to use it?

Eric Siegel: Which mass scale operation are we going to render more effective with these predictive scores that are output by the model. So for example, I’m doing these mass mailings and I want to target them better. I have a fraud detection team and I want to use their time better by serving them transactions, more likely to be fraud. I want to make better decisions about credit scoring as far as, you know, credit card or other loan applications. All these different kinds of business applications. Whatever it is, you need to decide within that realm. Very specifically, what are we doing now and how are we going to potentially change today’s operations, by informing them with or integrating into the process somehow the predictions that is the predictive scores, which are basically probabilities per individual of whatever outcome or behavior you’re predicting for the project, whichever serves to improve the efficacy of that large scale operations.

Eric Siegel: So, that’s the carrot at the end of the stick. That’s the deployment or the integration, what’s called the operationalization of the predictive model, right. At the end of the project or at the conclusion or the actual sort of deployment of the project. Bbt you, that’s the carrot at the end of the stick. But you start with that first very much and then you say, well, do I, what do I need to predict? And in order to predict that, do I have the right data available? You might do some preliminary data polls and get a sense of that. And then, okay, now we want to actually start doing the predictive modeling. We’re green-lighting the project. Now we need to see who do we have on board. This is something that, you know, this is within the realm of data scientists, which is a very subjective word, but oftentimes is used to refer to people who have experienced with predictive modeling and whoever is doing the actual number crunching and applying these software tools needs to have experience doing that, in the past.

Eric Siegel: So, sometimes you need to engage external resources consultancies and service or service providers and such, to help with that at least on a first project. Or are you doing more extensive training of some of your existing relatively technical staff or data oriented staff? So, as an individual you, you don’t sort of side, I’m going to do predictive analytics at my company. What you decide is my company would really benefit from predictive analytics. Let’s see how I can participate, how I can contribute. Because predictive analytics or these business applications of machine learning aren’t a technical endeavor, first and foremost. First and foremost, they’re an organizational change to existing processes. So it’s not just some kind of thing like, let’s put this technology in place and makes our website go faster. It’s not that. There’s engineering components to it, but they’re secondary to the fact that this is a change to organizational processes.

Eric Siegel: So you need to start with figuring out how is that process going to change? How would it be informed with predictions, right? Which aren’t necessarily like a crystal ball, but they’re better than guessing. And back in the napkin arithmetic, how is that going to help? So that organizational process, getting executive buy-in, getting buy in from operational managers who are, where the change of process is actually going to be taking place. And, and then enlisting the right team of people who can pull the right data, the analytics people, the actual data scientists or an external service provider. So there’s a team of several people. There’s involvement across the organization. And so that’s how you look at it. You don’t, it’s a, it’s a large scope process project.

Career Nation: Oh, that’s wonderful. And so this is a, this is, I love that, those examples, because these are a way to establish a discipline, a almost like a different approach to doing business within the company. And it really helps to sort of get grounded on, okay, what are the, what is the problem or the types of problem that we would like to solve, how big is that? What is the nature of that? Do we have enough data to solve that problem? And then based on those that sort of data inventory, we can figure out how do we apply some of these predictive models and then sort of drive those behaviors within the company and outside the company in many cases. That’s fascinating. And, and you clearly have had been having a lot of fun with this. You are, you’re running the predictive analytics world, which is a show in Las Vegas. You have written a wonderful book and we’ll, we’ll put the book link in the show notes below. And also, you are, you were, I think you were running the Dr. Data show. I saw a bunch of episodes there. So, tell us sort of what are the wonderful things you’ve been up to.

Eric Siegel: Yeah. If you want to hear more of the kind of stuff I’m describing about, well, how the technology works under the hood and, and some of the ins and outs. The Dr. Data show is a web series of 10 short episodes, so it’s 10 of the most interesting topics on machine learning, why it works, how it provides value, the ins and outs, how you evaluate to make sure the model has actually learned from the data rather than just kind of memorized it, or, or, or sort of found patterns that only apply in this particular set of examples. How do you know it’s actually learned in a way that’s universal, will apply in general, which is actually a pretty profound, almost philosophical question. But, the actual way you validate it isn’t philosophical, it’s very simple and pragmatic. You can actually just measure how well it works. So, all those kinds of ins and outs about the process are covered across these 10 episodes. And you can go to the or just Google it. So, that’s, that’s available online.

Career Nation: Wonderful. Eric, this has been awesome. And now this is a part of the show where we get to know a little bit better. And we would love to ask you some of your favorite things. So we’re getting into the favorite, quick fire round. Are you ready? It might, my favorite color is blue. Is that? So, that’s super helpful. So let me ask you a couple of specific questions. Other than your color, what is your favorite app and why?

Eric Siegel: Okay. I think there would be two. I would say ClassPass, which I think is only helpful if you live in an urban area. ClassPass is a way to, to sort of pick and choose and take, take an exercise class, at here and there without memberships and all that kind of stuff. So, I could just go and say, Oh, here’s a pretty good deal. I can just go take this one class tomorrow morning at 10 over at this Yoga studio or this sort of workout bootcamp place or what have you. And the reason I like it is because it, it, it, you’ve signed up for a certain number of credits you need to use per month and only some of them roll over. So, it helps enforce the discipline of getting yourself out there. Now I actually, am pretty disciplined.

Eric Siegel: I do sort of a, I go to the gym pretty much 365 days a year, but there’s, what is this from? So for me, the ClassPass thing is helping me, not just go to the gym and listen and listen, do work in my head by listening to podcasts or audio books or on the bike, like watching work-related videos. So, I’ll go take a yoga class and that’s a lot more challenging for me to sort of just be there and work in a slow, but physically assertive thing for an hour without actually getting work done at the same time. That’s actually hard for me, but it’s good for me. So that, I’d also say on the, on the flip side though, it’s also Audible, cause I, I don’t really read anymore. It’s all audio books.

Career Nation: Oh, that’s fascinating. I love that concept of ClassPass. Yeah, I’ll definitely download it. Do you have the other app as well as a favorite? Eric Siegel: The other, Career Nation: Well I think you mentioned two, right? So you’ve got a ClassPass. Audible is the other one? Got it. Cause I listen to audio books, audio books all the time in the gym. Yeah. Yeah. I love it man. Awesome. And what is your favorite quote? If you had to put up a quote on highway one-on-one or my favorite 680 or infamous 680. But, what would be your favorite quote?

Eric Siegel: My, my favorite quote, would have to be, spider man’s uncle who said, “With great power comes great responsibility.” And of course, in my case, I’m applying this to machine learning, which brings up a whole bunch of ethical issues. It is an extremely powerful technology that helps us automate mass scale decisions. It’s sort of the implementation of private and public public policies and automation and mechanization of those societal functions. It brings up a lot of ethical issues around social justice and how this affects all the people about whom decisions are being made. The way that we’re treated and served in modern society is dictated, more and more by predictive models that drive these decisions as, you know, they predict about us who will click, buy, lie or die, like in the title of my book and decide who to contact, investigate, incarcerate, or set up on a date.

Eric Siegel: Right? So it’s affecting us in all different ways. And some some of these decisions are extremely consequential with regard to who gets it, who gets approved for credit or a loan, or credit card. Who even gets offered it with marketing in the first place. And, and in some cases, you know how long you actually stay in prison in predictive policing. It’s all the same core technology, predictive modeling. And it’s to automate or semiautomate decision or inform or in one way or another to influence decisions, that are, you know, sometimes with a human in the loop and sometimes not depending on the area. So yeah, so, the spider man quote, you know, it’s paraphrase, paraphrasing Voltaire, but I know, I know originally from Spiderman and, and I, I take it to heart, actually, I think it’s really, sort of hits the nail on the head.

Career Nation: Oh, I totally agree. And there’s so many sort of layers to that that you, you mentioned. And you know, the, the, the, there’s opportunity on one hand where it, you know, we can do so many wonderful things with predictive, maybe we can predict diseases and things like that and really help a fellow human beings and communities and organizations and companies. And on the other hand, there’s probably never been a time earlier in history where so much sort of data and algorithms have been concentrated in only so many companies and companies that own those datasets, etc. So it’s, it’s going to be a very interesting now

Eric Siegel: world for sure. Yeah. I mean, I, I mean, I’m a proponent of the technology and extremely excited about its potential, positive value, not only for the bottomline in terms of profit. And although that often translates into benefits for individuals, but all sorts of social good applications. So it’s, it’s like a knife, right? It’s, we’re not going to outlaw it entirely. It can be used for good or bad. And so therefore there’s gotta be some management and oversight.

Career Nation: Oh, for sure. Which brings us to our next favorite question. What is your favorite book, Erik?

Eric Siegel: Well, the great novels, are the ones that stick with me. And I have the, so, you know, Herman Hesse’s Siddhartha and Steppenwolf, those ones really blew my mind and I didn’t read them for first time until I was in my early thirties. So, I had the fortune of, being kind a lopsided brain, through college, high school and college. I would kind of blow off all the reading as much as possible. My verbal wasn’t that high. I feel like I caught up though. And, maybe not, you know, to my geeky math side, but much better than it was back then. And so by my early thirties, I started, I was already sorta that developed, much more developed than you are when you’re 20, when I read these books for the very first time. So, that was a much more enriching experience.

Eric Siegel: I felt like I got all the levels a lot more than I would would’ve at the earlier age. So, those, that’s novels. But as far as sort of more business books. A couple of examples come to mind. Well, Geoffrey Moore Crossing the Chasm is like a pretty basic entrepreneurial book about, you know, if you’re, if you’re starting a new business or a new line of business, just the idea of a market niche and anecdotes about how that applies. That such a formative book and that one really stuck with me, especially coming from an academics background at that point. I’m real, I’m like, I’m a recovering academic, right. I, I, I, I, I’ve been in, in, in the industrial world since, 2001 or 2. And, and also just sort of being the sort of data geek side of things, but trying to see, okay, wait, there’s a business side and this, this, this one really, this one really stuck with me.

Eric Siegel: in terms of actual technical books, actually it’s a textbook and machine learning that stuck with me. And the title of the book is machine learning, but as the textbook by Tom Mitchell who was the founder and chair of the first machine learning department. Machine learning is usually within computer science, but it’s Carnegie Mellon has machine learning department. He came out with this textbook just in time for the first time I taught the graduate level course machine learning at Columbia university. And it’s just such a great way to sort of formulate and get a sense of this whole field, right? Not just get married to one particular thing with like deep learning kind of neural network that’s doing so well, but get a sense of what’s the overall field of, of, of machine learning and what are the, what are the universal concepts apply across all the different kinds of technical methods, predictive modeling methods, and sort of what are the universal requirements as far as the data preparation. So that sort of foundational structure and for understanding it as a field for me it was really set by that book.

Career Nation: Oh, those are wonderful, resources and, we’ll, we’ll track those down and put them in the show notes because it’d be really cool to take a look at that list. And, thanks for sharing that, Eric. And yeah. And why don’t, why don’t you share, your favorite restaurant?

Eric Siegel: Yeah, my favorite restaurant is just all the expensive sushi restaurants. But that’s a very, that’s a very specifically defined category of restaurant. And I’ve, I’ve, I’ve invented this, this category. It’s called expensive sushi. It’s not the kind of thing you’re going to do every day. But actually, one of the reasons I like it in a restaurant where the sushi is so good and it’s typically offered on omakaze, which, which means chef’s choice. It’s kind of like a tasting menu. You get smaller portions come gradually over a larger period of time, and then you end up taking more time to just appreciate how good each bite tastes, which is a very undervalued thing in our world, right? We’re, we’re constantly shoving Doritos into our mouth, right? So, nothing against Doritos, if Doritos are watching right now. But, so there’s this concept of mindful eating, right? And I read a book about it and I don’t remember who it was and the book was great, but you know, it’s not rocket science. It’s just think of, you know, actually pay attention to what your experience while you’re eating and that, that’s valuable for a lot for obvious reasons, if you think about it. So, and that sort of, that kind of restaurant experience, very conducive to that. And assuming with someone you like talking to, it’s a long meal though. It’s fun. Yeah.

Career Nation: Oh, that is so cool. I like that. And yes, apologies to Doritos and I liked that concept of mindful eating and that’s something that’s becoming more rare and it’s it’s time to get the mindful eating as well as the conversation back to the dinner table. I like that a lot. Yeah. Why don’t we shift gears a little bit Eric and talk a little bit about sort of your career and you had a phenomenal career. Very fascinating, colorful. You’ve done so many different wonderful things. How about we get into some of the techniques that you used in your career that you think are sort of unique and they’ve really helped you do basically either become better or organize projects better or you know, with your productivity you can move faster at work and things like that. How about some of the techniques that you feel that our audience would definitely appreciate?

Eric Siegel: You know, I’ve never, I’ve never tried to be a self-help guru and this, you know, thinking about, there’s so many things that I do that I could try to expound on that someone else might be find useful. So, generally, in my career, you know, as I said, I’m a former academic. And in academics you, you work really hard on concepts that it’s sort of seem ideal or abstractly sound. And that’s where my interest started with machine learning. Right? Which from a, just from a scientific point of view, I think it’s by far the most fascinating, interesting type of any kind of science technology or engineering. Cause you’re, you’re, it’s, these are methods, algorithms, computer programs that learn from example. So you’re actually automating the process of learning that is generalizing from some limit. It may be a long list, but it’s a limited number of examples. Eric Siegel: And how do you draw generalizations from that, that apply in general? That’s just such an interesting problem to try to solve. And there’s so many different facets, ins and outs to it. Although it’s the bottom line, it’s really simple to measure how well it works. It might be hard to design, how to do it, but when you’re trying it out, you could just try it on new data that, you know, you have this held a side data that it wasn’t used for the training part and say, well how well does it work on that? So you get instant gratification. You see how well it works. Fascinating area. Now does that mean it’s going to be viable when a, you know, a green academic like me 20 years ago steps out of the university and it’s like, I’m going to go use this and be a consultant.

Eric Siegel: Right? How do you convince people of this idea when they haven’t been indoctrinated into it abstractly and they’re just trying to run their business on a day-to-day process? That was a huge learning process for me, right? Where I, it’s not just about great ideas, but it’s about how you communicate them and how you socialize them and how, you know, how change management takes place. So it’s sort of like you know, and in my case, I was sticking to the ideals, to the principles. This is a great idea and it not only is fun and interesting scientifically, it really looks like it should be valuable. We can increase the profit of this marketing campaign by a factor of three is by not by changing any of the creatives or the product we’re selling, but just by targeting a little bit more effectively across customers predicted more likely to respond. That kind of business value proposition just sort of falls out naturally from it. Eric Siegel: Yeah. But, so that, that was in 2003. I first moved to the West coast and, and you know, started being a independent consultant in the business world. And you know, the world, back then we called it predictive analytics because machine learning was strictly an academic research and development terms. Nicely, that’s now become a more acceptable term in general. And the field certainly has taken off greatly. You know, when we launched Predictive Analytics World, that was the first, you know, conference, other than some run by the, the software vendors. But the first cross vendor conference outside the academic or research conferences, you know, focused on the commercial deployment. And we’ve stayed in the lead and we’ve stayed viable and only been growing since then, because we stay relevant and keep things going. But at the time, you know, we a little bit ahead of the curve. Eric Siegel: We didn’t know how many people would show up. Right. We, as it turned out, we hit the timing pretty well. It was, it was February, 2009. Even though we just hit a recession, we had a much bigger turnout than we had even hoped for because people were ready, this stuff was starting to warm up and now it’s hot. It might be a little too hot. There may be a little bit over promising here with the whole AI buzzword, but it’s very much realizing value in most commercial deployments. So, that would be one sort of takeaway as far as sort of stick to the principles that you believe in and sort of then look at the human side. How do you socialize it, how patient can you be, what are the tactics to do that, right? How, how, you know, how do you sort of, find that path, right? I mean, other than that, I’d say install Boomerang for Gmail, which is keep your inbox organized and tidy.

Career Nation: I love that. Yeah. Boomerang is one of my favorite tools as well. So, you know, what, what do you bring up is fascinating because a lot of times we come up with, let’s say, it could be a technology concept or a business concept and we are so convinced that this is going to change the world. And we’d go out there and pitch to customers and some customers bite and some don’t. And we sort of sort of, there’s something to be said about perseverance and tenacity to actually go out there and keep knocking on doors and actually working with customers. The other side of that that you also talked about a little bit and touched upon is sort of the human side of it. Which is, yes, the technology is great and the ROI is going to be great, but there’s this other part of it which is, you know, human beings are going to buy it. We at the end of the day, humans, we do business with human beings and it’s something about having that EQ and that emotional pull with customers as well, so that we, we figure out a way to how do I make this customer successful or the stakeholder successful in addition to just providing a great ROI for or great value to this company. I think that said, it’s a phenomenal learning.

Eric Siegel: Yep. And you can’t Ram it down their throats. So you can kind of make a graph that shows the promised ROI, but that doesn’t sell itself quite as quickly. It doesn’t sell itself on the schedule that you have in mind. Career Nation: That’s right. Because sometimes it takes time to realize the ROI and the value and quite frankly, people also want to see some visible returns in addition to just pure charts. And I think that’s where the magic is, Eric Siegel: right? So then over, you know, so I became that independent consultant, you know, hustling for clients for the first few years. In 2003 and conference started 2009 and its purpose was largely to do that and get as many brand name case studies with proven value, you know, from the trench stuff from, from, from the front lines as possible. My book is 2013 and then the updated in 2016. So, by 2013, having been in that world for 10 years, maybe I was overcompensating because as I mentioned earlier in our discussion today, I have 180, 183, a little mini case, like sort of one or two line case studies all, all in a compendium in the middle of the book and this color table, the central table because I was, I had trained myself to just work so hard to prove yes, this stuff actually works. It’s not just a good idea on paper. Look at all these examples where somebody got value, got success from it. You know, and it’s divided into seven or nine sub tables across all the different industries like marketing and financial sector, government, etc. So, that was sort of the result of that 10 years of me just sort of feeling like it was never enough to show yes, it works, it’s a good idea. Now everyone’s like expecting too much from it instead of too little.

Career Nation: Oh yeah. I mean it’s so wonderful and I’m sure you feel you must be feeling a lot of, Career Nation: some level of contentment and satisfaction that the thing that you started and you were one of the early pioneers of predictive analytics and you educated, consulted, helped, coached so many customers and clients. And now this thing is taking a life of its own. And now it’s a popular, I would say almost mainstream, you know, technology component, at least in Silicon Valley. And a lot of other tech companies, and you must be, you must be feeling, you know, hey, I’ve been vindicated. And you know, I was totally justified. I’m taking this to market that early. Sometimes in Silicon Valley, being early is sometimes being incorrect, but now you’ve been proven correct. And so it must be, must be a good feeling.

Eric Siegel: Yeah. And then the colleagues that I, that I, that I established in those early years of being consultant, you know, we’re still close colleagues, most of them participate at the conference as speakers and, or at least as attendees. And, so we all were sort of in the same boat. We’re kind of like, Oh, how do you explain this to people who are new to it or are nontechnical? And, and now we’re kind of looking at the world, like, what? Wow, this, this grew even faster than, you know, we even hoped over the last several years. It’s kind of amazing. But, you know, we always had no doubt that it should go this direction.

Career Nation: Eric, so this has been a phenomenal journey for you, right? You started with academia, consulting. You’re doing conferences, wrote a book and so many other things. What’s the future? What does the future hold for Eric? What is on the roadmap that you would like to share or if you’d like to make that a pleasant surprise. That’s cool as well.

Eric Siegel: Oh, well we’re continuing to grow the conference and keep it up to date and with all the hottest industry trends and case studies and stories. I’m working on a new updated online course about machine learning and the kinds of topics we talked about today. I have, I’ve never disclosed this, but I have a, so I have a second rap. So there’s already a rap we released a few years ago called ‘Predict This’ and it’s an educational, it’s the best ever educational rap music video about predictive analytics. You can go to, just three and a half minutes long. And we have another one, but that might take a couple of years by the time we actually, get it together, on a, on a special topics that I thought would be good. And I would say that outside sort of my consulting, my sort of central consulting career and the conference, it’s the ethical issues I mentioned. I’ve been writing op-eds on that in San Francisco Chronicle and Scientific American blog and some other places. You can see my list of about 10 op ed so far published. All it just go to, civilrightsdata and that gets you to that list, that linked list of articles. And I’m continuing to work on, on that very much. I have a lot more to say about, some of the social justice issues that underlie the deployment of predictive models.

Career Nation: Oh, that is so important. And thank you for leading the charge on that one, Eric. Now as we wrap up here, any parting thoughts, Eric, that you’d like to share with Career Nation? As you know, you know, we are an audience that’s, I would say, predominantly works in corporate America, a lot of us in Tech, business and we were sort of, some of us are early in career, middle of career, late in career. And so what, what would you like to share as your sort of parting words of wisdom for our audience? Eric Siegel: Yeah, I think that for those of you who are interested in machine learning or if you haven’t let go of the term AI yet after hearing me, by the way, I, I, I have a, a Dr. Data episode which is also on Big Think called AI is a big fat lie. So I am concerned about the misleading use of that term and many or most cases, but machine learning, supervised machine learning, very much a real thing. And if you’re interested in helping or getting involved with how it will provide value, your, your organization, my message to you would be, there are many roles, both managerial, nontechnical, involved in the operational deployment integration, the use, the consumption of the predictive model output, like its predictions. Its, its probability scores and outputs per individual or the supervising over the overall project. Or of course there’s the actual data side, the technical side, the data preparation, the corporate active modeling itself. Eric Siegel: There are so many different facets to the overall project and making sure that it’s running in a way collaboratively across viewpoints and across people in different roles at the organization. So, it’s not just one brilliant data scientist who knows how to do everything. Not at all. It can’t be, and there’s so many different ways you could potentially be involved. So, the core technology has many facets and its very involved, but the fundamental principles are not nearly as difficult to understand as you might imagine. They’re actually quite intuitive, which is, which the point of my book predictive analytics is to make that accessible and sort of unveil how it works under the hood. But then again, remind you, wait, this is about the value proposition to the business and how it gets deployed and used, not just the number crunching part. So, if you find the area promising or interesting or exciting, keep in mind as you kind of delve into it and learn some, that there are many different ways, that you could potentially be involved.

Career Nation: Yeah, there’s so many opportunities and Eric as you outlined and there’s so many roles in the world of predictive analytics and machine learning. It’s tech as well as non tech roles and those are quite frankly, we’re still early in the cycle and those are the sort of the leadership opportunities of the future and people who are getting involved with these concepts and frameworks and technologies have an opportunity to basically, you know, further their careers down the road. Exactly. Awesome. Eric, this has been fascinating. This is such a rich conversation. Thank you so much for your time. We wish you all the very best for predictive analytics world and we’ll drop a bunch of links and the notes below so that people can get in touch with you and know a little bit more about what are the wonderful things and projects that you’re up to. Awesome. Well thanks Abhijeet, it was great being on the show. Thanks for the great questions. Absolutely. Have a great day. Yeah, you too.