What do you think?
Rate this book
Predictive Analytics: The Power To Predict Who Will Click, Buy, Lie Or Die, Revised And Updated
Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, Revised and Updated
384 pages, Paperback
First published February 5, 2013
550 people are currently reading
4074 people want to read
About the author
Eric Siegel
13 books57 followersEric Siegel, Ph.D., is a leading consultant and former Columbia University professor who helps companies deploy machine learning. He is the founder of the long-running Machine Learning Week conference series and its new sister, Generative AI Applications Summit, the instructor of the acclaimed online course “Machine Learning Leadership and Practice – End-to-End Mastery,” executive editor of The Machine Learning Times, and a frequent keynote speaker. He wrote the bestselling Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, which has been used in courses at hundreds of universities, as well as The AI Playbook: Mastering the Rare Art of Machine Learning Deployment. Eric’s interdisciplinary work bridges the stubborn technology/business gap. At Columbia, he won the Distinguished Faculty award when teaching the graduate computer science courses in ML and AI. Later, he served as a business school professor at UVA Darden. Eric also publishes op-eds on analytics and social justice.
Eric has appeared on Bloomberg TV and Radio, BNN (Canada), Israel National Radio, National Geographic Breakthrough, NPR Marketplace, Radio National (Australia), and TheStreet. Eric and his books have been featured in BBC, Big Think, Businessweek, CBS MoneyWatch, Contagious Magazine, The European Business Review, Fast Company, The Financial Times, Forbes, Fortune, GQ, Harvard Business Review, The Huffington Post, The Los Angeles Times, Luckbox Magazine, MIT Sloan Management Review, The New York Review of Books, The New York Times, Newsweek, Quartz, Salon, The San Francisco Chronicle, Scientific American, The Seattle Post-Intelligencer, Trailblazers with Walter Isaacson, The Wall Street Journal, The Washington Post, and WSJ MarketWatch.
Eric has appeared on Bloomberg TV and Radio, BNN (Canada), Israel National Radio, National Geographic Breakthrough, NPR Marketplace, Radio National (Australia), and TheStreet. Eric and his books have been featured in BBC, Big Think, Businessweek, CBS MoneyWatch, Contagious Magazine, The European Business Review, Fast Company, The Financial Times, Forbes, Fortune, GQ, Harvard Business Review, The Huffington Post, The Los Angeles Times, Luckbox Magazine, MIT Sloan Management Review, The New York Review of Books, The New York Times, Newsweek, Quartz, Salon, The San Francisco Chronicle, Scientific American, The Seattle Post-Intelligencer, Trailblazers with Walter Isaacson, The Wall Street Journal, The Washington Post, and WSJ MarketWatch.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 160 reviews
April 5, 2013
"Predictive Analytics" is a summary of the state of the art in using computer models to predict individuals actions. I work in the industry and have developed predictive financial models. This book isn't aimed at people like me, at least not ones looking for a more technical, how-to explanation. Instead, this is more a survey of the field, including plentiful real-world examples and some high-level definitions. The definitions of lift, ensemble modeling, and uplift modeling I found new and interesting, although the explanation of uplift modeling is a bit of a miss. I think the value in the book is in the examples. There are many examples of PA in action, some short descriptions and some, like IBM's Watson, taking a full chapter. If you think better with examples, this is a good book but with a caveat. The examples are very high level and don't contain much if any technical detail. This is best for examples of what can be done, not how to do it. I also found it strange that I got the feeling that the author presents PA as something new, when based on his definitions and examples, some of these concepts and methods and projects are decades old. You get the impression that you are being sold something as NEW, when in fact it is just WHERE WE ARE NOW in the technology. This absolutely meets my criteria as a good business book -- it gives you the concepts to think in new ways about problems, and it provides examples that can be generalized into a number of different applications throughout a company.
March 30, 2013
I wish I could have predicted how much I would dislike this book. After reading just one chapter of Nate Silver's The Signal and the Noise this book comes across as amateurish. Too much noise, not enough signal.
June 21, 2021
Not thoughtfully written and shallowly propagandistic. It joins so much hype and adds little to the brimming pot. The last couple of chapters are more digestible, but still doesn't do much beyond illuminating the very basics of PA. Here's a bit of my review for book club:
I’ll start with positives: Love the suggestion that marketing departments that manufacture quasi-medical data should have to deal with it in a substantive way. That is a solution I haven’t heard proposed yet, but what if HIPAA, FERPA, etc. and all those “cumbersome bureaucratic” measures could weigh down every advertising outfit and data brokerage firm and social media giant that is able to manufacture sensitive data about people outside their domain? “You made it, you manage it.” (Author’s italics.) (Chapter 2, Good Prediction, Bad Prediction)
Another important caveat he does actually cover is the following: there can be heightened human trust in technical systems (as with judges and parole boards in Chapter 2, Machine Risk Without Measure). “What may render judges better informed could also sway them toward less active observation and thought, tempting them to defer to the technology...and grant it undue credence." So we can say that automated decision support systems can actually undermine the likelihood for the humans you’re organizing to interact in the way that you are trying to prompt them to.
So my substantive suggestion is more about how we want to treat “data science”. Based on what I’m gathering from this book, I suggest that people who practice predictive analytics should be called “mathematicians”, “statisticians”, or “predictive analysts”. As Siegel himself says, “PA’s mission is to engineer solutions,” and “Whatever works.” It seems to be about finding correlative relationships that help figure out how to get the right people to see the right ads. (But let’s be real — that’s largely what data scientists are spending their talent on.) It doesn’t sound like they’re doing science. It’s just a method that may be used in science, but is largely used elsewhere as well.
Siegel provides a helpful foil for me to make this argument. Because it was part of the established practice of science, the scientific study done by Gilbert and Karahalios of stock market and blog post anxiety measures received criticism until it found a way to establish causation, or predictive direction. The work that statisticians/predictive analysts are doing doesn’t come under this scrutiny if they are in business. The assertion that we don’t care what’s under the hood because the “black box” just gives us the predictions is anti-scientific. Predictive analysts don’t have to start with a theory, or even come to a theory when they find relationships in the data.
Sure, this method can be part of actual science, but it clouds up the meaning to call them scientists because they’re just doing math in law enforcement, healthcare, insurance, and human resources, and so forth. But calling it science artificially colors what is being done in these different domains — it’s possible that none of it is part of the scientific process because it may not ever be based on a theory, it may not be done in public or with the benefit of the scientific community’s oversight. It’s not necessarily building upon itself toward more insights.
Predictive analysts work towards enough prediction to make companies more money, predict more accidents, etc. But it doesn’t seem to ask why. It's not done as a search for how the world really works. PA is used to drive decisions, says Siegel. It’s a method that is about decision-making without wisdom or understanding. Siegel’s own repetition about the primacy of advertising assures me that as a method, it’s bound up with making profit in a way that even could undermine its ability to be used for science. “Benjamin Franklin forgot to include [advertising] when he proclaimed ’Nothing can be said to be certain, except death and taxes.” (Chapter 1)
But it unfairly receives the automatic clout assigned to anything with the word “science” in it.
I’ll start with positives: Love the suggestion that marketing departments that manufacture quasi-medical data should have to deal with it in a substantive way. That is a solution I haven’t heard proposed yet, but what if HIPAA, FERPA, etc. and all those “cumbersome bureaucratic” measures could weigh down every advertising outfit and data brokerage firm and social media giant that is able to manufacture sensitive data about people outside their domain? “You made it, you manage it.” (Author’s italics.) (Chapter 2, Good Prediction, Bad Prediction)
Another important caveat he does actually cover is the following: there can be heightened human trust in technical systems (as with judges and parole boards in Chapter 2, Machine Risk Without Measure). “What may render judges better informed could also sway them toward less active observation and thought, tempting them to defer to the technology...and grant it undue credence." So we can say that automated decision support systems can actually undermine the likelihood for the humans you’re organizing to interact in the way that you are trying to prompt them to.
So my substantive suggestion is more about how we want to treat “data science”. Based on what I’m gathering from this book, I suggest that people who practice predictive analytics should be called “mathematicians”, “statisticians”, or “predictive analysts”. As Siegel himself says, “PA’s mission is to engineer solutions,” and “Whatever works.” It seems to be about finding correlative relationships that help figure out how to get the right people to see the right ads. (But let’s be real — that’s largely what data scientists are spending their talent on.) It doesn’t sound like they’re doing science. It’s just a method that may be used in science, but is largely used elsewhere as well.
Siegel provides a helpful foil for me to make this argument. Because it was part of the established practice of science, the scientific study done by Gilbert and Karahalios of stock market and blog post anxiety measures received criticism until it found a way to establish causation, or predictive direction. The work that statisticians/predictive analysts are doing doesn’t come under this scrutiny if they are in business. The assertion that we don’t care what’s under the hood because the “black box” just gives us the predictions is anti-scientific. Predictive analysts don’t have to start with a theory, or even come to a theory when they find relationships in the data.
Sure, this method can be part of actual science, but it clouds up the meaning to call them scientists because they’re just doing math in law enforcement, healthcare, insurance, and human resources, and so forth. But calling it science artificially colors what is being done in these different domains — it’s possible that none of it is part of the scientific process because it may not ever be based on a theory, it may not be done in public or with the benefit of the scientific community’s oversight. It’s not necessarily building upon itself toward more insights.
Predictive analysts work towards enough prediction to make companies more money, predict more accidents, etc. But it doesn’t seem to ask why. It's not done as a search for how the world really works. PA is used to drive decisions, says Siegel. It’s a method that is about decision-making without wisdom or understanding. Siegel’s own repetition about the primacy of advertising assures me that as a method, it’s bound up with making profit in a way that even could undermine its ability to be used for science. “Benjamin Franklin forgot to include [advertising] when he proclaimed ’Nothing can be said to be certain, except death and taxes.” (Chapter 1)
But it unfairly receives the automatic clout assigned to anything with the word “science” in it.
August 13, 2013
This book is extremely introductory, which accounts for Siegel's 50,000-foot view of the topic. Yet, I came away feeling there could have been more details on the "how" of predictive analytics without destroying the book's aim of being an overview.
Rather than droning on about IBM's Watson, I thought Siegel could have spent a little more time explaining the logic behind building decision trees and preparing the training data. Instead, we get about 100 pages of fluff out of a 217-page text. A typical chapter starts with an interesting tease story and quickly veers off into a "let me tell you about my grandchildren" tangent before the reader gets to the real meat, what little there is, the book has to offer.
If you already know that PA involves probability, massive data sets, and decision trees, you already know 99% of what Siegel says in this book. As a matter of fact, Siegel sums up the whole book in a one page appendix of five principles. Just go to the bookstore and read that page and you will have the essence of this volume and will have saved yourself $20.
Rather than droning on about IBM's Watson, I thought Siegel could have spent a little more time explaining the logic behind building decision trees and preparing the training data. Instead, we get about 100 pages of fluff out of a 217-page text. A typical chapter starts with an interesting tease story and quickly veers off into a "let me tell you about my grandchildren" tangent before the reader gets to the real meat, what little there is, the book has to offer.
If you already know that PA involves probability, massive data sets, and decision trees, you already know 99% of what Siegel says in this book. As a matter of fact, Siegel sums up the whole book in a one page appendix of five principles. Just go to the bookstore and read that page and you will have the essence of this volume and will have saved yourself $20.
May 22, 2014
Okay at best. He clearly knows his stuff and has great experience to talk about. He also chooses interesting examples of predictive modeling that he hasn't worked on. But his style is self-absorbed and immature. If you are in the business, you will get something out of reading his book, but you probably won't enjoy it.
June 10, 2013
PREDICTIVE ANALYTICS by ERIC SIEGEL
Having no previous knowledge of predictive analytics, I was a little afraid this book might leave me bewildered. How wrong I was! My eyes were opened, my interest caught and held throughout this fascinating book.
There are many questions that come to mind when reading this book, but as you read on they are all very effectively answered by the author.
Predictive analytics are rooted in everyone’s daily lives and can have a substantial effect on their future actions. I like the way Eric Siegel explains, giving examples that can be related to, so that even a total novice like myself has some insight into this fascinating subject.
This book is a must for anyone working in marketing. Even if they have previously explored this area, this book will open their eyes to further insight and could prove to be invaluable. It is also a must for anyone wanting to understand how predictive analytics can work.
I particularly liked the chapter on The Data Effect. Predicting the mood of Blog posts was fascinating, as a blogger myself this held my interest. As for the Far Out, Bizarre and Surprising Insights, well you simply have to read it! I devoured every word! Can early retirement really decrease life expectancy? What does your web browsing signify? This book will reveal all and it is written in such a way to hold the readers interest from start to finish.
What effect do predictions have on the business world? What predictions do famous names such as Google, Facebook, Citybank and others make? There is so much to discover in this easy to read and understand book. Anyone interested in the world of analytics will find this fascinating.
I was surprised at how much I enjoyed this book. Very well explained Dr Siegel! I think this deserves five stars.
Having no previous knowledge of predictive analytics, I was a little afraid this book might leave me bewildered. How wrong I was! My eyes were opened, my interest caught and held throughout this fascinating book.
There are many questions that come to mind when reading this book, but as you read on they are all very effectively answered by the author.
Predictive analytics are rooted in everyone’s daily lives and can have a substantial effect on their future actions. I like the way Eric Siegel explains, giving examples that can be related to, so that even a total novice like myself has some insight into this fascinating subject.
This book is a must for anyone working in marketing. Even if they have previously explored this area, this book will open their eyes to further insight and could prove to be invaluable. It is also a must for anyone wanting to understand how predictive analytics can work.
I particularly liked the chapter on The Data Effect. Predicting the mood of Blog posts was fascinating, as a blogger myself this held my interest. As for the Far Out, Bizarre and Surprising Insights, well you simply have to read it! I devoured every word! Can early retirement really decrease life expectancy? What does your web browsing signify? This book will reveal all and it is written in such a way to hold the readers interest from start to finish.
What effect do predictions have on the business world? What predictions do famous names such as Google, Facebook, Citybank and others make? There is so much to discover in this easy to read and understand book. Anyone interested in the world of analytics will find this fascinating.
I was surprised at how much I enjoyed this book. Very well explained Dr Siegel! I think this deserves five stars.
June 22, 2015
000 stars! what horrifying writing, the middle school me wrote better essays than this guy. I couldn't go pass 8% of this book before quiting. why can't he just stay on point and rationally dissect each point one by one instead of floating all over the place and being lost in his long winded sentences. Dude, maybe you need a crash course in writing in tue social sciences.
October 26, 2018
Worth a read for those interested in analytics but want an entry level book.
October 18, 2017
This is a painfully badly written book with very layman explanation of the subject. I scanned through first 150 pages reading the same very examples mentioned by lots of other authors. Only 20 pages of this book were worth reading (and still were extremely difficult to consume). And if you're really interested in PA – use your time on Udacity courses – you'll get much more!
I would not recommend this book – mostly, it's a waste of your time.
I would not recommend this book – mostly, it's a waste of your time.
September 17, 2017
Provides a high level overview of the possibilities of predictive analytics. As someone aspiring to be able to do this type of work, I will definitely be accessing the resources provided throughout the book to obtain more specifics.
November 4, 2018
I give this book 3.5 stars. The first few chapters is familiar as the concepts and stories have already been covered in many Big Data books. Some of the information is useful but you have to trudge through some dull writing.
September 15, 2018
Not bad but a bit dated now
June 3, 2021
Watch a detailed review along with my favorite ideas and takeaways at:
https://youtu.be/88sQOiIK6zo
https://youtu.be/88sQOiIK6zo
February 17, 2020
I am starting to suspect that the authors get paid by the word, that's why they make their books so UNNECESSARY long.
August 23, 2018
Eric has a difficult job which is treading a fine line between being a business overview and a techie text book and I think he gets it right. If you are looking for a high level overview or a techie bible this isn't for you. Eric does walk you through where PA is being used today and how somethings things are not what they seem. There are some "decision trees" but that is more to get you thinking, rather than baffle you with tech. The client examples does help brief the subject to life and while this book is not the easiest of books to read, to be fair, I did read it on holiday.
August 23, 2018
Great book, love how it tells the story of AI growing in its industry applications. It gives some intuition of what's behind the models through examples of possible applications. Great start for newcomers in the field, especially those from other areas looking to work better with technical people.
July 19, 2014
As more and more companies try to harness the power of 'Big Data' - the latest business buzz word - books like Siegel's are helpful to get a grasp on just what it is. This book is less 'how to' than an attempt to explain what it is, and how it can work for you, with the latter point venturing a bit too close to hucksterism at times (hey, it is Siegel's field). Siegel does a good job explaining how valuable data is and convinces us that with smart, predictive modeling, data can change how we market to customers and increase our efficacy on the business front. I appreciated him tackling the different types of predictive models from a high-level and laying it out in a way that is entirely digestible to someone at arm's length.
A few drawbacks of the book are its limited case studies or not investing enough into the ones he does present. A Bloomberg cover article on Obama's celebrated analytic team prompted me to read this book and I was looking forward to a deep-dive into it, however Siegel shorts us here with a quick 5 page highlight. And a very minor point, the illustrations are too Microsoft Clipart to be taken serious.
A few drawbacks of the book are its limited case studies or not investing enough into the ones he does present. A Bloomberg cover article on Obama's celebrated analytic team prompted me to read this book and I was looking forward to a deep-dive into it, however Siegel shorts us here with a quick 5 page highlight. And a very minor point, the illustrations are too Microsoft Clipart to be taken serious.
May 20, 2019
A refreshing reminder of how much I love my job.
August 18, 2015
This was a good overview of the way machine learning is used to improve decision making across a wide range of disciplines from medical, marketing, and politics. It is written in choppy sequences that are separated by quotes, some of the quotes are actually just songs that the author made up. The intended audience of the book is not scientist level. But as an introduction to the field it is very good, and I could see how the methods I know could be used to make these models, even though the author never goes into detail about the math behind the models. After reading this book you will probably learn something about how predictive analytics affects you, but you will have no idea how to make a computer do machine learning.
February 25, 2016
- Re-read: no
- Grade: 3
As a PhD candidate in text mining… I can’t say I learnt much from this book. However, I still strongly recommend it for say lay men. The book is written in a vibrant tone, making you care. You will learn tons from it.
I didn’t know Nate Silver. It was good to know about his work on the Obama re-election, and the power that statistics and data predictions can have in our lives. Deciding our most primitives desires and life turning points.
Also a reminder for me: I need to work more on ensemble methods! The wisdom of the crowds!
---
Addendum!
- Big Data is not about the sheer data size, it’s about how freaking fast it grows every day!
- Predictive Analytics == Machine Learning, for lay men
- Grade: 3
As a PhD candidate in text mining… I can’t say I learnt much from this book. However, I still strongly recommend it for say lay men. The book is written in a vibrant tone, making you care. You will learn tons from it.
I didn’t know Nate Silver. It was good to know about his work on the Obama re-election, and the power that statistics and data predictions can have in our lives. Deciding our most primitives desires and life turning points.
Also a reminder for me: I need to work more on ensemble methods! The wisdom of the crowds!
---
Addendum!
- Big Data is not about the sheer data size, it’s about how freaking fast it grows every day!
- Predictive Analytics == Machine Learning, for lay men
November 26, 2014
This book was full of great examples and was written in a humorous and approachable way. I am somewhat confused by the reviews that say The Signal and The Noise by Nate Silver was better - I found that book desperately in need of an editor. The focus of The Signal and the Noise was also broader, explaining basic statistics, correlation vs. causation, etc. Predictive Analytics was a focused book filled with examples of PA being used successfully.
The cover art is awful and the font size a bit too small. The content is excellent, though.
The cover art is awful and the font size a bit too small. The content is excellent, though.
May 15, 2019
Полезное введение в историю машинного обучения и основных методов. Рада, что прочитала. Последние 90 страниц книги - это приложения ))
April 16, 2013
It's an easy reading book for quite heavy topic. Even though most of the topics are not new to me but at least it taught me how to explain predictive analytics in easy term.
One thing that I learn most is the last chapter about uplift modeling: not predicting the response, but only focusing those who can be influence through contact.
Overall, it's a recommended book for business leaders who wants to double or triple their ROE using analytics.
One thing that I learn most is the last chapter about uplift modeling: not predicting the response, but only focusing those who can be influence through contact.
Overall, it's a recommended book for business leaders who wants to double or triple their ROE using analytics.
January 4, 2014
A rather interesting book that outlines different prediction methods, but he never gets into the mathematical aspect of predictive analytics. Overall it's a great book for beginners getting into the field and contains ample amounts of information regarding predictions. It also gets into what predictive analytics is being used for today and by what types of companies. It was a rather enlightening read on the subject.
July 29, 2014
The content is actually a good, high-level, non-technical overview of the field and the ways data can be used in business.
But the writing. Oh goodness, the writing. So many paragraphs feel like the work of a high-schooler just out of "essays 101". Chapters begin with word clouds and quotes (which normally make me shy from a book), and some of those quotes are from the author! I had to force myself to make it to the end.
But the writing. Oh goodness, the writing. So many paragraphs feel like the work of a high-schooler just out of "essays 101". Chapters begin with word clouds and quotes (which normally make me shy from a book), and some of those quotes are from the author! I had to force myself to make it to the end.
January 23, 2016
Lots of filler. This probably could have been condensed to 50 pages or less. I was expecting the book to be a little more technical, or to get into any kind of detail on HOW to conduct predictive analytics (PA). Instead, the author presented multiple examples of how PA has been used or could be used to predict business outcomes. I probably would have liked the book more if I had absolutely no knowledge of the subject area.
October 16, 2018
This book provides a solid, but high-level, overview of the field of predictive analytics. It is not a practitioner's guide if that is what you are seeking. Rather it explores various categories of applications of the technology, citing numerous interesting examples of each. It is not a very long book, especially given the numerous charts, tables and graphics included. However, it is a good starting point for those who want to establish some familiarity of this field.
July 25, 2017
This is a nice, entertaining book that gives someone an overview about what big data analytics are and how they can be used. it is written so that a novice can understand. I enjoyed the quirky quotes and the various case studies. That said, this book will not help someone who wants to delve deeper into how to actually create these algorithms.
August 5, 2015
You've probably heard the story of a Target store figuring out a teenager was pregnant before her family did, and you may have seen Watson beat Ken Jennings at Jeopardy. You may even have some knowledge of A/B testing. This book not only explains how these work -- without getting technical, but with conceptual clarity -- but shows a vast array of related applications.
May 29, 2017
Siegel's jokey and friendly tone help make this a simple "beginners guide" to PA. by breaking down the often complicated concepts with colloquial language and providing real-life case studies, it becomes an easier read than it should be a good crash course in the basic applications and tenants of predictive analytics.
Displaying 1 - 30 of 160 reviews