No. 252 of 500 No. 31 in AI for Beginners

Predictive Analytics

The Power to Predict Who Will Click, Buy, Lie, or Die

By Eric Siegel

Print length
368 pages
Publisher
Wiley
Reading level
Beginner
ISBN-13
978-1119145677
Published
January 20, 2016
Written for
Managers, marketers, students and general readers wanting a plain-language introduction to predictive machine learning.

Synopsis

Eric Siegel, a former Columbia University computer science professor and the founder of the Predictive Analytics World conference series, explains for general readers how machine learning turns the data organizations already collect into predictions about individuals: who will click an ad, buy a product, default on a mortgage, commit fraud or fall ill. With a foreword by Thomas H. Davenport, the book is organized around named effects, including the prediction effect, the data effect and the ensemble effect, which show why even modestly accurate predictions pay off when applied across millions of decisions.

Chapters move from deployment and ethics, with cases involving Target, HP and law enforcement, to data quality and false discoveries, how a model learns from mortgage data at Chase, how combining models helped teams compete for the Netflix Prize, how IBM's Watson answered Jeopardy! questions, and how uplift modeling identifies the customers or voters a campaign can actually persuade, with examples from Telenor, U.S. Bank and the Obama campaign. A color insert catalogs 182 real applications across nine industry groups, and a revised and updated edition followed in 2016. Readers come away knowing what prediction can and cannot do, which questions to put to a data science team, and where privacy concerns arise.

Scorecard

Editorial score

How it scored

Editorial score: 69.3 / 1005 weighted criteria

  • Influence30%

    22.5/30
  • Authority25%

    18.8/25
  • Practical value20%

    12.0/20
  • Currency15%

    7.5/15
  • Readability10%

    8.5/10
  • Editorial score

    69.3/100

How the rank was computed

Rank No. 252 of 500 comes from a composite of 54.8: observable signals 23.7 of 55 points plus editorial review 31.1 of 45 points.

All 11 inputs
InputPoints
Observable signals23.7/55
Reach (Amazon ratings volume)10.5/16.5
Consensus across best-of lists8.3/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted4.2/8.3
Amazon sales rank0.7/5.5
Recency0.0/2.8
Editorial review31.1/45
Author authority8.4/11.3
Influence10.1/13.5
Practical value5.4/9.0
Readability3.8/4.5
Currency3.4/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A lively, nontechnical tour of how organizations use machine learning to predict individual behavior, still useful though its examples have aged.

Strengths

  • Memorable case studies from Chase, Netflix and IBM
  • Explains machine learning without equations or code
  • Serious treatment of ethics and privacy

Keep in mind

Its case studies largely date from the early 2010s, and it gives little attention to deep learning or generative AI, so technical readers will find it basic.

Key takeaways

  1. A little prediction goes a long way: even modest accuracy gains produce large returns when applied across millions of customer, patient or citizen decisions.
  2. Combining many models in an ensemble usually beats any single model, a lesson Siegel illustrates through the Netflix Prize competition.
  3. Uplift modeling predicts who will change behavior because of an intervention, letting marketers and campaigns target persuadable people instead of likely buyers.
  4. Models can infer sensitive facts, such as pregnancy, from ordinary transaction data, so deploying them carries real privacy and civil liberties obligations.

Notable facts

About the author

Eric Siegel

Eric Siegel, PhD, is a former Columbia University professor who won the Distinguished Faculty award for teaching graduate computer science courses in machine learning and AI, and later taught at UVA Darden. He founded the Predictive Analytics World and Machine Learning Week conference series, is executive editor of The Machine Learning Times, and wrote The AI Playbook (MIT Press, 2024).

Questions

What does Predictive Analytics by Eric Siegel cover?
It explains, without equations, how organizations use machine learning to predict individual behavior, such as clicks, purchases, fraud, loan defaults and health outcomes. Siegel walks through real cases from Chase, Target, Netflix, IBM's Watson and the Obama campaign, introduces ideas such as ensembles and uplift modeling, and discusses the privacy and civil liberties questions that prediction raises.
Is Predictive Analytics still worth reading in 2026?
As a conceptual introduction, yes. Its explanations of why prediction pays off, how models learn from data, why ensembles work and how uplift modeling targets persuadable people remain accurate. Its examples date from the early 2010s and it does not cover deep learning or generative AI, so readers should pair it with Siegel's The AI Playbook or another recent book on deployment.
Who should read Predictive Analytics?
Business managers, marketers, policy staff and students who need to understand what predictive machine learning does, and how to discuss it with data scientists, are its core audience. It assumes no math or programming. Practicing data scientists will find the concepts familiar, though its case studies can help them explain projects to executives and colleagues.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.