Competing on Analytics: Updated, with a New Introduction

The New Science of Winning

Thomas Davenport, Jeanne Harris

• Publisher: Harvard Business Review Press
• Year: 2017
• Print ISBN: 9781633693722
• eBook ISBN: 9781633693739
• Topic: Business
• Subtopic: Decision Making
• Index: Business

The New Edition of a Business Classic

This landmark work, the first to introduce business leaders to analytics, reveals how analytics are rewriting the rules of competition.

Updated with fresh content, Competing on Analytics provides the road map for becoming an analytical competitor, showing readers how to create new strategies for their organizations based on sophisticated analytics. Introducing a five-stage model of analytical competition, Davenport and Harris describe the typical behaviors, capabilities, and challenges of each stage. They explain how to assess your company's capabilities and guide it toward the highest level of competition. With equal emphasis on two key resources, human and technological, this book reveals how even the most highly analytical companies can up their game.

With an emphasis on predictive, prescriptive, and autonomous analytics for marketing, supply chain, finance, M&A, operations, R&D, and HR, the book contains numerous new examples from different industries and business functions, such as Disney's vacation experience, Google's HR, UPS's logistics, the Chicago Cubs' training methods, and Firewire Surfboards' customization. Additional new topics and research include:

• Data scientists and what they do
• Big data and the changes it has wrought
• Hadoop and other open-source software for managing and analyzing data
• Data products—new products and services based on data and analytics
• Machine learning and other AI technologies
• The Internet of Things and its implications
• New computing architectures, including cloud computing
• Embedding analytics within operational systems
• Visual analytics

The business classic that turned a generation of leaders into analytical competitors, Competing on Analytics is the definitive guide for transforming your company's fortunes in the age of analytics and big data.

PART 1

THE NATURE OF ANALYTICAL COMPETITION

CHAPTER ONE

THE NATURE OF ANALYTICAL COMPETITION

USING ANALYTICS TO BUILD A DISTINCTIVE CAPABILITY

In 1997 a thirty-something man whose résumé included software geek, education reformer, and movie buff rented Apollo 13 from the biggest video-rental chain on the block—Blockbuster—and got hit with $40 in late fees. That dent in his wallet got him thinking: why didn’t video stores work like health clubs, where you paid a flat monthly fee to use the gym as much as you wanted? Because of this experience—and armed with the $750 million he received for selling his software company—Reed Hastings jumped into the frothy sea of the “new economy” and started Netflix, Inc.

Pure folly, right? After all, Blockbuster was already drawing in revenues of more than $3 billion per year from its thousands of stores across America and in many other countries—and it wasn’t the only competitor in this space. Would people really order their movies online, wait for the US Postal Service (increasingly being referred to as “snail mail” by the late 1990s) to deliver them, and then go back to the mailbox to return the films? Surely Netflix would go the route of the many internet startups that had a “business model” and a marketing pitch but no customers.

And yet we know that the story turned out differently, and a significant reason for Netflix’s success today is that it is an analytical competitor. The online content creation and distribution company, which has grown from $5 million in revenues in 1999 to $8.3 billion in 2016, is a prominent example of a firm that competes on the basis of its mathematical, statistical, and data management prowess. Netflix streams a wide range of content—including movies, television shows, documentaries, and original programming—to over 93 million subscribers in 190 countries worldwide. Every minute, Netflix customers stream 69,444 hours of video. Customers watch their cinematic choices at their leisure; there are no late fees.

Netflix employs analytics in two important ways, both driven by customer behavior and buying patterns. The first is a movie-recommendation “engine” called Cinematch that’s based on proprietary, algorithmically driven software. Netflix hired mathematicians with programming experience to write the algorithms and code to define clusters of movies, connect customer movie rankings to the clusters, evaluate thousands of ratings per second, and factor in current website behavior—all to ensure a personalized web page for each visiting customer.

Netflix also created a $1 million prize for quantitative analysts outside the company who could improve the Cinematch algorithm by at least 10 percent. It was an innovative approach to crowdsourcing analytics, even if the winning algorithm was too complex to fully adopt. But no doubt Netflix’s data scientists learned from the work and improved the company’s own algorithms. CEO Reed Hastings notes, “If the Starbucks secret is a smile when you get your latte, ours is that the website adapts to the individual’s taste.”¹ Netflix analyzes customers’ choices and customer feedback on the movies they have viewed—over 1 billion reviews of movies they liked, loved, hated, and so forth—and recommends movies in a way that optimizes the customer’s taste. Netflix will often recommend movies that fit the customer’s preference profile but that aren’t in high demand. In other words, its primary territory is in “the long tail—the outer limits of the normal curve where the most popular products and offerings don’t reside.”²

Now that Netflix is solidly in the business of creating new entertainment, the company has used analytics to predict whether a TV show will be a hit with audiences before it is produced. The most prominent example of Netflix’s predictive efforts is House of Cards, the company’s first original series. The political drama stars Kevin Spacey and is now entering its fifth season. Netflix has spent at least $200 million producing it thus far, so it’s a big decision. The company doesn’t release viewership figures, but the show is widely regarded as a home run. And it’s not by accident. Netflix employed analytics to increase the likelihood of its success. It used attribute analysis, which it developed for its movie recommendation system, to predict whether customers would like the series, and has identified as many as seventy thousand attributes of movies and TV shows, some of which it drew on for the decision whether to create it:

• Netflix knew that many people had liked a similar program, the UK version of House of Cards
• It knew that Kevin Spacey was a popular leading man
• It knew that movies produced or directed by David Fincher (House of Cards’ producer) were well liked by Netflix customers

There was certainly still some uncertainty about investing in the show, but these facts made for a much better bet. The company also used predictive analytics in marketing the series, creating ten different trailers for it and predicting for each customer which one would be most likely to appeal. And of course, these bets paid off. Netflix is estimated to have gained more than 3 million customers worldwide because of House of Cards alone.

And while we don’t know the details of Netflix’s analytics about its other shows, it seems to be using similar approaches on them. Virtually all of the original shows Netflix produced were renewed after their first seasons—the company’s batting average is well over .900. In addition, Netflix has had many shows nominated for Emmys and has won its fair share as well.

Like most analytical competitors, Netflix has a strong culture of analytics and a “test and learn” approach to its business. The chief product officer, Neil Hunt, notes,

From product management all the way down to the engineering team, we have hired for and have built a culture of quantitative tests. We typically have several hundred variations of consumer experience experiments running at once. For example, right now we’re trying out the “Netflix Screening Room,” which lets customers see previews of movies they haven’t seen. We have built four different versions of that for the test. We put twenty thousand subscribers into each of four test cells, and we have a control group that doesn’t get the screening room at all. We measure how long they spend viewing previews, what the completion rate is, how many movies they add to their queue, how it affects ratings of movies they eventually order, and a variety of other factors. The initial data is quite promising.³

Reed Hastings has a master’s in computer science from Stanford and is a former Peace Corps math teacher. The company has introduced science into a notably artistic industry. As a BusinessWeek article put it, “Netflix uses data to make decisions moguls make by gut. The average user rates more than 200 films, and Netflix crunches consumers’ rental history and film ratings to predict what they’ll like . . . ‘It’s Moneyball for movies, with geeks like Reed [Hastings] looking at movies as just another data problem,’ says Netflix board member Richard N. Barton.”⁴

In its testing, Netflix employs a wide variety of quantitative and qualitative approaches, including primary surveys, website user testing, concept development and testing, advertising testing, data mining, brand awareness studies, subscriber satisfaction, channel analysis, marketing mix optimization, segmentation research, and marketing material effectiveness. The testing pervades the culture and extends from marketing to operations to customer service.

Netflix may seem unique, but in many ways it is typical of the companies and organizations—a small but rapidly growing number of them—that have recognized the potential of business analytics and have aggressively moved to realize it. They can be found in a variety of industries (see figure 1-1). Some are not widely known as analytical competitors. Others, like Netflix, Caesars Entertainment in the gaming industry, or the Oakland A’s in baseball, have already been celebrated in books and articles. Some, such as Amazon and Google, are digital powerhouses that have harnessed the power of the internet to their analytical engines. Others, such as AB InBev and Procter & Gamble, have made familiar consumer goods for a century or more. These companies have only two things in common: they compete on the basis of their analytical capabilities, and they are highly successful in their industries. These two attributes, we believe, are not unrelated.

FIGURE 1-1

---

Analytic competitors are found in every industry

---

What Are Analytics?

By analytics, we mean the extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive decisions and actions (see the box “Analytics Definitions” for some key terms). The analytics may be input for human decisions or may drive fully automated decisions.

ANALYTICS DEFINITIONS

Analytical competitor: An organization that uses analytics extensively and systematically to outthink and outexecute the competition.

Analytics: The extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive decisions and actions:

• Descriptive analytics (aka business intelligence [BI] or performance reporting) provides access to historical and current data. It provides the ability to alert, explore, and report using both internal and external data from a variety of sources.
• Predictive analytics uses quantitative techniques (e.g., propensity, segmentation, network analysis and econometric forecasting) and technologies (such as models and rule-based systems) that use past data to predict the future.
• Prescriptive analytics uses a variety of quantitative techniques (such as optimization) and technologies (e.g., models, machine learning and recommendation engines) to specify optimal behaviors and actions.
• Autonomous analytics employs artificial intelligence or cognitive technologies (such as machine learning) to create and improve models and learn from data—all without human hypotheses and with substantially less involvement by human analysts.

As figure 1-2 shows, analytics may be descriptive, predictive, prescriptive, or autonomous. Each of these approaches addresses a range of questions about an organization’s business activities. The questions that analytics can answer represent the higher-value and more proactive end of this spectrum.

FIGURE 1-2

---

Potential competitive advantage increases with more sophisticated analytics

---

In principle, analytics could be performed using paper, pencil, and perhaps a slide rule, but any sane person using analytics today would employ a computer and software. The range of analytical software encompasses relatively simple statistical and optimization tools in spreadsheets (Excel being the primary example, of course), traditional statistical software packages (e.g., Minitab or Stata), complex data visualization and descriptive analytics suites (Qlik, Tableau, MicroStrategy, Oracle Hyperion, IBM Cognos), comprehensive descriptive, predictive and prescriptive analytics software (SAS, IBM), predictive industry applications (FICO), and the reporting and analytical modules of major enterprise systems (SAP BusinessObjects and Oracle). Open-source statistical programming capabilities (e.g., R, Python) are rapidly evolving to address both traditional statistical analysis and massive unstructured data. And as we’ll describe later in the book, good analytical capabilities also require good information management capabilities to acquire, transform, manage, analyze, and act on both external and internal data. Some people, then, would simply equate analytics with analytical information technology. But this would be a huge mistake—as we’ll argue throughout this book, it’s the human and organizational aspects of analytical competition that are truly differentiating.

Why Compete on Analytics?

At a time when companies in many industries offer similar products and use comparable technology, high-performance business processes are among the last remaining points of differentiation. Many of the previous bases for competition are no longer available. Unique geographical advantage doesn’t matter in global competition, and protective regulation is largely gone. Proprietary technologies are rapidly copied, and breakthrough innovation in products or services seems increasingly difficult to achieve. What’s left as a basis for competition is to execute your business with maximum efficiency and effectiveness, and to make the smartest business decisions possible. And analytical competitors wring every last drop of value from business processes and key decisions. Analytics are even increasingly being embedded into their products and services.

Analytics can support almost any business process. Yet organizations that want to be competitive must have some attribute at which they are better than anyone else in their industry—a distinctive capability.⁵ This usually involves some sort of business process or some type of decision, or perhaps a distinctive product offering. Maybe you strive to make money by being better at identifying profitable and loyal customers than your competition, and charging them the optimal price for your product or service. If so, analytics are probably the answer to being the best at it. Perhaps you sell commodity products and need to have the lowest possible level of inventory while preventing your customer from being unable to find your product on the shelf; if so, analytics are often the key to supply chain optimization. Maybe you have differentiated your products and services by incorporating some unique data and proprietary algorithms. Perhaps you compete in a people-intensive business and are seeking to hire, retain, and promote the best people in the industry. There too, analytics can be the key.

On the other hand, perhaps your operational business processes aren’t much different from anybody else’s, but you feel you compete on making the best decisions. Maybe you can choose the best locations for your stores—if so, you’re probably doing it analytically. You may build scale through mergers and acquisitions, and select only the best candidates for such combinations. Most don’t work out well, according to widely publicized research, but yours do. If so, you’re probably not making those decisions primarily on intuition. Good decisions usually have systematically assembled data and analysis behind them.

Analytical competitors, then, are organizations that have selected one or a few distinctive capabilities on which to base their strategies, and then have applied extensive data, statistical and quantitative analysis, and fact-based decision making to support the selected capabilities. Analytics themselves don’t constitute a strategy, but using them to optimize a distinctive business capability certainly constitutes a strategy. Whatever the capabilities emphasized in a strategy, analytics can propel them to a higher level. Capital One, for example, calls its approach to analytical competition “information-based strategy.” Caesars’ distinctive capabilities are customer loyalty and service, and it has certainly optimized them with its analytically driven strategy. GE is differentiating its industrial services processes by using sensor data to identify problems and maintenance needs before they cause unscheduled downtime.

Can any organization in any industry successfully compete on analytics? This is an interesting question that we’ve debated between ourselves. On the one hand, virtually any business would seem to have the potential for analytical competition. The cement business, for example, would seem to be as prosaic and non-analytical an industry as one could find. But the global cement giant CEMEX has successfully applied analytics to its distinctive capability of optimized supply chains and delivery times. We once believed that the fashion business might never be analytical, but then we found numerous examples of analytics-based predictions about what clothing styles and colors might sell out this season.

On the other hand, some industries are clearly more amenable to analytics than others. If your business generates lots of transaction data—such as in financial services, travel and transportation, or gaming—competing on analytics is a natural strategy (though many firms still don’t do it). Similarly, if you can draw on the wealth of data available on the internet or on social media to get a unique insight into your customers and markets, competing on analytics is a great way to differentiate yourself. If your business model is based on hard-to-measure factors like style (as in the...

Table of contents

• Cover
• Title
• Copyright
• Contents
• Foreword
• Introduction: Four Eras in Ten Years
• Part 1 The Nature of Analytical Competition
• Part 2 Building an Analytical Capability
• Notes
• Index
• Acknowledgments
• About the Authors