Analytics May/June 2014 by Analytics Magazine

H T T P : / / W W W. A N A LY T I C S - M A G A Z I N E . O R G

DRIVING BETTER BUSINESS DECISIONS

M AY / J UNE 2014 BROUGHT TO YOU BY:

SPORTS ANALYTICS WHAT’S A NICE DEFENSE CONSULTANCY DOING IN THE SPORTS SPACE?

ALSO INSIDE: • How to measure anything • The big ‘V’ of big data • Powerful decision-making Executive Edge Verisk Digital Services President Henna Karna on ‘going digital’

INS IDE STO RY

Quantified warriors What’s a nice defense consultancy company such as the Perduco Group doing in the sports analytics space? That’s the question I asked Stephen Chambal, co-founder of Perduco, after attending his session on “opportunities, barriers and lessons learned” in sports analytics at the recent INFORMS Conference on Business Analytics & Operations Research in Boston. As Chambal notes in his article on the same topic in this issue of Analytics, there’s a “simple” answer (sports are fun) and a “real” answer (his company’s core capabilities and the business opportunities the sports industry presents, combined with a couple of chance encounters, triggered Perduco’s strategic push into the sports arena). As it turns out, the defense community and the sports community are not that far apart in terms of their ultimate goals. They’re both interested in prevailing on the “battlefield,” whether it’s a desert in the Middle East or a basketball court in Madison Square Garden, and they’re both interested in the so-called “quantified warrior” – the ability to monitor and assess a soldier’s/professional athlete’s condition and to understand how to optimize their performance on their 2

A N A LY T I C S - M A G A Z I N E . O R G

respective battlefields. It all makes for a fascinating story, but before I give away too much of it, click here. Chambal’s session was just one of many sessions I attended at the Boston conference. From Tom Davenport’s opening keynote talk on “Analytics 3.0” (look for an article on that topic in a future issue of Analytics magazine, but in the meantime, click here to see his thoughts on predictive analytics in this issue) to the Oscar-esque Edelman Gala, from the four dozen poster presentations to the non-stop series of networking events, the conference was first rate. From my perspective as the editor of Analytics as well as OR/MS Today (the membership magazine of INFORMS), there’s nothing more energizing than attending a conference such as the Boston event, and I suspect that holds true for anyone involved in the analytics community. If you couldn’t make it to Boston or even if you did and are craving another analytics fix, fear not. INFORMS will present its inaugural Big Data Conference on June 22-24 in San Jose, Calif.

– PETER HORNER, EDITOR peter.horner@ mail.informs.org W W W. I N F O R M S . O R G

OPTIMIZE YOUR BUSINESS WITH UNPRECEDENTED SPEED IDEA

IN A FEW HOURS

MISSION CRITICAL ENTERPRISE APP

IN A FEW MONTHS

PUBLISHED INSTANTLY TO YOUR ENTERPRISE OPTIMIZATION APP STORE

PROOF OF CONCEPT

IN A FEW DAYS

OPTIMIZATION APP

IN A FEW WEEKS

To learn more about AIMMS Optimization Apps, visit aimms.com. info@aimms.com | +1 425 458 4024

C O N T E N T S

DRIVING BETTER BUSINESS DECISIONS

MAY/JUNE 2014 Brought to you by

FEATURES 28

‘HOW TO MEASURE ANYTHING’ By Douglas W. Hubbard and Douglas A. Samuelson Latest edition of book takes another look at seven arguments, new and old, in search of the value of business ‘intangibles.’

POWERFUL BUSINESS DECISION-MAKING By Alex Romanenko and Alex Artamonov Big data is a hot topic, but harnessing its full potential can be elusive. The case for an analytics-driven business transformation.

THE BIG ‘V’ OF BIG DATA By Pramod Singh, Ritin Mathur, Arindam Mondal and Shinjini Bhattacharya The three keys – information infrastructure, information management and insights – to unlocking the hidden “value” of big data.

ADVENTURES IN CONSULTING By Stephen Chambal What’s a defense consulting company doing in sports? Core capabilities, chance encounter score a business opportunity.

44 4

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

N Di ew m AS en P sio V2 na 01 lM 4w in excel ode ith lin g

AnAlytic Solver PlAtform easy to Use Predictive and Prescriptive Analytics

How can you get results quickly for business decisions, without a huge budget for “enterprise analytics” software, and months of learning time? Here’s how: Analytic Solver Platform does it all in Microsoft Excel, accessing data from PowerPivot and SQL databases. Sophisticated Data Mining and Predictive Analytics Go far beyond other statistics and forecasting add-ins for Excel. Use classical multiple regression, exponential smoothing, and ARIMA models, then go further with regression trees, k-nearest neighbors, and neural networks for prediction, discriminant analysis, logistic regression, k-nearest neighbors, classification trees, naïve Bayes and neural nets for classification, and association rules for affinity (“market basket”) analysis. Use principal components, k-means clustering, and hierarchical clustering to simplify and cluster your data.

Help and Support to Get You Started Analytic Solver Platform can help you learn while getting results in business analytics, with its Guided Mode and Constraint Wizard for optimization, and Distribution Wizard for simulation. You’ll benefit from User Guides, Help, 30 datasets, 90 sample models, and new textbooks supporting Analytic Solver Platform. Analytic Solver Platform goes further than any other software with Active Support that alerts us when you’re having a problem, and brings live assistance to you right where you need it – inside Microsoft Excel. Find Out More, Download Your Free Trial Now Visit www.solver.com to learn more, register and download a free trial – or email or call us today.

Simulation, Optimization and Prescriptive Analytics Analytic Solver Platform also includes decision trees, Monte Carlo simulation, and powerful conventional and stochastic optimization for prescriptive analytics.

Tel 775 831 0300 • Fax 775 831 0314 • info@solver.com

DRIVING BETTER BUSINESS DECISIONS

DEPARTMENTS 2 Inside Story 8 Executive Edge 12 Analyze This! 16 Healthcare Analytics 20 Forum 24 INFORMS Honors 62 Predictive Analytics 66 Conference Preview 72 Five-Minute Analyst 78 Thinking Analytically Analytics (ISSN 1938-1697) is published six times a year by the Institute for Operations Research and the Management Sciences (INFORMS), the largest membership society in the word dedicated to the analytics profession. For a free subscription, register at http://analytics.informs.org. Address other correspondence to the editor, Peter Horner, peter.horner@mail.informs.org. The opinions expressed in Analytics are those of the authors, and do not necessarily reflect the opinions of INFORMS, its officers, Lionheart Publishing Inc. or the editorial staff of Analytics. Analytics copyright ©2014 by the Institute for Operations Research and the Management Sciences. All rights reserved.

INFORMS BOARD OF DIRECTORS President Stephen M. Robinson, University of Wisconsin-Madison President-Elect L. Robin Keller, University of California, Irvine Past President Anne G. Robinson, Verizon Wireless Secretary Brian Denton, University of Michigan Treasurer Nicholas G. Hall, Ohio State University Vice President-Meetings William “Bill” Klimack, Chevron Vice President-Publications Eric Johnson, Dartmouth College Vice President Sections and Societies Paul Messinger, CAP, University of Alberta Vice President Information Technology Bjarni Kristjansson, Maximal Software Vice President-Practice Activities Jonathan Owen, General Motors Vice President-International Activities Grace Lin, Institute for Information Industry Vice President-Membership and Professional Recognition Ozlem Ergun, Georgia Tech Vice President-Education Joel Sokol, Georgia Tech Vice President-Marketing, Communications and Outreach E. Andrew “Andy” Boyd, University of Houston Vice President-Chapters/Fora David Hunt, Oliver Wyman

INFORMS OFFICES www.informs.org • Tel: 1-800-4INFORMS Executive Director Melissa Moore Meetings Director Laura Payne Marketing Director Gary Bennett Communications Director Barry List Headquarters INFORMS (Maryland) 5521 Research Park Drive, Suite 200 Catonsville, MD 21228 Tel.: 443.757.3500 E-mail: informs@informs.org

ANALYTICS EDITORIAL AND ADVERTISING Lionheart Publishing Inc., 506 Roswell Street, Suite 220, Marietta, GA 30060 USA Tel.: 770.431.0867 • Fax: 770.432.6969 President & Advertising Sales John Llewellyn john.llewellyn@mail.informs.org Tel.: 770.431.0867, ext. 209 Editor Peter R. Horner peter.horner@mail.informs.org Tel.: 770.587.3172 Assistant Editor Donna Brooks donna.brooks@mail.informs.org Art Director Jim McDonald jim.mcdonald@mail.informs.org Tel.: 770.431.0867, ext. 223 Advertising Sales Sharon Baker sharon.baker@mail.informs.org Tel.: 813.852.9942

N Di ew m AS en P sio V2 na 01 lM 4w od ith excel eli ng

AnAlytic Solver PlAtform from Solver to full-Power Business Analytics in

The Excel Solver’s Big Brother Has Everything You Need for Predictive and Prescriptive Analytics From the developers of the Excel Solver, Analytic Solver Platform makes the world’s best optimization software accessible in Excel. Solve your existing models faster, scale up to large size, and solve new kinds of problems. From Linear Programming to Stochastic Optimization Fast linear, quadratic and mixed-integer programming is just the starting point in Analytic Solver Platform. Conic, nonlinear, non-smooth and global optimization are just the next step. Easily incorporate uncertainty and solve with simulation optimization, stochastic programming, and robust optimization – all at your fingertips.

Comprehensive Forecasting and Data Mining Analytic Solver Platform samples data from Excel, PowerPivot, and SQL databases for forecasting and data mining, from time series methods to classification and regression trees, neural networks and association rules. And you can use visual data exploration, cluster analysis and mining on your Monte Carlo simulation results. Find Out More, Download Your Free Trial Now Analytic Solver Platform comes with Wizards, Help, User Guides, 90 examples, and unique Active Support that brings live assistance to you right inside Microsoft Excel. Visit www.solver.com to learn more, register and download a free trial – or email or call us today.

Ultra-Fast Monte Carlo Simulation and Decision Trees Analytic Solver Platform is also a full-power tool for Monte Carlo simulation and decision analysis, with a Distribution Wizard, 50 distributions, 30 statistics and risk measures, and a wide array of charts and graphs.

Tel 775 831 0300 • Fax 775 831 0314 • info@solver.com

EXE CU TIVE E D G E

Digital value: Greater than the sum of its parts Going digital, in simple terms, means transitioning from an “inside-out” to an “outside-in” approach to business.

BY HENNA A. KARNA

Over the past decade, corporations across industry verticals have invested significant capital to upgrade their infrastructure to analyze customer data collected within a multichannel environment. And so the obvious next question is – what now? The answer can be derived from the combination of capabilities, created by the investment in these newer technologies – specifically, the digital engagement of the customer, where customer represents both external consumers and internal stakeholders. Whether to create a network of external customers to drive the creation of a dynamically optimized yet evolving set of products or to drive efficiency in internal operations, digital presents corporations with both challenges and opportunities. Going digital, in simple terms, means transitioning from an “inside-out” to an “outside-in” approach to business. With the exception of a handful of super-enterprises, most corporations are focused on “What can I offer to customers?” and “What can I deliver to customers?” That focus is important but traditional “inside-out” thinking. On the other hand, corporations with an “outside-in” orientation ask questions such as: “What do customers consciously – and subconsciously – want?” “How can I affirm, if not expand, my control over the value chain with new ideas and products?” “How do I redefine business processes with an operating model that informs key answers to the aforementioned questions?”

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

New Solver for Office 365 Excel: Free for Your Tablet or Phone New

With our Solver App for Office 365, SharePoint 2013, and Excel 2013, you have all the capabilities of the Excel Solver on your desktop, laptop, tablet or phone. It works in the Excel Web App with all popular Web browsers, and solves your model “in the cloud” on Windows Azure. And it’s free and available now! Just visit solver.com/app or the Office Store online, or use Insert Apps for Office in Excel 2013.

Analytic Solver Platform: Multidimensional Models with PivotTables New

Analytic Solver Platform 2014 brings multi-dimensional optimization modeling to Excel. Easily create dimensions or index sets and cubes of computed values, using regular Excel formulas – extended to operate over multiple dimensions. Use PivotTables, created in Excel or from databases with PowerPivot, to populate your model with data. Easily create new PivotTables of optimization results.

Find Out More, Download Your Free Trial Now. Visit www.solver.com to learn more, register and download a free trial – or email or call us today. Frontline Solvers – The Leader in Spreadsheet Analytics

Tel 775 831 0300 • Fax 775 831 0314 • info@solver.com

EXE CU TIVE E D G E

digital encompasses both external expanTHE DIGITAL CUSTOMER Technological advances have created sion that is customer-centric and internal the digital customer – one who is empow- optimization that improves efficiency withered by technology, distracted with technol- in the organization. Linking the two is imogy and social through technology. As we portant for a seamless, effective operating move forward, communication with custom- model. ers will be continual and bidirectional, capturing customers’ sentiments and built upon THE FOUR QUADRANTS OF DIGITAL self-service as well as co-creation of prodAt the highest level, the impact of digital ucts and services. Supported by an efficient can be segregated into four quadrants, deoperating model, digitally enhanced tech- fined by interaction vectors (unidirectional nology effectively enables value creation for or bidirectional) and the strategic input vecthe business. tors (outside-in or inside-out). Bidirectional At the highest level, a company would interaction refers to a duplex interaction have to engage its audiences more frequently, more personally and across multiple channels. Digital engagement should mimic the personalization of faceto-face interactions, offering a collaborative environment in which customers can participate in idea generation and product development, and the organization is better able to understand and predict customers’ wants and needs. But keep in mind, Figure 1: The four quadrants of digital. a more holistic view of 10

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

model between the customer and the corporation, effectively providing a means by which true information exchange, not just information transfer, can occur. Outsidein refers to the ability to extract input from outside the corporation to drive decisionmaking and broader strategic direction. Inside-out, on the other hand, refers to strategic decision-making based on a model driven by perspectives within the corporation (primarily a supply-constrained view). As described in the four-quadrant view shown in Figure 1, both outside-in and inside-out input models are relevant consideration factors depending on the objectives (external expansion or internal optimization) of the corporation pursuing a digital transformation. The highlighted examples (customer touch points, captive conversion, internal collaboration, performance management) provide high-level examples of key opportunities for corporations looking to capitalize on digital. WHERE TO BEGIN On a practical level, executing a digital engagement strategy should be gradual but still involve all key facets of the companyâ&#x20AC;&#x2122;s infrastructure. Although a digital platform is heavily supported by technology and is rooted in IT, it should be a priority for the entire business. Paramount for digital success is a broad top-down mandate that spans A NA L Y T I C S

marketing, sales, service, operations, finance and IT. In terms of implementation, inertial elements can slow the internal adoption of digital with business-as-usual perspectives within the company, and externally through an untrained customer base. To manage those constraints, companies should adopt and institute a gradual but focused program, concentrating initially on the interactions most intuitively handled through a digital interface. Such early interactions are best exploited in a service and support (e-service) model that provides a natural problem-resolution incentive for customers to engage. As a result of a digital-service-first model, corporations can obtain insightful data from a large set of interaction types to help identify those easiest to migrate to digital. Payment of bills is one the most common interaction types for this purpose. The challenges of the present, however, donâ&#x20AC;&#x2122;t change the fact that digital engagement is the blueprint for future corporate operating models. To that end, organizations that embrace, implement, and refine their thinking around this imminent paradigm will be the likely standouts in the capital markets in the years ahead. Dr. Henna A. Karna is president of Verisk Digital Services, the digital business unit of Verisk Analytics.

M A Y / J U N E 2 014

ANALY ZE T H I S !

Grad school desires vs. real-world demands The business world is increasingly data-rich, but one must have the ability to sort that data out before any of this analysis can take place. This means getting comfortable with programming and data preparation.

BY VIJAY MEHROTRA

I was recently promoted from “associate professor” to “professor.” There were two notable things about this promotion: • This is the first time in my life that I had ever received a formal promotion from an existing employer (all other job title changes have been the result of changing employers). • I am now officially no longer a “junior” faculty member, a fact confirmed by the impending arrival of my 50th birthday later this year.” Anyway, a few weeks after being notified of my promotion, I flew off to Minnesota to visit St. Olaf College, my undergraduate alma mater. While my primary purpose in going was to attend a meeting of the school’s Alumni Board, the real highlight was the chance to meet with current St. Olaf students, to have a chance to see the world through their eyes, and to offer up any relevant wisdom gleaned from my own experience. The first St. Olaf student I met with was a senior math major, a bright young man who had until recently been planning to pursue a Ph.D. in economics. He was now set on further studies in operations research, and he was trying to decide where to go to grad school. I winced. The second student was a young woman in her junior year who was majoring in mathematics and biology with a concentration in statistics. She had seized the opportunity to visit with me, in large part because

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

she had just begun to explore graduate programs in O.R. I tried to talk her out of it. Bear in mind that some years ago, I had been in the same place that these current St. Olaf students now were. Blessed with a lot of good choices, I had chosen to go forth to study operations research. In fact, so too did my college classmates Hai Chu and Karen Donohue, and I am grateful to have had the chance to bask in their (reflected) professional success in operations research. So why would I advise these youngsters not to follow in our glorious footsteps? Let’s start with some important specifics. First of all, for both of these students, the decision to go to graduate school seemed to serve many purposes: an opportunity to challenge themselves; a chance to improve prospects for both financially and intellectually rewarding careers; and a socially acceptable path with parents, peers and professors. Also, from our conversations, it appeared that both had been initially attracted to operations research by my friend Steve McKelvey, a professor who has been inspiring Olaf math majors since my own student days. Finally, it was quickly apparent to me that the primary motivation was the chance to meaningfully apply their (current and future) skills, rather than any particular passion for O.R. itself. A NA L Y T I C S

Graduate programs in operations research certainly have many virtues, and I will always be deeply indebted to the one that took me in. There will always be some for whom this is a clear and obvious right next step, students who are passionate about the methods and hungry to learn more about them. Yet for the generally quantitatively strong undergraduate who is interested in applying her technical skills within the business world, my postcollegiate recommendations are based on a few simple premises: 1. The business world is increasingly data-rich, but one must have the ability to sort that data out before any of this analysis can take place. This means getting comfortable with programming and data preparation, which we know typically takes up more than 50 percent of the time on most “real-world” projects. 2. Optimization is great, but really good answers quickly are actually better, especially if the environment is rapidly changing or the objective itself lacks a well-defined functional form. 3. You often can’t optimize a system without first predicting future demand, and that forecasting is itself a significant challenge. 4. You are unlikely to do any of this great work totally on your own, so developing the skills needed to work effectively with others is too important to be left to chance. M A Y / J U N E 2 014

ANALY ZE T H I S !

You are unlikely to do any of this great work totally on your own, so developing the skills needed to work effectively with others is too important to be left to chance.

With all this as background, I suggested that the students pursue one of the following: • An M.S. degree in analytics that focuses on preparing students for effectively working across the analytics project cycle, which in addition to content found in traditional O.R. programs also includes training in problem discovery and framing, data capture, preparation and analysis, predictive modeling, business communication, teamwork and project skills. • A Ph.D. program in an academic area of interest (economics, biology, physics and psychology were all discussed), which would include additional rigorous technical training, require them to combine the experience of acquiring deep domain knowledge with challenges in data acquisition and analysis, and help them to more deeply develop their ability to learn independently while working in collaboration with an advisor (and ideally as part of a research group). The second St. Olaf student I met, who is still more than a year from graduating, agreed to give our conversation some serious thought. I was pleased that she had at least listened to what I had to say. However, for the first student, just weeks from commencement, the die is cast. He’s going to Cornell to study O.R. Kids today! What can you do? Vijay Mehrotra (vmehrotra@usfca.edu) is a professor in the Department of Business Analytics and Information Systems at the University of San Francisco’s School of Management. He is also a long-time member of INFORMS.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Make optimized decisions. Even if your data is incomplete. Robust optimization at both the solver and modeling level. Now part of FICO® Xpress Optimization Suite. Missing data and the challenges of harnessing Big Data have introduced a lot of uncertainty into the process of solving complex optimization problems. FICO has solved that problem. We’ve enhanced FICO® Xpress Optimization Suite with features to handle the difficulty of uncertainty introduced by predictive analytics data. This robust optimization guarantees feasible solutions in the face of unknowns. Still not feeling certain? Visit FICO at the 2014 INFORMS Conference: The Business of Big Data, June 22-24 in San Jose, CA and learn how robust optimization handles your missing data challenges.

HEALT H CARE A NA LY T I C S

Rise of the empowered patient consumer – courtesy of analytics About 44 million people in the United States have no health insurance and 38 million have inadequate insurance. For a lot of people there is no “shared responsibility” for health per se. Most people don’t “own” the care of their health; it was provided and mostly paid for by someone else.

BY RAJIB GHOSH

As of this writing, 7.5 million people have signed up for their own health insurance policies via Healthcare.gov or 14 state-run health insurance exchanges. In addition about three million people have enrolled in state Medicaid programs. Private enrollment outside of those insurance marketplaces is also growing and could be substantial. In other words – all signs indicate that more and more people are apparently taking control of their own health – the holy grail of consumer-driven healthcare. Clearly, if we seek control over our cost of insurance, we have to be careful about our personal health choices. Shouldn’t that be the case anyway? Well, that’s not how we do things in the United States. Based on a report by the Congressional Budget Office, more than 50 percent of Americans or 156 million people were covered by employer-sponsored insurance plans in 2013 [1]. Government-run programs such as Medicare and Medicaid cover 31 percent of the population who are poor, disabled or over 65 years old [2]. The government safety net is one of the most prized albeit expensive possessions of the American public.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Despite the fact that more people are taking money out of that safety net than are putting in and a threat that the Medicare fund will be depleted by 2037 – the American public is unwilling to do anything drastic about the safety net. Still, that leaves quite a large number of people who either have to buy insurance on their own or remain uninsured. One estimate shows that about 44 million people in the United States have no health insurance and 38 million have inadequate insurance. While those numbers are huge for a developed nation, for a lot of people there is no “shared responsibility” for health per se. Most people don’t “own” the care of their health; it was provided and mostly paid for by someone else.

Despite the fact that more people are taking money out of that safety net than are putting in and a threat that the Medicare fund will be depleted by 2037, the American public is unwilling to do anything drastic about the safety net.

CHANGING WORKFORCE, CHANGING INSURANCE COVERAGE All of that is changing. Some of it started to change when the availability of employer-sponsored healthcare coverage started to decline a decade ago. According to a 2010 report, the number of people with employer-sponsored health insurance was down 10.6 percent from what it was in 2000. By 2013, the decline was even greater as the recession, job losses and rising costs that forced some small employers to ditch employee group insurance altogether were all contributing factors. Meanwhile, the American workforce has been changing, too. For example, 20 percent to 30 percent of workers in Fortune 100 organizations today are freelancers or “contingent A NA L Y T I C S

M A Y / J U N E 2 014

HEALT H CARE A NA LY T I C S

Providers are adopting information digitization and healthcare analytics, mostly in the form of descriptive business intelligence tools that make fancy post-mortem charts. Predictive analytics is still far-fetched.

workers.â&#x20AC;? By 2020, the number is expected to rise to 50 percent [3], and the number of people covered under employer-sponsored health insurance will become a smaller percentage of the overall population. More people will have to pay for insurance on their own â&#x20AC;&#x201C; from the exchange marketplaces or otherwise. For many who will not have to pay their total insurance bill, the cost sharing will be higher or the coverage will be less or even inadequate. Those that can afford it might have to supplement their insurance with personal policies. WHERE IS THE ANALYTICS? As part of the Affordable Care Act widely known as Obamacare, the U.S. government is trying to drive performance efficiencies and improved quality of care through providers and in the delivery system using programs such as value-based purchasing, readmission penalties and meaningful use of electronic health record systems. In response to these initiatives, providers are adopting information digitization and healthcare analytics, mostly in the form of descriptive business intelligence tools that make fancy post-mortem charts. Predictive analytics is still far-fetched. Health Leaders Media recently identified the top three strategic drivers for providers in 2014, which includes clinical decision support and clinical performance tracking. Both require heavy use of analytics. Payers are taking on more risks to increase their medical-loss-ratio using analytics that can identify patient cohorts with higher risk

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

exposures. But where is the consumer in this change? Are we not supposed to “drive” better, more efficient healthcare for us? Today, that drive is limited to asking for provider cost transparencies or insurance plan shopping. Apart from the quantified selfers, most of us are happy with our annual physical check ups that cost our health system $8 billion a year according to a 2012 study analysis [4], but that does nothing to address serious and expensive illnesses or premature mortality. Few, if any, individuals use analytic insights to proactively know what our current or future risk exposure is or what behavior we should abandon to prevent higher downstream medical costs. Needless to say, we are not sophisticated enough to analyze our now forbidden (by FDA) 23andMe genetic test report to know our genetic predisposition toward future medical expenses. A FUTURE OF EMPOWERED CONSUMER PATIENTS For the latter, we are yet to have predictive analytics, which can take our individual physiological measures and a myriad of other factors and inform us what we need to do to avoid out-ofpocket medical costs three to five years downstream that won’t be covered by our insurance plan. This brings up an A NA L Y T I C S

interesting idea of fusing our health insurance information with our physiological data and genetics – not under the watchful eyes of insurance providers but for us and only for us. Think of it as our personal financial risk dashboard using the powers of predictive analytics! It will be even better if we are able to tweak our behavior data and then see the impact in our risk dashboard. That will be real empowerment for us as consumers. Rajib Ghosh (rghosh@hotmail.com) is an independent consultant and business advisor with 20 years of technology experience in various industry verticals where he had senior level management roles in software engineering, program management, product management and business and strategy development. Ghosh spent a decade in the U.S. healthcare industry as part of a global ecosystem of medical device manufacturers, medical software companies and telehealth and telemedicine solution providers. He’s held senior positions at Hill-Rom, Solta Medical and Bosch Healthcare. His recent work interest includes public health and the field of IT-enabled sustainable healthcare delivery in the United States as well as emerging nations. Follow Ghosh on twitter @ghosh_r. NOTES & REFERENCES 1. Avik Roy, 2010, “Obama Officials In 2010: 93 Million Americans Will Be Unable To Keep Their Health Plans Under Obamacare,” Forbes. 2. “Income, Poverty and Health Insurance Coverage in the United States,” 2011, government publication. 3. Thomas Fisher, 2012, “The Contingent Workforce and Public Decision Making.” 4. Sharon Begley, 2013, “Think preventive medicine will save money? Think again,” Reuters.

M A Y / J U N E 2 014

FO RUM

Analyzing analysts: dreamer vs. pragmatist Analysts and EPM project managers in all industries face a common struggle: acceptance of their ideas, methods and findings by often suspicious work colleagues and managers, some of whom exhibit substantial resistance to change.

BY GARY COKINS

Do you have two imaginary voices that are on each of your shoulders telling you opposite messages? I do. And the voices are conflicting. One is a message of positive hope and possibilities, and the other one is of negative discouragement. The topic and context for each message involves the frustratingly slow adoption rate for applying analytics and progressive enterprise performance management (EPM) methods. Examples of EPM methods are the balanced scorecard with key performance indicators (KPIs), channel and customer profitability analysis, driver-based rolling financial forecasts and lean management techniques. I much better enjoy the inspiring messenger compared to the naysayer one. Who wouldn’t? But I have two ears, so I must listen to both voices. The negative voice is the clear-eyed pragmatist. The positive voice is the creative wild-eyed dreamer. Using a jail prisoner analogy, the pragmatist sees the prison window bars as barriers while the dreamer sees the stars in the night sky. What I am writing about is the struggle that analysts and EPM project managers in all industries have. It is with the acceptance of their ideas, methods and findings by often suspicious work colleagues and managers, some of whom exhibit substantial resistance to change. (You know the type. Their motto is, “We don’t do it that way here.”)

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

BEHAVIORAL CHANGE MANAGEMENT Organizations seem hesitant to adopt analytics and EPM methods. Is it evaluation paralysis or brain freeze? Most organizations make the mistake believing that applying analytics and EPM methods are 90 percent math and 10 percent organizational change management with employee behavior alteration. In reality it is the other way around – it is more likely 5 percent math and 95 percent about people.

A problem with removing behavioral barriers to deploy analytics and EPM methods is that almost none of us have training or experience as organizational change management specialists. We are not sociologists or psychologists. However, we are learning to become like them. Our focus should be on the “why to implement” and its motivating effects on organizations rather than the “how to.” The challenge is how to alter people’s attitudes. One way to remove cultural barriers is to acknowledge a problem that all

JUNE 16-19, 2014

CHICAGO www.pawcon.com Delivering on the promise of

big data PRODUCED BY

A NA L Y T I C S

M A Y / J U N E 2 014

FO RUM

organizations suffer from an imbalance with how much emphasis they should place on being smart rather than being healthy. Most organizations overemphasize trying to be smart by hiring MBAs and management consultants with a quest to achieve a run-it-by-thenumbers management style. These types of organizations miss the relevance of how important is to also be healthy – assuring that employee morale is high and employee turnover is low. To be healthy they also need to assure that managers and employees are deeply involved in understanding the leadership team’s strategic intent and direction setting. Healthy behavior improves the likelihood of employee buyin and commitment. Analytics and EPM methods are much more than numbers, dials, pulleys and levers. People matter . . . a lot. When organizations embark upon applying or expanding its use of analytics and EPM methods, I believe they need two plans: (1) an implementation plan, and (2) a communication plan. The second plan is arguably much more important than the first. WHAT DEFINES SUCCESS? Overcoming barriers is no small task. Cultural and behavioral obstacles not only include that resistance to 22

A N A LY T I C S - M A G A Z I N E . O R G

change but also include some co-workers who fear the consequences of knowing the truth. But driving social change in others is achievable. It requires motivation to want to make a difference in an organization’s performance and to provide others with insights and foresight for them to make better decisions. Overcoming barriers also requires influencing others to be more open-minded. After all, one can never choose an alternative that has not even been considered. Although I am an optimist by nature, I am also a perpetual worrier. I cannot shut out the pessimistic voice on my one shoulder constantly warning me about what can go wrong. Maybe that is OK because that voice forces me to think about contingency plans to cope with unplanned and unexpected events and outcomes. I believe we all need both voices. Just do not shut down the positive voice. It is a corny line, but where there is a will there is a way. Try not to only see the prison window bars but also the stars. Gary Cokins (gcokins@garycokins.com) is the founder of Analytics-Based Performance Management LLC, an advisory firm. A member of INFORMS, he is an internationally recognized expert, speaker and author in advanced cost management and performance improvement systems. He was previously a principal consultant with SAS. For more of Cokins’ unique look at the world, visit his website at www.garycokins.com. A version of this article appeared in Information Management.

W W W. I N F O R M S . O R G

INFORMSCONFERENCE

BIG DATA

THE

BUSINESS

June 22-24

2014

San Jose, California

Tutorials, Case Studies Spanning These Topics • Big data 101: how to navigate the big data ecosystem • Lessons-learned on real-world implementations • Building and managing data science teams • From scoping the problem to advanced analytics, visualization and supporting the decision process • Identifying, storing, searching, cleaning the data you have

Keynote Speaker Bill Franks Chief Analytics Officer Teradata Corporation Putting Big Data to Work

• Gaining insight from new data sources • Selecting the right big data technology • Ethics and privacy requirements • Emerging technologies and trends

meetings. informs.org/bigdata2014 Conference Co-Chairs:

Margery H. Connor Diego Klabjan Chevron Corporation Northwestern University

INFO RM S H O N O R S

CDC wins INFORMS Edelman Award T

he U.S. Centers for Disease Control and Prevention (CDC), which collaborated with Kid Risk, Inc. to use analytics and operations research to combat the remaining pockets of polio around the world, won the 2014 Franz Edelman Award for Achievement in Operations Research and the Management Sciences. Following a series of judged presentations, the award was presented at the Edelman Awards Gala held in conjunction with the INFORMS Conference on Analytics & O.R. in Boston. INFORMS is the premier organization for advanced analytics professionals. Dr. Bruce Aylward, World Health Organization, assistant director-general of Polio, Emergencies and Country Collaboration, said, “This work has been fundamental to so much of what’s happened in the polio eradication program over the last few years, 24

A N A LY T I C S - M A G A Z I N E . O R G

and it has helped to support many of our decisions over the last decade and to bring the world much, much closer to one where future generations will never know the terror of this disease.” “Through collaborations with Kid Risk, Inc. and other partners, CDC is helping to identify the best strategies to further polio eradication and achieve the endgame,” added Dr. Mark Pallansch, director of the Division of Viral Diseases in the National Center for Immunization and Respiratory Diseases at CDC. The Franz Edelman competition attests to the contributions of analytics and operations research in the profit and nonprofit sectors. Since its inception in 1972, cumulative dollar benefits from Edelman finalist projects have reached over $213 billion. As a spearheading partner of the Global Polio Eradication Initiative (GPEI), W W W. I N F O R M S . O R G

the CDC annually contributes over $100 million of its budget and significant human resources to polio eradication activities for which it maintains high standards for developing evidence-based policies and expectations of cost-effective use of its resources. In 2001, the CDC launched a collaboration with Kid Risk, Inc. to use a range of operations research and management science tools combined with the best available scientific evidence and field knowledge to develop integrated analytical models for the evaluation of the global risks, benefits, and costs of polio eradication policy choices. The analytical results from the collaboration significantly furthered polio eradication in many ways, including more rapid response to outbreaks and reaffirmation that pursuing eradication instead of control is the “best buy” to prevent cases of paralysis and to save lives and money. Recognition of polio eradication as a major program in need of stable financing helped support a fundraising effort in 2013 that raised over $4 billion from donors to finish the job. The team foresees increased integration of operations research and management science tools to perform simultaneous probabilistic and dynamic modeling for other complex global health challenges, including other vaccine-preventable diseases like measles and rubella. A NA L Y T I C S

Members of the Edelman Award-winning team from the CDC and Kid Risk, Inc.

Along with the CDC, the other finalists competing in the 2014 Franz Edelman Award Competition included teams from Alliance for Paired Donation, The Energy Authority, Grady Health System, Australia’s NBN and Twitter. MAYO CLINIC EARNS INFORMS PRIZE Mayo Clinic, the innovative healthcare organization that has used analytics throughout its organization to provide economical, quality services in an era of ballooning medical costs, was named the 2014 winner of the INFORMS Prize. The prize was presented at an awards gala held in conjunction with the 2014 INFORMS Conference on Business Analytics and Operations Research in Boston. “Operations research is deeply rooted in Mayo Clinic’s culture,” says Mayo Clinic President & CEO John Noseworthy, M.D. M A Y / J U N E 2 014

INFO RM S H O N O R S

Mayo Clinic’s century-long history of using systems thinking, analytics and operations research traces its roots back to Dr. Henry Plummer, who developed the first integrated, paper medical record as a platform to organize and share patient information within a group practice of medicine.

Representatives from the Mayo Clinic accept the INFORMS Prize.

“These disciplines help us improve patient outcomes and experience while controlling rising health care costs, one of the biggest financial challenges facing our country today.” The INFORMS Prize recognizes effective integration of operations research into organizational decision-making. The award is given to an organization that has repeatedly applied the principles of O.R. in pioneering, varied, novel and lasting ways. Mayo Clinic’s century-long history of using systems thinking, analytics and operations research (O.R.) traces its roots back to Dr. Henry Plummer, who developed the first integrated, paper medical record as a platform to organize and share patient information within a group practice of medicine. This served as the foundation of Mayo Clinic’s culture of applying engineering and O.R. principles. Mayo Clinic continues to make significant investments to ensure a sophisticated advanced analytics and O.R. infrastructure. With more than 500 practitioners of O.R. and analytics, Mayo Clinic is able to continually leverage analytical methods to enhance strategic planning, care process redesign, patient experience, inventory management and project management – leading to

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

important patient benefits as well as financial savings. Examples include optimization models for patient scheduling, queuing theory for effectively transporting patient, systems dynamics for strategic capital allocation planning, simulation modeling to redesign pharmacies that reduce patient waiting and discrete event simulation models for blood management. In addition to applying advanced analytics and O.R. to its business, Mayo Clinic has made significant strides to educate analytics_Layout 1 4/25/14 12:51 PM Page 1

staff and disseminate what they’ve learned. Mayo Clinic’s leadership recognizes analytics and engineering as key contributors to the organization’s sustained excellence, market differentiation, and superior customer experience. Looking to the future, senior leaders consider these disciplines to be vital in addressing the formidable challenges in healthcare today and tomorrow. Past recipients of the award include Intel, UPS, HP, IBM, Ford, Procter & Gamble and GE Research.

Stand Out. Put yourself in a lucrative new career. Apply now for a master’s degree in business analytics or supply chain management. • Intensive nine month programs • World-renowned faculty • Experiential projects with industry clients • Personalized professional development

www.leeds.colorado.edu/ms 303-492-8397 leedsms@colorado.edu

A NA L Y T I C S

M A Y / J U N E 2 014

VA L U E OF BU S I N E S S I NTA NG I BLE S

Explaining ‘How to Measure Anything’ Latest edition of book takes another look at seven arguments, new and old.

BY DOUGLAS W. HUBBARD (left) AND DOUGLAS A. SAMUELSON

nalytics professionals and decision-makers are often stymied by the lack of good metrics on which to base decisions. But everything that matters has observable consequences and, with a bit of (often trivial) math, these observations provide the grounds for reducing uncertainty. Even imperfect information has a computable value for decisions. These ideas were summarized in the book “How to Measure Anything: Finding the Value of ‘Intangibles’ in Business” 28

A N A LY T I C S - M A G A Z I N E . O R G

[Hubbard, 2007, 2010, 2014], written by one of the authors of this article. With 65,000 copies of the book sold in five languages, the message seems to strike a chord. The client list of the author’s firm, Hubbard Decision Research (HDR), and the thousands of individuals who have registered on the book’s website indicate a diverse audience. They include engineers, human resources, software developers, information-security specialists, scientists from many fields, managers in many industries, actuaries and teachers. It W W W. I N F O R M S . O R G

appears that the challenge of measuring what – at first – appear to be “intangible” is common for many analysts and managers in organizations of all types. A third edition has just been released, with an accompanying workbook to facilitate classroom teaching and self-study. The third edition also allowed the author to include cases from new clients and to respond to the most common challenges sent in by readers in the seven years since the first edition. As in the two earlier editions, readers learn how to frame the measurement problem and how to avoid measuring the wrong things, and they see the value of relying on their quantitative models over pure intuition. However, even for the most fervent advocates of quantitative methods among our clients and readers, we find that they can easily be bogged down by some of the same obstacles as the skeptics of quantitative methods. Even though we make what seems to us to be a strong argument for the correct way to approach these issues and even though clients say they “conceptually” agree with the argument, they sometimes still seem to repeat, unknowingly, some of the same errors. A NA L Y T I C S

What follows are seven areas where we build on the message of the previous edition by adding new cases, new research and new responses to the challenges we continue to observe among our readers and clients. 1. It’s still true, anything can be measured. We haven’t found a real “immeasurable” yet, although many things initially appear to be. In the past several years, HDR has developed measures of the risk of a mine flooding, drought resilience in the Horn of Africa, the market for new laboratory devices, the risks of cyberattacks and the value of industry standards, to name a few. The other author of this article (Samuelson) measured the asset value of information technology [Samuelson, 2000] and the value of deterrence in security situations [unpublished]. In each of these cases something was perceived to be virtually impossible to measure and, yet, the authors were able to show that we can use informative observations and simple, established mathematical approaches to reduce uncertainty enough to make decisions. As in earlier editions, the book explains the three reasons anything is ever M A Y / J U N E 2 014

ME AS URIN G A NY T H I NG

Almost everyone can be trained to be an expert “probability estimator.”

perceived to be immeasurable, and why all three are mistaken. 2. Do the math. A key point in every edition of the book was that we measure to feed quantitative decision models, and that even naïve quantitative models easily outperform human experts in a variety of estimation and decision problems. In a meta-study of 150 studies comparing expert judgment to statistical models, the models clearly outperformed the experts in 144 of the cases [Meehl, 1975]. More and more research confirms this. The third edition adds the findings of Philip Tetlock’s giant undertaking to track more than 82,000 forecasts of 284 experts over a 20-year period. From this, Tetlock could confidently state, “It is impossible to find any domain in which humans clearly outperformed crude extrapolation algorithms, less still sophisticated statistical ones” [Tetlock, 2006]. The book reviews additional research to show that, unless we do the math, most people, even statistically trained experts, are susceptible to common inference errors. 3. Just about everyone can be trained to assess odds like a pro. Almost everyone can be trained to be an expert “probability estimator.” Building on the work of others in decision psychology [Lichtenstein and B. Fischhoff, 1980], HDR started providing “calibrated probability assessment” training in the mid-1990s. The third edition included data from more than 900 people calibrated by HDR. The data consistently

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

shows that virtually everyone is extremely overconfident before the training (e.g., over a large number of trials, when they say they are 90 percent certain, they may have less than a 60 percent chance of being correct). But HDR also found that about 80 percent of individuals can be trained to be nearly perfectly calibrated (they are right just as often as they expect to be). In other words, they can be trained in about half a day to be as good as a bookie at putting odds on uncertain events. This skill becomes critical in the process of quantifying someone’s current uncertainty about a decision. 4. Calculating information values avoids “the measurement inversion.” A defined decision should always be the objective of measurement. Uncertain variables in such a decision have a computable expected value of information (EVI); that is, what is it worth if we had less uncertainty about this? When HDR compared the EVI to clients’ past measurement habits, virtually always what got measured and what needed to be measured were very different things. With the third edition, HDR has conducted more than 80 major decisions analysis, and the results are consistent with earlier findings: This phenomenon appears to pervade every industry and profession from software development to pharmaceuticals, A NA L Y T I C S

real estate to military logistics, and environmental policy to technology startups. It appears that the intuition managers follow to determine what to measure routinely leads them astray; they tend not to measure the very things for which they have the poorest information and would therefore benefit most from more data. Hubbard calls this practice “the measurement inversion,” and it appears that the best guarantee to avoid this problem is simply to know the information values of uncertainties relevant to a decision. 5. A philosophical dilemma: Does probability describe the object of observation or the observer? When someone says, “but how do I know what the exact probability is?” they are implicitly adopting a particular definition of the word “probability.” Since the author observed the challenges some readers were having with this issue, the newest edition of “How to Measure Anything” expands more on it. We generally take a Bayesian position on the interpretation of probability – that is, probability is used to quantify the uncertainty of an observer, not a state of the thing being observed. This stands in contrast to the “frequentist” point of view, which treats a probability as a kind of idealized frequency of occurrence in some objective system. Somewhat ironically, the validity of applying subjective M A Y / J U N E 2 014

ME AS URIN G A NY T H I NG

probabilities to uncertain outcomes has been tested with frequentist methods. That is, extremely large trials have been conducted where individuals’ probabilities were compared to observed outcomes. As mentioned earlier, the authors and other researchers have verified that people who are trained as “calibrated probability assessors” can repeatedly assign probabilities that, after sufficient trials, align with the observed frequencies. Since probability is your state of uncertainty, and since you can be calibrated, you can always state a probability – in the Bayesian sense. 6. Statistical significance doesn’t mean what you think, and what it does mean you probably don’t need. Another issue the author was observing that was getting in the way of useful measurements was that there was a widely held, but very vague, understanding of the concept of “statistical significance.” The new edition addresses pervasive misunderstandings of this idea and then makes the case that, even when it is understood, it isn’t directly relevant to most real decision-making problems. The book contains examples in which just five sample observations, or in some cases even just one observation, substantially reduce uncertainty. We have had clients who looked at a small sample and – without attempting any math – questioned the statistical significance of 32

A N A LY T I C S - M A G A Z I N E . O R G

the sample and the results. But sample size alone is not sufficient to determine statistical significance. Nor does statistical significance mean the chance that a claim is true, nor that if we fall short of statistical significance we have learned nothing. The newest edition argues that the entire concept is not necessary when even small reductions in uncertainty can have significant economic value. 7. You need less data than you think, and you have more data than you think. A client or reader who says, “I would like to measure this but we just don’t have enough data” is very likely making a series of erroneous assumptions. As in the previous point on statistical significance, managers may seriously underestimate how much uncertainty reduction they get from a small amount of data. In fact, we have never seen anyone who made this claim who had actually calculated the uncertainty reduction from a given set of data and computed the value of it to the decision, to ascertain that the uncertainty reduction had no value. Managers also underestimate how much data they really have. One example of this, discussed in the third edition of the book, is the “uniqueness fallacy.” This is the tendency to believe that only highly similar if not identical examples are informative. The latest edition includes cases where experts insisted that since each situation is W W W. I N F O R M S . O R G

unique, they cannot extrapolate from historical data. Then – and without a hint of irony – they will claim that therefore they must rely on their experience. Of course, as the book argues, expertise and science are both based on past observations – one of these with much more selective recall and tendency for flawed inferences than the other. Managers make just such a mistake whenever they say that they can’t make estimates about implementing a new technology because it is so unique – even though they have a long history of implementing new technologies. Using that same logic, your insurance company couldn’t possibly compute a life insurance premium for you because you’re unique and because you haven’t died yet. In fact, insurance actuaries know how to extrapolate from larger, more heterogeneous populations. The third edition also expands on developments in how big data, social media, mobile phones and personal measurement devices are making the “we don’t have enough data” excuse much harder to justify. SUMMARY You can, in fact, measure anything, in our view, but doing so is sometimes a challenge even for those who are convinced the claim is true. We simply need to recognize that the perceived challenge A NA L Y T I C S

results from some of the same old, entrenched misconceptions. Your problem is most likely not as unusual as you think; there are sources of information you can use, if you think creatively about how to apply them; calibrated experts can make good estimates of their uncertainty about the data points they provide; and calculating the expected value of information can focus you on collecting the most useful additional data, not wasting effort and resources on data that won’t help much. Douglas W. Hubbard (dwhubbard@ hubbardresearch.com) is president of Hubbard Decision Research in Glen Ellyn, Ill., and an internationally recognized expert in measurement and decision analysis. Douglas A. Samuelson (samuelsondoug@yahoo.com), D.Sc., is president and chief scientist of InfoLogix, Inc., a consulting and R&D company in Annandale, Va., and a contributing editor of OR/MS Today and Analytics magazines. He is a longtime member of INFORMS.

NOTES & REFERENCES 1. Douglas W. Hubbard, 2007, “How to Measure Anything: Finding the Value of ‘Intangibles’ in Business,” Wiley; third edition, 2014. 2. S. Lichtenstein and B. Fischhoff, 1980, “Training for Calibration,” Organizational Behavior and Human Performance, Vol. 26, No. 2, pp.149-171. 3. Paul Meehl, 1986, “Causes and Effects of My Disturbing Little Book,” Journal of Personality Assessment, Vol. 50, pp. 370-375. 4. Douglas A. Samuelson, 2001, “Information Technology Benefits Assessment,” Encyclopedia of Operations Research and the Management Sciences, Second Edition, Springer. (A revised version also appears in the third edition, 2013.) 5. Philip E. Tetlock, 2006, “Expert Political Judgment: How Good Is It? How Can We Know?” Princeton, N.J.: Princeton University Press.

M A Y / J U N E 2 014

B IG DATA

Using analytics to make powerful business decisions Big data is a hot topic, but harnessing its full potential can be elusive.

BY ALEX ROMANENKO (left) AND ALEX ARTAMONOV ig data is generating a powerful buzz. For firms that know how to harness it, big data can offer a significant competitive advantage. However, much of the ongoing hype has been focused on gaining insights from these vast amounts of accumulated information while a more intriguing question lingers outside of the spotlight: How can these insights be translated into powerful business decisions?

A N A LY T I C S - M A G A Z I N E . O R G

So far, big data alone has not been developed into anything near its touted value, primarily because there is a disconnect between the vast volume of data and the managers who make and implement business decisions. Analytics can bridge this gap by applying algorithms, generating and presenting recommendations for optimal, practical and achievable business decisions in a user-friendly format. Operations research â&#x20AC;&#x201C; the scientific discipline of using analytical methods to W W W. I N F O R M S . O R G

Source: A.T. Kearney analysis

make better decisions â&#x20AC;&#x201C; encapsulates these analytical techniques, applicable at all stages of a companyâ&#x20AC;&#x2122;s operations, to help improve overall profitability given particular business objectives and constraints. In turn, a variety of user-friendly visualization tools and techniques can help management focus on their most important key performance indicators (KPIs). This article explores some of the most common applications for analytical and visualization techniques and highlights the benefits that can be achieved. THE POWER OF OPTIMAL DECISIONS Analytical techniques allow us to understand and stimulate demand, develop A NA L Y T I C S

an efficient production plan, effectively source and allocate production resources, and lower distribution costs. Across all industries, many companies are excelling at applying these techniques, recognizing them as necessary to maintain a competitive advantage. Analytics can have a sizable impact across all areas of operations (see Figure 1). SALES AND MARKETING Demand forecasting. Being customer-oriented and demand-driven are modern business prerequisites. Although demand sensing and predicting future behavior are crucial activities that directly influence sales, required inventory levels and customer service, M A Y / J U N E 2 014

B U S IN E S S DE C I S I O N -M A K I NG

More accurate forecasting, supported by collaborative silo-penetrating processes, can reduce working capital up to 20 percent and reduce out-of-stock events by up to 6 percent.

many companies still use the wrong tools, including spreadsheets and black-box enterprise resource planning (ERP) algorithms, which are not necessarily fine-tuned for individual SKUs and may be especially ill-suited for slowmoving items with no sales in some periods. Forecasting is also often ignored at the pointof-sale level, which is harder to do but can be used to improve distribution-center forecasts and cross-department collaboration. Choosing right-time series models, tightening modeling parameters using mathematical optimization and adjusting processes to become demand-driven can result in substantial operational improvements. More accurate forecasting, supported by collaborative silo-penetrating processes, can reduce working capital up to 20 percent and reduce out-of-stock events by up to 6 percent. Marketing optimization. Demand can be stimulated by driving up sales with brand-recognition campaigns or by promoting individual goods and services. Sometimes, these promotional campaigns are either too broad or poorly timed and very often offer higher discounts than necessary to achieve extra sales volumes. Marketing optimization approaches maximize the effectiveness of these campaigns within marketing budget constraints. Alternatively, they can inform decision-makers about the right budget level to achieve a certain sales volume. These techniques routinely improve the marketing budget by 10 percent while allowing for achieving the business objectives.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Optimal pricing. Once the desired level of demand is attracted, it can be further managed through optimal pricing. Optimal pricing balances both margin and volume so that transaction profitability is maximized amid business constraints such as production and distribution limitations. Determining the optimal pricing level requires a good understanding and quantification of the underlying demand for goods and services and of the profitability of any given transaction. At the portfolio level, optimal pricing, implemented based on multivariate constraint optimization techniques, allows driving some segments for volume and others for margin growth, aligned with the business strategy. Profitability improvements as a result of applying these techniques can reach 2 percent to 5 percent of revenue. Once demand is understood and the right level to maximize the profitability of sales transactions is attracted, it is time to analyze how demand triggers other activities in the company. In traditional systems, demand only affects decisions at the nearest stock locations and their replenishment through outbound logistics. However, demand signals from the point of sale can also be taken into account on upstream stages of the supply chain to more accurately decide how to allocate stock across the system, which needs to be modeled holistically. A NA L Y T I C S

SUPPLY CHAIN Strategic network. Whether it is after inorganic growth, mergers and acquisitions, moving sourcing to low-cost countries or resourcing transport providers, footprint and flow path restructuring is a crucial activity for supply chain managers. The most common optimization applications are establishing where to get raw materials, what to produce where, how much and where to store it, who to deliver to, and what assets are required across the whole network. However, real life presents interesting modeling and implementation challenges. Examples of these include convincing stakeholders of the need to holistically optimize the end-to-end supply chain, considerations for production scheduling, demand sensing across all network tiers, non-linear costs for warehousing and transport, non-linear relationships between the quality of raw materials and the quality of finished goods, the integration of less quantifiable and known elements such as competition into models, and the double objectives of costs and CO2 emissions. Typical network optimization projects save 5 percent to 10 percent, but higher benefits â&#x20AC;&#x201C; up to 20 percent â&#x20AC;&#x201C; are not unprecedented. Supply chain operations. Supply chains are inherently dynamic because of the uncertainties of customer demand, M A Y / J U N E 2 014

B U S IN E S S DE C I S I O N -M A K I NG

Source: A.T. Kearney analysis

lead times and other unforeseen events. Simulation is the best way to model dynamic operations to improve business policies that balance customer service levels, multi-echelon inventories, and the use of transport and production assets while keeping costs under control. Because company operations is an area where the devil is in the details and multiple trade-offs exist, much effort goes into modeling to capture this complexity on the required level of detail. Benefits include overall service-level improvements and 20 percent to 30 percent inventory reductions. 38

A N A LY T I C S - M A G A Z I N E . O R G

Holistic supply chain. When overhauling the supply chain, strategic and lower-detail operations designs are both beneficial because they leave no stone unturned. Operational design also helps to show the customer how a redesigned network will work in real life. Although there are many ways to make the optimization and simulation methods work together, a model with two feedback loops is ideal (see Figure 2). Optimized flow path design is fed into a dynamic simulation model, which is then tightened with a simulation-optimization loop. If an adjustment is needed to the high-level network W W W. I N F O R M S . O R G

design, the outcome from operational testing is passed back into strategic optimization. Once the structural backbone of the supply chain and product flow is modeled, the next step is to improve operational complexity and distribution. OPERATIONS Production scheduling. In many production processes, setup times are conditional on the sequence in which operations are performed on a single machine. Batch size is therefore a crucial decision because it determines how often changeovers need to be done and determines product availability in the warehouse, which affects dispatch to customers. Scheduling problems become even more complex when the same job can be performed on different machines with varying degrees of efficiency. Using complex combinatorial math optimization with column generation to optimize production schedules, product throughput and asset utilization can be improved by up to 5 percent. Site design. Simulation is again an ideal tool when production, transport or other capabilities need to be designed or redesigned to support operational changes, including increased throughput, alternative job-shop configurations and production scheduling. Flexible and A NA L Y T I C S

detailed models can be built to simulate many scenarios to determine the most suitable configurations, especially if optimization add-ons are used. For example, if production capacity will increase significantly, a simulation can quickly find the most feasible logistics and warehouse layouts to cope with this expansion. Asset management and failure prediction. Modern operations are characterized by a vast and complex assets base. When an asset fails, the operating routine can be significantly disrupted, which can deal a big blow to revenue. Methods that can predict an asset failure and then plan and schedule the corresponding maintenance can lower the risk of failure and improve the return on assets. Approaches that use predictive analytics can lower maintenance and repair costs by 10 percent. Because production and operations decisions lead directly to dispatching goods and services to points of sale and end customers, distribution is the next area to consider. DISTRIBUTION Transport routing. The traveling salesman problem is one of the most notorious optimization tasks when a vehicle needs to visit numerous geographically distributed locations. However, M A Y / J U N E 2 014

B U S IN E S S DE C I S I O N -M A K I NG

Source: A.T. Kearney analysis

optimization can help plan this task. Several constraints need to be taken into account to reflect operational reality â&#x20AC;&#x201C; including terminal handling capacities, demand availabilities and sequence priorities â&#x20AC;&#x201C; to represent operations with terminal networks that use less-than-fulltruckload shipments. Within a few hours, a huge number of alternatives can be evaluated to identify the best schedule, resulting in asset utilization improvements of 2 percent to 5 percent. Transport loading. Transport loading is another interesting application of math optimization that belongs to the 40

A N A LY T I C S - M A G A Z I N E . O R G

family of knapsack problems, where different loading setups are possible. This gets even trickier if every load is unique; for example, if the problem needs to be solved on a daily basis because customer orders change or if loading sequence matters because some items can be transported on top of others but some cannot. Typical project benefits are 2 percent to 4 percent of increased vehicle space usage. Many companiesâ&#x20AC;&#x2122; operations are executed by external service providers, and this is where efficient procurement becomes essential. Procurement cuts W W W. I N F O R M S . O R G

Source: A.T. Kearney analysis

across all business functions and, if executed well, can significantly improve the bottom line. PROCUREMENT Sourcing. For many sourcing events, optimization can help match expressive or non-standard offers with the business requirements. Potential suppliers often come up with volume or package discounts, step-change pricing, alternative offers, capacity constraints or other ways to showcase their strengths. However, these complex offers cannot be taken at face value, and optimization is required to A NA L Y T I C S

assemble the puzzle pieces into a coherent picture that covers business requirements and minimizes purchasing costs. Another benefit of this approach is that a sensitivity analysis can be used to estimate the costs of business constraints and challenge business stakeholders on the ones that are less crucial to business (see Figure 3). For example, typical sourcing events for transport services result in an 8 percent to 12 percent cost reduction. Sourcing with strategic network design. Sourcing optimization conducted jointly with strategic network design can be especially beneficial. In M A Y / J U N E 2 014

B U S IN E S S DE C I S I O N -M A K I NG

this case, network optimization uses true market quotes rather than approximated lane and warehousing costs. In many cases, it uncovers hidden market potential (see Figure 4). The benefits for sourcing transport and warehousing together with simultaneous supply network optimization are often in the range of 10 percent to 15 percent. Component choice. When the production process is flexible, such as if raw materials vary or different formulations can be used to arrive at the same result, optimization can help determine the most cost-efficient way to make products. This is especially useful if costs for raw components are volatile, and different vendors can supply materials of various quality and resulting costs. Optimization projects that explore production flexibility can minimize total costs of goods sold by 1 percent to 3 percent on raw material purchasing. WEAVING ANALYTICS INTO THE FABRIC OF BUSINESS Developing sophisticated models is impractical if business stakeholders donâ&#x20AC;&#x2122;t use them. Gaining their buy-in is vital. To capture incremental business

benefits on a regular basis, analytical solutions must be institutionalized and incorporated into daily decision-making. Visualization is one way to help stakeholders focus on their KPIs by presenting information in a user-friendly format. Every analytics project needs a visualization component that reflects insights, complexities and interdependencies so that the advanced analytical algorithms are not perceived as black boxes (see accompanying sidebar story, pages 43-44). This, in turn, increases trust in the results. Regardless of market conditions, forward-thinking players that use analytics perform better than their competitors. They know that capturing a competitive advantage requires going beyond ERP upgrades. By bridging the gap between decision-makers and the vast volume of data, an analytics-driven business transformation can ensure that optimal decisions are an integral part of every business unit. Alex Romanenko (alex.romanenko@atkearney.com) leads A.T. Kearneyâ&#x20AC;&#x2122;s Analytics Practice in London, which develops and delivers analytics-based solutions in the United Kingdom and around the world. Alex Artamonov (alex.artamonov@atkearney.com), a manager with A.T. Kearney, leads supply chain transformation projects that use analytics to support strategic and operational decision-making.

Request a no-obligation INFORMS Member Benefits Packet For more information, visit: http://www.informs.org/Membership

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Transforming data into graphics to gain buy-in The most sophisticated analytical mod-

An array of visualization tools can

els are meaningless if business stake-

transform advanced algorithms into eas-

holders avoid using them when making

ily understandable graphics. For example,

strategic decisions. Turning data into a

Tableau Softwareâ&#x20AC;&#x2122;s (tableausoftware.com)

more aesthetic, user-friendly format builds

interactive dashboard can present footprint

trust in complex analytical results.

optimization results and compare scenari-

Figure 5-1: Info graphic view of analytics jobs.

A NA L Y T I C S

os to help stakeholders make informed decisions (see Figure 5-1).

M A Y / J U N E 2 014

B U S IN E S S DE C I S I O N -M A K I NG

Figure 5-2

Gephi (gephi.org), an open-source graph visualization and manipulation software program, can be used to highlight complexity and interlinks between supply chain tiers to derive a cost-efficient and sustainable setup (see Figure 5-2).

Circos (circos.ca), which takes data from a table and converts it into a circular layout, can show the links between three types of blending products and their components to come up with the most economical component purchases (see Figure 5-3).

Figure 5-3

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Visual Analytics Opportunity at your fingertips.

The answers you need, the possibilities you seek—they’re all in your data. SAS helps you quickly see through the complexity and find hidden patterns, trends, key relationships and potential outcomes. Then easily share your insights in dynamic, interactive reports.

Try Visual Analytics and see for yourself

sas.com/VAdemo

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. © 2014 SAS Institute Inc. All rights reserved. S120597US.0214

VO L UME, VE LO C I T Y, VA R I E T Y A N D ...

The big ‘V’ of big data Keys to tapping the hidden “value” of big data.

BY (l-r) PRAMOD SINGH, RITIN MATHUR, ARINDAM MONDAL AND SHINJINI BHATTACHARYA “There is a big data revolution. However, what is revolutionary is not the quantity of data alone. The big data revolution is that now we can do something with the data.” – Professor Gary King, Harvard University

hile storage and computational capacity is essential and a given, it is important to note that improved statistical and computational methods are creating opportunities like never before when dealing with big data. Today, large amounts of data are available that individuals, businesses and governments can manage 46

A N A LY T I C S - M A G A Z I N E . O R G

and leverage for information that can lead to insights. However, it is essential to analyze information in a cost-effective manner. High volume, variety and velocity of data cannot translate into value unless management makes a concerted effort. Unlocking this value that businesses can get from big data involves three key elements: W W W. I N F O R M S . O R G

Figure 1: Three key elements to unlocking the value of data. 1. Information infrastructure (ingest and store efficiently): This element is about creating infrastructure that can capture, store, replicate and scale information at speed. 2. Information management: An information ecosystem to manage, secure, govern and leverage information seamlessly across an organizationâ&#x20AC;&#x2122;s information assets. 3. Insights: Correlate and use this data in conjunction with existing business data (usually structured data) and analyze using descriptive and prescriptive analytics to aid decision-making. The last few years have seen a lot of focus and attention on infrastructure and information management. Exciting new technologies, frameworks and A NA L Y T I C S

methodologies have evolved to address the needs of these elements. For example, infrastructure technologies have greatly improved, as have the innovations that benefit data centers. They range from fast and efficient servers to data center solutions that can capture, store, replicate and scale information at high speeds. Information management has seen the most rapid evolution and change. Managing information through distributed technologies (file systems such as Hadoop) has changed information storage to provide low latency, high speed, highly available systems. Innovations in areas such as traditional enterprise data warehousing environments through indatabase, in-memory capabilities and M A Y / J U N E 2 014

B IG DATA

For business analytics to be successful in meeting an organization’s needs for decision support, it fundamentally needs to be able to consider all data sets relevant to solving a particular business question.

compression techniques have resulted in organizations being able to get insights from data quicker. As organizations come to terms with managing big data and harnessing them through systems and converge their usage with traditional data sources, the business analytics to guide decision-making is itself evolving. This article will focus on how analytics has been influenced by big data and what practices will emerge in years to come through observations within Hewlett Packard. UNITED WE ARE “BIG,” DIVIDED WE MAY BE SMALL For business analytics to be successful in meeting an organization’s needs for decision support, it fundamentally needs to be able to consider all data sets relevant to solving a particular business question. Traditional business intelligence (BI) and enterprise data warehouse (EDW) environments focus on the usual data generated from business operations. This is data generated through point of sale transactions, customer data, financial, business planning data, inventory management systems, etc. Businesses today, however, also have access to two other key forms of data. The first of these can be loosely categorized as “human information”; this form of data comes from having increased knowledge of customers through e-mail, social media and other marketing channels, but also from an organization’s institutional data in the form of documents and customer support call records, as well as video, audio and image sources. This data tends to be unstructured in format.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Figure 2: Bringing data together. The second new type of data comes from machine data. This is data generated from an increasingly interconnected world of devices and systems. Examples range from data generated by sensors, smart meters, RFID tags, security and intelligence systems, IT logs (application and Web servers), etc. This data tends to be largely semi-structured or unstructured. Business systems in BI and EDW environments are not architected to handle the volume and variety of â&#x20AC;&#x153;human informationâ&#x20AC;? nor the volume and velocity of machine generated data. Today, organizations need to bring all their data together for advanced analytics. For example, at HP, structured data from a customerâ&#x20AC;&#x2122;s purchase history, demographics and warranty data can be

A NA L Y T I C S

combined with unstructured data coming from customer support records and social media for a more focused customer engagement strategy. BIG DATA AND ANALYTICS: PROCESS, PURPOSE, PRACTICE As information and data assets of an organization come together and combine with external data, analytical techniques and analysis will have a larger role to play. In general, the characteristics of big data that most influence the analytics process are related to the variety and volume of data. However, velocity, which is handled through business intelligence practices, is considered distinct from core analytics practices for the purposes of this article. The analytics process is usually represented as a set of activities

M A Y / J U N E 2 014

B IG DATA

â&#x20AC;&#x201C; preparing data, developing analytical models based on analytical techniques to solve for the business question, validating the model and deploying it. These are essentials of analytics and represent the elements of big data that organizations today need to pay attention to. Following is a closer, more technical look at each of these activities. DATA PREPARATION: SAMPLING. Sampling has been the backbone of analytical processes with the premise of using information of a sample to infer on a larger population. Historically, sampling has been a core part of analytical processes due to limitations on collecting data on populations and then analyzing it in aggregate. Sample accuracy, of course, depends on several factors and is predicated on the minimization of various biases in the sampling methodology. While there are arguments both for and against sampling in big data environments, from a data scientistâ&#x20AC;&#x2122;s perspective, a few aspects of the data used in an analytics process need to be well understood. First, one canâ&#x20AC;&#x2122;t always use large storage and computing power to analyze a population unless the marginal business returns are higher due to the addition of more data sets. Second, some 50

A N A LY T I C S - M A G A Z I N E . O R G

specialized application areas do need population rather than sample. For example, in case of analysis of cyber security threats out of a large data set, our interest lies in finding the outliers and anomalies in the data. Where millions of rows of data may not give any value, a particular row (one individual) may be very useful and could save huge losses for the business. Another example is when one has to identify the five top social media influencers in cloud computing; here we might consider population as an important element. And last, irrespective of whether we choose to use a sample or a population, it is still vitally important for a data scientist to understand and question the data source and collection methods so that a selection bias in data may be averted. As such, the need and relevance of sampling in big data applications is contextual and depends on the question being solved and the source of the data. ANALYTICS TECHNIQUE: REGRESSION. One of the most common analytical techniques used by analysts today are related to regression. Regression is commonly understood as a statistical process for estimating relationship between variables. The techniques help predict the value of a dependent variable, given values of independent variables W W W. I N F O R M S . O R G

and are widely used for prediction and forecasting. Most common methods of regression such as “ordinary least squares” and “maximum likelihood estimation” require that the number of variables be less than the number of observations. In a big data environment, where increasing newer data sets are being incorporated, the number of independent variables available often greatly exceeds the number of observations. A case in point is the study of genes, where the different types of genes are the independent variables and the number of patients in a study is the observations. Another good example is texture classification of images where the variables are the pixels and observations are the number of images available for observation. In addition to this, the analyst also has to address some very important issues. For example, do the new variables really help improve the accuracy of the prediction? In general, not all variables contribute to an improved accuracy of the model. Typically, only a few of the large number of potentially influential factors account for most of the variation. To handle this complexity of variable selection brought about by increasing number of data sets available for analysis through big data techniques, a few methods have gained attention and

adoption, such as subset selection for regression, penalized regression, Biglm, Revolution R and Distributed-R Vertica, and the split-and-conquer approach. For a more technical discussion of each of these methods, click here. ANALYTICS TECHNIQUE: CLUSTERING. Segmentation, using clustering techniques, is a common method used to reveal natural structure of data. Cluster analysis involves dividing the data into useful as well as meaningful groups where objects in one group (called a cluster) are more similar to each other than to those in other groups. In general, a clustering technique should have the following characteristics to be suitable for use in a big data environment: It should be able to capture clusters of various shapes and sizes, effective treatment for outliers and be able to efficiently execute the algorithms for large data sets. Most partitional and hierarchical methods that rely on centroid-based approaches to clustering do not work very well in large data sets where the underlying data supports clusters of different sizes and geometry. Techniques such as DBSCAN (Density Based Spatial Clustering of Application with Noise) [1] can help find clusters with arbitrary shapes. It works by determining M A Y / J U N E 2 014

B IG DATA

the density associated with a point by counting number of points in a region of a specified radius around a point. Points with a density above a threshold are classified as core points, while noise points are defined as non-core points that don’t have core points within the specified radius. Noise points are discarded and clusters are formed around core points. This very idea of density-based identification of a cluster helps in creating clusters of various shapes. CURE (Clustering with Representatives) [2] also does well at capturing clusters of various shapes and sizes, since only the representative points of a cluster are used to compute its distance from other clusters. The clustering algorithm starts with each input point as a separate cluster, and at each successive step merges the closest pair of clusters. The representative points help in capturing the different physical shape and size of the clusters. A clustering technique called BIRCH (Balanced Iterative Reducing and Clustering Using Hierarchies) [3] is effective in managing outliers. It works by first performing a “pre-clustering phase” in which dense regions of points are represented by compact summaries, and then a centroid-based hierarchical clustering algorithm is used to cluster the set of summaries (which is much smaller than the original data sets). 52

A N A LY T I C S - M A G A Z I N E . O R G

SUMMARY Big data has influenced the entire spectrum of analytics – from data ingestion, storage, preparation, modeling and deployment. Organizations are exhibiting rigor in trying to harness big data through innovations in information infrastructure and information management. New core analytics techniques and practices are also changing to accommodate for the challenges of volume and variety of data associated with big data. A new era of volume, velocity and variety is leading the way for value creation in organizations like never before. Dr. Pramod Singh is director of Digital and Big Data Analytics at Hewlett-Packard and a member of INFORMS. Ritin Mathur is senior manager of Big Data Analytics at HP. Arindam Mondal and Shinjini Bhattacharya are data scientists at HP. All four are based in Bangalore, India. REFERENCES 1. Levent Ertoz, Michael Steinbach and Vipin Kumar, “DBSCAN: Finding Clusters of Different Sizes, Shapes and Densities in Noisy, High Dimensional Data,” paper, Department of Computer Science, University of Minnesota. 2. Sudipto Guha, Rajeev Rastogi and Kyuseok Shim, “CURE: An Efficient Algorithm for Large Databases,” Proceedings of the ACM SIGMOD Conference, 1998. 3. Tian Zhang, Raghu Ramakrishnan and Miron Livny, BIRCH presentation, 2009.

Subscribe to Analytics It’s fast, it’s easy and it’s FREE! Just visit: http://analytics.informs.org/

W W W. I N F O R M S . O R G

Wanted: Big Data Experts. Introducing the SMU Cox Master of Science in Business Analytics. Across a variety of industries, there’s a need for big data problem solvers. And the SMU Cox School of Business is filling it. We’re training the best and brightest to translate big data into the big picture. All in less than a year’s time. Whether you’re looking to hire the most wanted or become one of them, look no further. Learn more at coxmsba.com.

SMU is an Affirmative Action/Equal Opportunity Institution.

ADVEN T U RE S I N C O NS U LT I NG

Whatâ&#x20AC;&#x2122;s a defense consulting company doing in sports? BY STEPHEN CHAMBAL he Perduco Group is a small defense consulting company in Dayton, Ohio, near Wright Patterson Air Force Base (WPAFB). The company specializes in high-end data analytics with core competencies in data architecting, business intelligence and business analytics. Perduco was formed in 2011 and has grown rapidly to 21 employees. Twenty of those employees are working in the defense space, and one, Dr. Jacob Loeffelholz, is the lead for a strategic push into the sports domain. Now, what is a defense consulting company doing with a director for sports analytics? The easy answer is simple: sports are fun. However, the real

A N A LY T I C S - M A G A Z I N E . O R G

answer is more tied to core capabilities, business opportunity, two chance encounters and a willingness to believe in what youâ&#x20AC;&#x2122;re doing! BIG, UGLY DATA Perducoâ&#x20AC;&#x2122;s data competency is tied to integrating and aggregating big, ugly data. This includes the use of enterprise data architecting to build the data infrastructure required when solving client problems. On the analytics side, the company specializes in advanced or predictive analytics and has broad-based expertise and experience in the field of operations research (O.R.). This core capability in O.R., coupled with an ability to visualize W W W. I N F O R M S . O R G

Football heatmap: visualizing player performance in any situation.

Peyton Manning Heat Map and communicate results, gives Perduco a competitive advantage in the defense space. This same advantage can be applied to other business areas such as energy, healthcare and finance. However, a chance encounter and 30-minute discussion brought a new opportunity – sports – to the forefront. I retired from the United States Air Force in 2011 and partnered with Toyzanne Mason to form The Perduco Group. My final assignment in the Air Force was spent serving on faculty at the Air Force Institute of Technology (AFIT) at WPAFB. On a return visit to AFIT in 2012, I stopped in to visit with Dr. Ken Bauer, professor in the Department of Operational Sciences at AFIT. We traded updates, and the topic turned to sports. Ken mentioned an article A NA L Y T I C S

he had recently published with Jacob on predicting NBA outcomes using artificial neural networks. I knew Jacob from his time at AFIT and had even helped Jacob find a job with a defense consulting company (not Perduco) after graduation. What he did not know, however, was that Jacob’s article had just topped 1,000 full downloads from the Journal of Quantitative Analysis in Sports. For those not familiar with academic literature, this is an uncommonly high number. I left AFIT that day knowing two things: Perduco was going into the sports domain, and Jacob was the right person to lead this push. I first called my business mentor and vice president for Perduco, Chris Mason, and brought up the idea of expanding Perduco into the sports sector. As a M A Y / J U N E 2 014

ADVEN T U RE S I N C O NS U LT I NG

VFT player evaluation: converting subjective assessments into quantifiable rankings. sports zealot, Chris was very interested but needed to understand the business case before committing to this decision. Jacob and I met to discuss the potential opportunities and the growing application of analytics in sports. Jacob highlighted a number of areas where analytics could be leveraged in the sports domain. More importantly, they both recognized the limited use of advanced analytics from the O.R. discipline, which could quickly be adapted to solve many problems in sports, as many of the challenges are similar to those being faced in defense. In fact, Perduco was already working in a number of these crossover areas. 56

A N A LY T I C S - M A G A Z I N E . O R G

‘QUANTIFIED WARRIOR’ The defense community is very interested in what is called “quantified warrior,” which is an ability to monitor and assess a soldier’s condition and to understand how to optimize their performance on the battlefield. As one can imagine, athletic teams are very interested in maximizing their players’ performance “on the battlefield.” A second example is related to overhead monitoring and the collection of surveillance data to understand and predict activities on the ground. The intelligence community is continuously engaged in this type of effort in order to provide better assessments of enemy activity to commanders on the ground. W W W. I N F O R M S . O R G

Scout scheduling: maximizing the value of a scout on the road. STATS Inc. recently installed SportVU, an overhead surveillance system in all professional basketball arenas to track player movements on the court at all times. The collection systems allow coaches or team â&#x20AC;&#x153;commandersâ&#x20AC;? to better analyze team and opponent behavior and evaluate performance on the court. There are countless other examples of defense crossover areas, but in every case, O.R. is a critical requirement in solving these problems. The business opportunities were there, and in the summer of 2012, Perduco made the corporate decision to hire Jacob as the lead for all things sports. Jacobâ&#x20AC;&#x2122;s first task was to determine where A NA L Y T I C S

the company should focus in this very broad business area called sports analytics. This led to a one-year path of exploration, investigating a number of ways to generate revenue through the application of advanced analytics and O.R. This also led to a second chance encounter which would result in Perduco finding their sports focus in the summer of 2013. Before jumping ahead, there is useful insight into covering two major explorations, which are, in some ways, still being pursued within the company. The first push into the sports sector came in the form of team consultation. Anyone passionate about sports analytics M A Y / J U N E 2 014

ADVEN T U RE S I N C O NS U LT I NG

has read the book or seen the movie “Moneyball.” The idea of helping teams pick players, make trades, analyze strategy and win more games is what sports geeks get excited about – the sexy side of sports analytics. Perduco aggressively pursued direct team consultation and met with a large number of organizations at both the professional and collegiate levels and across multiple sports including football, basketball, baseball and hockey. Professional basketball quickly became the target of interest, and Perduco developed a number of prototype applications to demonstrate the benefits of O.R. capabilities to team management. Scout scheduling optimization, aggregate player evaluation and prediction of player performance are just a few of the solutions presented to professional teams. Although these solutions were well received, Perduco gained little traction with respect to establishing formal consulting agreements with organizations. TEAM CONSULTATION Perduco discovered a number of challenges when pursuing direct team consultation. Teams are very protective of their data and even more protective of the questions related to team and player challenges. Even with a willingness to sign non-disclosures and protect data sources, teams are hesitant to fully share information with respect to organizational decisions. 58

A N A LY T I C S - M A G A Z I N E . O R G

Additionally, most teams were interested in how the capabilities presented were being used by other teams. The idea being, if no one else was using the tools, there was no forcing function for them to adopt new analytical methods. Put another way, teams are very interested in being “first to be second.” The challenge then became getting the first team on board, which highlighted the most critical issue limiting the expansion into sports – relationships. Relationships are the foundational building blocks required to make any business successful. Perduco had many strategic relationships in defense, but building a similar network of connections in professional sports would take time to develop. The company continues to pursue team consultation, and in January of 2013, an NBA Western Conference team began using Perduco’s scout scheduling optimization tool and is now seeing great benefits with their use of this capability. The second major push into sports came in the area of business-to-business consulting. The connections are easier to find, and the relationships are easier to build. Companies supporting the sports industry are very interested in new capabilities and are willing to discuss challenge areas where they see a need for analytical solutions. New or advanced capabilities give companies a competitive advantage in a rapidly W W W. I N F O R M S . O R G

changing and dynamic market space. Perduco met with a number of sports companies, including STATS Inc., and continued to gain exposure for their predictive analytic expertise. The business case is working, but it takes considerable time and resources to develop solutions at the level required for these companies to market to their client base. Business-to-business was expected to be the major push in sports until a second chance encounter presented Perduco with a new option and a major shift in company focus.

ontinuing ducation

ESSENTIAL PRACTICE SKILLS FOR ANALYTICS PROFESSIONALS Topic areas:

» Problem Framing

» Developing the Work Plan » Testing Recommendations

Next time I get a visionary assignment that needs some clarity, I’ll be using what I learned in this course to work towards a great solution! - Caroline Alexander, Fed Ex Corporation

A NA L Y T I C S

ANALYTICS

PROFESSIONALS

INFORMS Continuing Education program offers intensive, two-day in-person courses providing analytical professionals with key skills, tools, and methods that can be implemented immediately in their work environment.

Topics areas:

This course will be held San Jose, CA – June 20-21, 2014

”

Learn more about these courses at: informs.org/continuinged

COURSES FOR

DATA EXPLORATION & VISUALIZATION

» Presenting Results » Impact Assessment

“

Prior to my 20-year reunion, I reconnected with an old friend and classmate from the Air Force Academy, Troy Neihaus. We talked over the phone and during this call, Troy mentioned Perduco’s website and noticed the push into sports analytics. He had a friend who was part-owner of a company in the fantasy football space and offered to make the connection. The next day, Perduco was introduced to Full Time Fantasy, a company that runs a very popular fantasy football website FFToolbox.com. The first phone call lasted only 30 minutes,

» Accessing Data » Understanding Raw Data » Cleaning & Transforming Data » Exploring & Visualizing Data » Dimension Reduction

“

The datasets on which the exercises are based are taken from real-life scenarios, are fun to work with and very challenging. The course provides a general framework for tackling data analysis and the instructors highlight the pitfalls one can made along the process. - Ivan Hernandez, Stevens Institute of Technology

”

This course will be held San Jose, CA – June 25-26, 2014

M A Y / J U N E 2 014

ADVEN T U RE S I N C O NS U LT I NG

but the decision was made to shift focus, integrated capabilities in the summer of resources and energy into web-based 2014. A number of capabilities are befan support for fantasy sports. ing developed, tested and implemented into the FFToolbox infrastructure. These FANTASY SPORTS capabilities will provide users access to Jacob, working with FFToolbox, advanced analytics for analyzing players, quickly researched the fantasy sports simulating drafts and visualizing statistilandscape and identified numerous op- cal information in an easy to use and unportunities for advanced analytics and derstandable format. The goal is to put data visualization. The two companies the power of O.R. in the hands of active began developing a fantasy sports players strategy to move forwho may not otherwise ward and outlined both have access to these the execution and busitypes of mathematical ness plan to make this happen. The in- algorithms. FFToolbox gives their custegration of FFToolbox’s expertise in tomers a competitive advantage of their fantasy sports and Perduco’s depth in own when competing in daily, weekly O.R. provides enormous opportunity and season-long fantasy sports games. to bring advanced and unique capabili- The fantasy sports market is an exploties to the market space. Furthermore, sive industry with millions of potential FFToolbox is already recognized as a customers, each looking for an edge and leader in the space with nearly 10 mil- bragging rights when competing against lion users visiting their website on an family, friends, co-workers and unknown annual basis. This footprint provides opponents across the world. immediate exposure for Perduco analytPerduco has spent nearly two years ics at a level never imagined this early committed to expanding into the sports in the company’s expansion into sports. domain. Perduco’s core capabilities in In fact, the building footprint in sports operations research and its passion to has led to the spin off of Perduco Sports bring advanced analytics to the sports from the parent company, The Perduco industry has kept the company moving Group, which happened in the spring of forward at all times. The expected busi2014. ness areas have shifted here and there, Full Time Fantasy and Perduco but with each shift came increased opare on schedule to release their first portunities for success. The chance 60

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

encounter with Bauer in summer of 2012 led to the hiring of Loeffelholz and the launch of sports analytics within Perduco. The chance encounter with Neihaus in summer of 2013 led to the introduction of Perduco to Full Time Fantasy and the launch into the fantasy sports market. Believing in chance encounters and being ready and willing to react when they happen has been a driving force behind this adventure. If the pattern continues, the next chance encounter should be coming soon in summer of 2014 and could ignite the viral power of the Internet

and launch Perduco into an overnight success, 24 months in the making! Dr. Stephen Chambal (Stephen.chambal@ theperducogroup.com) is vice president of The Perduco Group and is responsible for strategic business development and high-end recruiting. Chambal retired from the United States Air Force in 2011 after more than 24 years of service. In his final military assignment, he served as the director of Operational Analysis for the Air Force Institute of Technology. Chambal enlisted in the Air Force in 1986 and obtained his commission from the Air Force Academy in 1993. He held various assignments within the scientific analysis career field, including test, space and special programs. Chambal holds a Ph.D. in operations research and has authored and co-authored numerous articles, white papers and conference presentations.

BECOME A CERTIFIED ANALYTICS PROFESSIONAL ®

www.informs.org/Build-Your-Career/Analytics-Certification

Be among the first to become a Certiﬁed Analytics Professional (CAP®). Make plans now to take the profession's first analytics certification exam.

BENEFITS OF CERTIFICATION

• Advances your career potential by setting you apart from the competition • Drives personal satisfaction of accomplishing a key career milestone • Helps improve your overall job performance by stressing continuing professional development • Recognizes that you have invested in your analytics career by pursuing this rigorous credential • Boosts your salary potential by being viewed as experienced analytics professional CBT • Shows competence in the principles and practices of analytics

APPLICATIONS

• Prepare to apply by reviewing Candidate Handbook & Study Guide • Arrange now to secure academic transcript and confirmation of “soft skills” to send to INFORMS

COST

• $495 INFORMS Members • $695 Non-Members • Bundled rates with meetings available

QUESTIONS? certification@mail.informs.org

A NA L Y T I C S

STARTING JUNE 1

UPCOMING CAP® EXAMS MAY 17 Baltimore, MD

YOUR SCHEDULE Kryterion Test Centers Computer Based Test (CBT)

MAY 22 Cincinnati, OH MAY 26 Ottawa, ON JUNE 19 Chicago JUNE 21 San Jose, CA

NOW INTRODUCING COMPUTER-BASED TESTING It is now more convenient than ever to schedule your CAP exam in more than 700 Kryterion test centers in more than 100+ countries. No more waiting for INFORMS to come to you! To find the location closest to you, check this site: http://www.kryteriononline.com/host_locations/

M A Y / J U N E 2 014

Special Advertising Supplement

INFORMS would like to thank and salute the following sponsors of the Franz Edelman Awards Gala who contributed at the Leadership Level and Executive Level. Leadership Level ($10,000) IBM Executive Level ($5,000) AIMMS Alix Partners Booz Allen Hamilton Institute for Information Industry Metron Revenue Analytics Spinnaker University of Cincinnati

Smarter technology for a Smarter Planet:

Anyone can make a decision. The question is: Can you make the right one?

Want to see the power of CPLEXÂŽ and Decision Optimization in action? Visit us at the booth to see how IBMâ&#x20AC;&#x2122;s innovative technology and services are fundamentally redefining how real-world companies meet their performance goals.

OPTIMIZE YOUR BUSINESS WITH UNPRECEDENTED SPEED IDEA

IN A FEW HOURS

MISSION CRITICAL ENTERPRISE APP

IN A FEW MONTHS

PUBLISHED INSTANTLY TO YOUR ENTERPRISE OPTIMIZATION APP STORE

PROOF OF CONCEPT

IN A FEW DAYS

OPTIMIZATION APP

IN A FEW WEEKS

To learn more about AIMMS Optimization Apps, visit aimms.com. info@aimms.com | +1 425 458 4024

S T R AT E G Y & O R G A N I Z AT I O N

TECHNOLOGY

E N G I N E E R I N G & O P E R AT I O N S

A N A LY T I C S

2012 INFORMS Innovation in Analytics Award Winner

Make sense of your data.

Achieve operational efficiencies.

Exceed your organizational goals.

Ready for what’s next. Now more than ever, making sense of large, complex data sets is critical to an organization’s success. Supported by extensive industry experience, our analytics experts are armed to help clients make informed decisions and turn choices into action. These insights help our clients anticipate, identify, and address their specific needs—from achieving cost savings to reducing fraud, waste, and abuse. We leverage these analytics capabilities to help clients across US and international governments in defense, intelligence, and civil sectors, as well as commercial health, financial services, and energy and utilities industries. Whether you’re managing today’s issues or looking beyond the horizon, count on us to help you be ready for what’s next. Ready for what’s next. www.boozallen.com/analytics

Congratulations to the 2014 Franz Edelmen Finalists!

SCIENTIFIC SOLUTIONS

Applied Mathematics, Physical Modeling, and Computer Science

www.metsci.com

The Lindner Master of Science in Business Analytics

TOPTop 20 PROGRAM in the country

â&#x20AC;&#x201D;InformationWeek

Learn more: business.uc.edu/analytics UCLindnerCollege

@UCBusAnalytics | @LindnerCollege

Center for Business Analytics

Salutes the Sponsors of the 2014 Edelman Gala

RECOGNIZING DISTINCTION IN APPLICATIONS OF ANALYTICS, OPERATIONS RESEARCH, AND MANAGEMENT SCIENCE LEADERSHIP LEVEL

EXECUTIVE LEVEL

CHAMPION LEVEL

SUSTAINER LEVEL

www.informs.org www.informs.org/aboutanalytics

March 31, 2014 Boston, MA

PR ED IC TIVE A N A LY T I C S

Most swans are white: Living in a predictive society My view – and Siegel’s, I would guess – is that this predictive activity has generally been good for humankind. In the context of healthcare, crime and terrorism, it can save lives. In the context of advertising, using predictions is more efficient … In politics, it seems to reward those candidates who respect the scientific method.

BY THOMAS H. DAVENPORT 62

Eric Siegel’s book – “Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die”– deals with quantitative efforts to predict human behavior. One of the earliest efforts to do that was in World War II. Norbert Wiener, the father of “cybernetics,” began trying to predict the behavior of German airplane pilots in 1940 – with the goal of shooting them from the sky. His method was to take as input the trajectory of the plane from its observed motion, consider the pilot’s most likely evasive maneuvers, and predict where the plane would be in the near future so that a fired shell could hit it. Unfortunately, Wiener could predict only one second ahead of a plane’s motion, but 20 seconds of future trajectory were necessary to shoot down a plane. In Siegel’s book, however, you will learn about a large number of prediction efforts that are much more successful. Computers have gotten a lot faster since Wiener’s day, and we have a lot more data. As a result, banks, retailers, political campaigns, doctors and hospitals, and many more organizations have been quite successful of late at predicting the behavior of particular humans. Their efforts have been helpful at winning customers, elections, and battles with disease. My view – and Siegel’s, I would guess – is that this predictive activity has generally been good for

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

humankind. In the context of healthcare, crime and terrorism, it can save lives. In the context of advertising, using predictions is more efficient, and could conceivably save both trees (for direct mail and catalogs) and the time and attention of the recipient. In politics, it seems to reward those candidates who respect the scientific method (some might disagree, but I see that as a positive). However, as Siegel points out – early in the book, which is admirable – these approaches can also be used in somewhat

harmful ways. “With great power comes great responsibility,” he notes in quoting Spider-Man. The implication is that we must be careful as a society about how we use predictive models, or we may be restricted from using and benefiting from them. Like other powerful technologies or disruptive human innovations, predictive analytics is essentially amoral, and can be used for good or evil. To avoid the evil applications, however, it is certainly important to understand what is possible with predictive analytics, and you will

How will you stand out from the crowd?

A membership in INFORMS will help! • Certification for Analytics Professionals • Online access to the latest in operations research and advanced analytics techniques • Networking Opportunities available at INFORMS Meetings and Communities • New Members receive one free Subdivison membership in 2014

A NA L Y T I C S

visit http://join.informs.org

M A Y / J U N E 2 014

PR ED IC TIVE A N A LY T I C S

certainly learn that if you keep reading. This book is focused on predictive analytics, which is not the only type of analytics, but the most interesting and important type. I don’t think we need more books anyway on purely descriptive analytics, which only describe the past, and don’t provide any insight as to why it happened. I also often refer in my own writing to a third type of analytics—“prescriptive”— that tells its users what to do through controlled experiments or optimization. Those quantitative methods are much less popular, however, than predictive analytics. This book and the ideas behind it are a good counterpoint to the work of Nassim Nicholas Taleb. His books, including “The Black Swan,” suggest that many efforts at prediction are doomed to fail because of randomness and the inherent unpredictability of complex events. Taleb is no doubt correct that some events are black swans that are beyond prediction, but the fact is that most human behavior is quite regular and predictable. The many examples that Siegel provides of successful prediction remind us that most swans are white. Siegel also resists the blandishments of the “big data” movement. Certainly some of the examples he mentions fall into this category – data that is too large or unstructured to be easily managed by conventional relational databases. But the 64

A N A LY T I C S - M A G A Z I N E . O R G

point of predictive analytics is not the relative size or unruliness of your data, but what you do with it. I have found that “big data often equals small math,” and many big data practitioners are content just to use their data to create some appealing visual analytics. That’s not nearly as valuable as creating a predictive model. Siegel has fashioned a book that is both sophisticated and fully accessible to the non-quantitative reader. It’s got great stories, great illustrations and an entertaining tone. Such non-quants should definitely read this book, because there is little doubt that their behavior will be analyzed and predicted throughout their lives. It’s also quite likely that most non-quants will increasingly have to consider, evaluate and act on predictive models at work. In short, we live in a predictive society. The best way to prosper in it is to understand the objectives, techniques and limits of predictive models. And the best way to do that is simply to read Siegel’s book. Thomas H. Davenport (www.tomdavenport.com) is a visiting professor at Harvard Business School, the President’s Distinguished Professor at Babson College, co-founder of the International Institute for Analytics and the co-author of “Competing on Analytics” and several other books on analytics. He is a member of INFORMS. This foreword by Professor Davenport is excerpted with permission of the publisher, Wiley, from “Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die” (February 2013) by Eric Siegel and is reprinted for promotional considerations.

W W W. I N F O R M S . O R G

Is your solver provider

getting in the way of your success?

At Gurobi, we make it easier to achieve your optimization goals. Faster solutions with superior algorithms and powerful tuning Maximum productivity with flexible interfaces and modeling language support Great technical support from easy-to-reach optimization experts No surprises with our transparent pricing and flexible licensing

Gurobi Optimizer 5.6

State-of-the-Art Mathematical Programming Solver Experience state-of-the-art optimization algorithms backed by industry-leading customer support.

Gurobi.com +1 713-871-9341 info@gurobi.com

Get started with a FREE TRIAL today!

CO N FERE N C E P R E V I E W

INFORMS Big Data Conference June 22-24 event in San Jose, Calif., to focus on the business of big data. INFORMS is launching a new topical conference that will put the focus squarely on the business of big data – how organizations can transition from being data-rich to decision-smart. The INFORMS Big Data Conference will be held June 22-24 at the San Jose Convention Center in San Jose, Calif. The conference is organized around tracks on “Big Data Case Studies,” “Big Data 101” and “Emerging Trends in Big Data.” Case studies of big data projects that illustrate the complete journey from business problem to analytics solutions will be a major component of the conference. Sessions on big data 101 will offer tutorials on how to navigate the big data ecosystem, how to select and use the right technologies, as well as the challenges of building data science teams. Critical topics such as ethics and privacy requirements will also be addressed. Other sessions on the program will look into the future, exploring emerging technologies and business trends. 66

A N A LY T I C S - M A G A Z I N E . O R G

Bill Franks

Michael Svilar

The committee has hand-picked speakers to address defined topics in an intensive program geared to the interests of business decision-makers, IT managers and analytics professionals. Poster sessions, technology workshops, tutorials and panel discussions will also be part of the program, as well as facilitated networking opportunities. Bill Franks, chief analytics officer at Teradata Corporation, and Michael Svilar, managing director, delivery lead and capability lead at Accenture, will deliver keynote presentations. At Teradata and throughout his career, Franks W W W. I N F O R M S . O R G

has focused on translating complex analytics into terms that business users can understand and then helping organizations implement the results effectively within their processes. He is the author of the book â&#x20AC;&#x153;Taming the Big Data Tidal Waveâ&#x20AC;? and holds a patent in forecasting analytics. For more than 30 years, Svilar has run analytics projects across multiple industries including retail, communications, financial services, automotive, consumer packaged goods and electronics.

Early registration rates will be available until May 23. Rooms are being held at the Marriott San Jose at a discounted rate until May 26. Conference organizers anticipate that rooms will sell quickly and advise attendees to make reservations early, well before the cut-off date. After that date, reservations will be accepted at prevailing rates on a room available basis. Visit http://meetings.informs.org/bigdata2014/ for more information on the INFORMS Big Data Conference.

Would like to congratulate the 2014 Edelman Award winner U.S. Centers for Disease Control and Prevention with KidRisk, Inc.

Click here to relive the excitement of the winner announcement on YouTube.

A NA L Y T I C S

M A Y / J U N E 2 014

CO N FERE N C E P R E V I E W

Program tracks and speakers Track: Big Data Case Studies •G oogle – Behdad Masih, quantitative analyst •H ERE, a Nokia Business – Toby Tennent, product development manager •E xponent®, Inc. – Juergen Klenk, principal scientist •K abbage, Inc. – Pinar Donmez, chief data scientist •U .S. Census Bureau – Cavan Capps, big data lead •M erck & Co., Inc. – John E. Koch, director of informatics • I ntel Corp. – Link C. Jaw, Internet-ofthings solutions group •J P Morgan Intelligent Solutions – Govind Nagubandi, data scientist •A nalytics Media Group (AMG) Alan Papir, software engineer

•K aiser Permanente – Anton J. Mobley, MS, security data scientist •B ooz Allen Hamilton – Peter Guerra, BS, principal • I BM – Kevin Foster, MS, big data solutions architect •K overse, Inc. – Paul Brown, MS, CEO and founder •V erizon Wireless – Anne Robinson, Ph.D., director, supply chain strategy & analytics •B ooz Allen Hamilton – Brian Keller, Ph.D., data scientist

Track: Emerging Trends in Big Data • I BM – Rob High, BA, chief technology officer, Watson solutions, IBM fellow and vice president •L inkedIn Corp. – Simon Zhang, MD &

•T eradata Aster – Lee Paries, vice

MBA, senior director, business analytics

president, Central & Western U.S.

•S AS – Paul Kent, B Commerce (WITS),

•B ell Laboratories – Marina Thottan, director •A lpine Data Labs – William C. Ford, lead data scientist

vice president, big data •U niv. of California Berkeley – Ion Stoica, Ph.D., professor, Univ. of California Berkeley; CEO, Databricks; CTO, Conviva

Track: Big Data 101 •A mazon – Vikram Garlapati, manager, solutions architect •N orthwestern University – Diego Klabjan, professor, industrial engineering & management sciences; director, master of science in analytics program

A N A LY T I C S - M A G A Z I N E . O R G

•K aggle – Anthony Goldbloom, B Commerce, founder and CEO •S AP – Chris Hallenbeck, global vice president •P ivotal – Kaushik Das, senior principal data scientist • GraphLab – Shawn Scully, data scientist

W W W. I N F O R M S . O R G

Organizing Committee General Chair Candace A. Yano University of California-Berkeley Program Chair Philip Kaminsky University of California-Berkeley Plenary/Keynotes Chair Shmuel S. Oren University of California-Berkeley Invited Sessions Co-Chairs Hyun-Soo Ahn Damien Beil University of Michigan Sponsored Sessions Co-Chairs Alper Atamturk Zuo-Jun Max Shen University of California-Berkeley Contributed Sessions Co-Chairs Rachel Chen University of California-Davis Steven Nahmias Santa Clara University Practice Program Co-Chairs Vijay Mehrotra San Francisco State University Warren Lieberman Veritec Solutions Thomas Dag Olavson Google, Inc.

November 9-12, 2014 Hilton San Francisco & Parc 55 Wyndham San Francisco, California

Submission Deadline: May 15, 2014 Submit Early, Capacity Limited! Registration opens June 1, 2014

meetings2.informs.org/sanfrancisco2014

Interactive Sessions Co-Chairs Hari Balasubramanian Ana Muriel Univ. of Massachusetts-Amherst Tutorials Co-Chairs Alexandra M. Newman Colorado School of Mines Janny Leung Chinese University of Hong Kong Arrangements Co-Chairs Julia Miyaoka Theresa M. Roeder San Francisco State University

INFO RM S IN I T I AT I VE

Continuing ed courses set for San Jose INFORMS’ popular continuing education courses will be held June 20-21 (“Essential Practice Skills for Analytics Professionals”) and June 25-26 (“Data Exploration & Visualization”) before and after, respectively, the 2014 INFORMS Conference on the Business of Big Data in San Jose, Calif. Register early to save $100.

Essential Practice Skills for Analytics Professionals Taught by Dr. E. Andrew Boyd, Texas A&M, University of Houston, and Houston Public Media Participants will learn practical tools for integrating their analytical skills into realworld problem solving for businesses and other organizations. The course provides

Data Exploration & Visualization Taught by Stephen McDaniel and Eileen McDaniel, Freakalytics, LLC Attendees will experience first-hand the power of visualization in exploring data as an adjunct to tried and trusted analytics methods. Using a hands-on lecture and case study approach, attendees will walk away with a proven framework for data exploration that is directly relevant to real-world problems in many fields. Incorporating popular free and leading software tools to facilitate hands-on learning, attendees will be able to directly apply the methods and techniques from this course in their work, regardless of their visualization tool of choice. At the end of the course, participants are

approaches that can be applied immediately to a wide variety of settings, whether within

expected to:

a participant’s own organization or for an

• have confidence to explore new data using

external client. By the end of the course,

the exploration/visualization approach;

participants will:

• have the ability to approach and deploy

• learn to link their subject-matter expertise

interactive visualization;

to the challenges of messy, unstructured

• understand how to identify practically

problems, organizational noise, and non-

meaningful discoveries;

technical decision makers; and • understand best-practice techniques,

• experience the use of state-of-the-art

including: problem statement summaries,

visualization software; and

issue trees, interview guides, work

• think creatively about data and insights.

plans, sensitivity analysis, stress-testing recommendations, story-boarding, slide-

Course dates: June 25-26, San Jose, Calif.

craft, delivering presentations and fielding

For more information on the course

Q&A sessions. Course dates: June 20-21, San Jose, Calif.

and available discounts, visit www.informs.org/ continuinged.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Connect with the earned expertise of business forecasters and practical research from top academics from around the globe.

Each issue of Foresight contains articles that you’ll use in your day-today work, whatever types of forecasting you do.

““

Here’s what our readers say:

Issue 26 Summer 2012 THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING

“The information is relevant to practitioners and is presented in a way that is not overly academic but with significant credibility.” Thomas Ross, Financial Analyst, Brooks Sports

5 Setting Internal Benchmarks Based on a Product’s ForecaStaBIlIty DNa

“Foresight make(s) important research findings available to the practitioner.”

18 Regrouping to Improve Seasonal Product Forecasting 32 Forecasting Software that Works For – Not against – Its Users

38 Book Review Abundance: The Future Is Better Than You Think 41 reliably Predicting Presidential elections

Anirvan Banerji, Economic Cycle Research Institute

“...an important forum for practitioners to share their experiences....”

Issue 27 Fall 2012 THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING

Dan Kennedy, Senior Economist, Connecticut Department of Labor

“I find Foresight very useful! I use it as a teaching resource to bring theoretical forecasting techniques to life for the students.” Dr. Ilsé Botha, Senior Lecturer, University of Johannesberg

5 Special Feature: Why Should I Trust Your Forecasts? 23 Tutorial: The Essentials of Exponential Smoothing 29 S&OP: Foundation Principles and Recommendations for Doing It Right

40 New Texts for Forecasting Modelers

Put Foresight to work to improve your forecasts and rally support for the people, processes and tools that accurate forecasting requires.

Subscribe today! forecasters.org/foresight/subscribe/

Issue 29 Spring 2013 THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING THE INTERNATIONAL JOURNAL OF APPLIED FORECASTING

5 Forecasting revenue in Professional Service Companies 14 Forecast value added: A Reality Check on Forecasting Practices 19 s&oP and Financial Planning 26 cPFr: Collaboration Beyond S&OP 39 Progress in Forecasting rare events 50 Review of "global trends 2030: alternative Worlds"

FIVE- M IN U T E A N A LYST

Buffet’s billiondollar basketball bracket bet

A collection of Tribbles, which despite their harmless appearance, quickly grow to fill the space. Much like Factorials.

BY HARRISON SCHRAMM, CAP 72

Warning: Factorials and powers are ubiquitous in this article. Like Tribbles from Star Trek, expressions like “64” look cute and innocent, but they are some of the most deadly mathematical beasts known to man.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

PHOTO BY STEPHEN SLADE. COURTESY OF UNIVERSITY OF CONNECTICUT

Shabazz Napier (13) leads UConn to victory over Kentucky in the 2014 NCAA Championship game. I was bemused to read this article [1] in Slate magazine detailing the odds of Warren Buffet’s basketball challenge, which may be found here [2]. [Buffet offered a billion dollars to anyone who submitted a perfect bracket (i.e., correctly predicting the winner of all 63 games) of “March Madness,” otherwise known as the NCAA Men’s Basketball Championship Tournament.] A billion dollars – even with taxes – is a lot of

money. How hard is it to come up with a perfect bracket? There is only one perfect bracket in a world with many potential brackets, so we first need to find out how many possible brackets there are. The NCAA is a single elimination tournament, which means that each team plays until they lose. In a single elimination tournament, each round is made up of n teams, with n / 2 games n /2 played. Therefore, there are 2 possible outcomes in the first round. Knowing that the tournament starts with 64 teams, there are possible outcomes for the first round. Using similar calculations at each round, there are possible outcomes, only one of which is correct. For comparison’s sake, 1 billion is a thousand million, or 109 so the odds of winning the basketball challenge 9 2 are around 1: (10 ) or one in a billion billion 3 . So, it appears your odds of winning are not very good at all. Frequently, one can get a feel for the value of a gambling game by the “fair price” that one would be willing to pay to play the game; specifically, the value that would make one indifferent between playing the game and just keeping their money. For this game, a “fair” price would be nine million attempts per penny! M A Y / J U N E 2 014

FIVE- M IN U T E A N A LYST

SOME EXCURSIONS Suppose I could have one piece of information. A likely choice would be: “How much better off would we be if we knew the eventual winner?” In this case, we would reduce the number of possibilities in each round by 1; and the number of combinations would “only” be , which is 64 times better than the original estimate. Sixty-four is generally reckoned to be a small number when compared 18 to 10 . If you knew the Final Four, you would be considerably better off, at , or 520,000 times better than the original bet, which is to say that in the scheme of things, you are no better off at all. Now suppose that you had to pay $1 to play this game instead of it being free, but you think you are pretty good at predicting basketball games. You would need to have 72 percent accuracy in your ability to pick basketball games to be risk neutral for a dollar (i.e., 72 percent accuracy increases your odds of winning to 1 in 1 billion). As bad as these odds are, here’s a game that is even worse, which I will call the Georgetown Wager (after my colleague who challenged me to come up with a tougher game). Suppose that you are given the 64 teams that will play, but the games are randomized and you 74

A N A LY T I C S - M A G A Z I N E . O R G

have no information about their parings; you know that there are 64 teams in the first round, of which 32 will win, and so on, but you don’t know who will play who. In this version, you have to figure out for the number of possibilities, (1) where the “choose” function, if you expand (1) out by hand and cancel terms4, you will find that it is: . If the odds of Buffet’s Billion Bracket are bad, the Georgetown Wager is patently absurd; the odds of winning are , which are roughly on the order of winning Buffet’s game twice in a row! You may be asking: “So, if this is such a good bet for the house, why don’t I run a similar lottery?” Because I don’t have a billion dollars, and I’m not willing to lose. Remember that the “house” has to be willing to pay out the fee in the extraordinarily rare event that someone won. Events that are “statistically impossible” are still “possible.” and while it is extraordinarily unlikely that someone will win, there is no law of physics that prevents someone from winning. W W W. I N F O R M S . O R G

The perfect first round buyout: Suppose you had a perfect first round, in that you guessed the first 32 games correct. Congratulations. Strictly speaking, the risk-neutral buyout price from the bank’s perspective (i.e., Mr. Buffet) is approximately $2. Now, this figure presumes that you got to this point by dumb luck, and you will certainly claim – and the house may believe – that you got to this point because you are very good at predicting basketball games. After all, to have a 50 percent

probability of predicting the first round correctly, you would need to have ~98 percent per-game prediction accuracy. So, should you play? Sure, go ahead. Expected value calculations presume that you are going to do something else with the money; this is true for large amounts but typically not for small. So it depends on what else you would do with the money. In this particular example, you won’t pay anything, except some advertising e-mails, so if that works for you, go ahead.

video learning center Your one-stop shop to view top presentations from key INFORMS meetings NOW ONLINE! 2014 Edelman Presentations 2013 Analytics Conference and Annual Meeting 2012 Analytics Conference and Annual Meeting 2011 Analytics Conference and Annual Meeting 2010 Practice Conference and Annual Meeting 2009 Annual Meeting

Your latest member benefit lets you learn from the best on your schedule.

http://livewebcast.net/INFORMS_Video_Learning_Center A NA L Y T I C S

M A Y / J U N E 2 014

FIVE- M IN U T E A N A LYST

If a similar wager cost $1 to play, it would depend on what else you would do with the money. Foregoing a late afternoon soda to buy a ticket for this game, if you enjoy talking about it, would be OK. Dumping out your life savings in order to play games is a terrible idea (we’ve written about this before, see July 2013 [5]). The point is that we all do lots of things where the odds of winning are practically zero. This is not necessarily a bad thing. If you derive “pleasure” out of daydreaming about winning a billion dollars or have fun arguing basketball scores with your friends, go for it! Just do so with eyes open, knowing that it is incredibly unlikely that you will win. And don’t forget, there are also 20 first-prize winners, regardless of whether the grand prize is given or not, valued at $100,000. While no billion, this is no small amount of money, and most importantly, does not require you to be perfect, simply better than the other players who enter. If you think you are good at filling out your bracket, then perhaps you should enter with the hopes that you win the first prize. Here, the odds are no worse than 1:750,000, which is a number that you can start to comprehend! A note on calculation. I used R to do the large calculations in this article. 76

A N A LY T I C S - M A G A Z I N E . O R G

Professionals always need to be concerned about numerical stability and floating point precision, which may be the subject of a subsequent article. If I did not have a good computational platform or was doing this 50 years ago, I would resort to Sterling’s Approximation, It’s amazing to think about all of the computation that we simply take for granted. Finally, knowing that

is very handy. (If you need a proof, start writing out numbers in binary) Harrison Schramm (harrison.schramm@gmail. com) is an operations research professional in the Washington, D.C., area. He is a member of INFORMS and a Certified Analytics Professional (CAP).

NOTES & REFERENCES 1. http://www.slate.com/articles/sports/sports_ nut/2014/03/billion_dollar_bracket_challenge_why_ it_s_a_bad_idea_to_enter_warren_buffett.html 2. https://tournament.fantasysports.yahoo.com/ quickenloansbracket/challenge/?qls=BDB_B14qlb03. qlredirect 3. When your exponents have exponents, the numbers are really huge! 4. Comment: If you want real understanding in mathematics, there is no substitute for expanding by hand. This is how the mathematicians of 50 years ago did things, and there is goodness in it, even today. 5. http://www.analytics-magazine.org/july-august2013/838-five-minute-analyst-carnival-game

W W W. I N F O R M S . O R G

Analytics SAS and Hadoop take on the Big Data challenge. And win.

Why collect massive amounts of Big Data if you can’t analyze it all? Or if you have to wait days and weeks to get results? Combining the analytical power of SAS with the crunching capabilities of Hadoop takes you from data to decisions in a single, interactive environment – for the fastest results at the greatest value.

Read the TDWI report

sas.com/tdwi

THIN K IN G A N A LY T I CA LLY

Spy catcher

BY JOHN TOCZEK John Toczek is the senior director of Decision Support and Analytics for ARAMARK Corporation in the Global Operational Excellence group. He earned a bachelor of science degree in chemical engineering at Drexel University (1996) and a masterâ&#x20AC;&#x2122;s degree in operations research from Virginia Commonwealth University (2005). He is a member of INFORMS.

Your government has lost track of a high profile foreign spy and they have requested your help to track him down. As part of his attempts to evade capture, the spy has employed a simple strategy. Each day the spy moves from the country that he is currently in to a neighboring country. The spy cannot skip over a country (for example, he cannot go from Chile to Ecuador in one day). The movement probabilities are equally distributed among the neighboring countries. For example, if the spy is currently in Ecuador, there is a 50 percent chance he will move to Colombia and a 50 percent chance he will move to Peru. The spy was last seen in Chile and will only move about countries that are in South America. He has been moving about the countries for several weeks. Question: Which country is the spy most likely hiding in and how likely is it that he is there? Send your answer to puzzlor@gmail.com by June 15. The winner, chosen randomly from correct answers, will receive a $25 Amazon Gift Card. Past questions can be found at puzzlor.com.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

OPTIMIZATION

www.gams.com

High-Level Modeling The General Algebraic Modeling System (GAMS) is a high-level modeling system for mathematical programming problems. GAMS is tailored for complex, large-scale modeling applications, and allows you to build large maintainable models that can be adapted quickly to new situations. Models are fully portable from one computer platform to another.

State-of-the-Art Solvers GAMS incorporates all major commercial and academic state-of-the-art solution technologies for a broad range of problem types.

GAMS Integrated Developer Environment for editing, debugging, solving models, and viewing data.

A Water Management Decision Support System (DSS) for the Indus Basin Large non-linear optimization models developed in GAMS are a centerpiece of the water management DSS for Pakistan's Indus Basin. The system is used for agricultural investment planning and water management, but also to investigate the impacts of climate risks on water and agriculture. An international consortium led by National Engineering Services Pakistan (NESPAK) recently extended this system. Major features include: • The GAMS models seamlessly interact with an Oracle database, which feeds both model data and results into a Geographic Information System (GIS). • Users from government, industry, and consulting groups use the web-based application to calculate water requirements and cropping patterns. • The results are available in various formats: maps, graphs, and tables.

Europe GAMS Software GmbH info@gams.de USA GAMS Development Corporation sales@gams.com http://www.gams.com

For further information please contact Khalid Mahmood - Khalid.Mahmood@nespak-wr.com