#Analytics 2016: Richard E. Rosenthal Early Career Connection nominations now open!

The Richard E. Rosenthal Early Career Connection (ECC) program, to be held during the INFORMS Analytics 2016 conference in Orlando, FL April 10-12, 2016, is now accepting nominations. Nominations are due March 4, 2016.

Last year we redesigned the program to allow more interactions between participants and also between participants and conference attendees. Innovations that we plan to continue this year include inviting INFORMS leadership to the ECC reception for mingling with participants on Sunday, reserved tables for ECC participants and select conference attendees (invited based on ECC participants' research and work specialties) for Monday's lunch and Tuesday gathering during the coffee hour. 

From the website: 

The Early Career Connection (ECC) is a program of special events designed for professionals who are new to their academic or industry careers. The program facilitates networking and introduces participants to well-established researchers and practitioners for more effective communication. Participants benefit from a discount on registration to the conference, as well as the networking events exclusive to ECC participants. The discounted registration rate is $615.


The mission of ECC is to provide early-career professionals with new perspectives into some of the most critical problems facing industry today, enabling them to broaden their research agendas. The goal is for these analytics and OR leaders of the future to have an opportunity, early in their careers, to apply their outstanding analytical talents to important business problems.

Those nominated and selected for this honor will receive a reduction of the conference registration fee. Awardees are expected to participate fully in the ECC events, as well as the conference sessions and social programs. We also strongly encourage all ECC nominees to submit a paper or poster to the Select Presentations or Poster Presentations (however, this is not a requirement for acceptance to the ECC).

To Nominate an ECC Participant

Please send an email to by March 4, 2016, containing the following information. The nominator must be from the same organization as the nominee. 

blue-checkmark Nominee’s name, email and telephone number

blue-checkmark Nominator’s name, email and telephone number

blue-checkmark Type of degree, year of nominee’s degree, and institute that awarded the degree

blue-checkmark If nominee has a master’s degree, starting date at company

blue-checkmark Organization and department

blue-checkmark Nominee’s position at the organization

blue-checkmark A brief paragraph from the nominator explaining why this person is being nominated (50-150 words)

blue-checkmark A brief paragraph from the nominee describing their relevant research or analytics/OR project (50-150 words)

Analytics vs emotional intelligence

I wonder what my readers would say are the two main trends in the business management community today, but for me these trends are the rise of analytics (we are at Analytics 3.0 now, apparently) and the importance of emotional intelligence, pioneered by Daniel Goleman, as well as the role of grit as key to success. Goleman has written several important books on the types of skills leaders should have now to success, but this 2004 HBR article he authored gives you a preview of the four groups of EQ strengths: self-awareness, self-regulation, motivation, empathy, and social skill.

Back in 2004, Goleman wrote: "Although a certain degree of analytical and technical skill is a minimum requirement for success, studies indicate that emotional intelligence may be the key attribute that distinguishes outstanding performers from those who are merely adequate." This reminded me of what people say about tenure at a research institution: you have to be an excellent researcher but you only need to meet the minimum requirements as a teacher (be "not-terrible-enough") to succeed as a professor. It seems that what Goleman wrote in 2004 could be paraphrased as: a leader succeeds by knowing the viable minimum about the technical or analytical side of his business and excelling at emotional intelligence.

Perhaps this was understandable eleven years ago; yet I can't help but wonder if the trends of analytics and emotional intelligence will collide, or if they will merge to yield tomorrow's industry superstar, the emotionally-intelligent data scientist. Or are people who like both data and people barely more common than unicorns? (This would make them in even higher demand in the marketplace, then!) This would be a data scientist who understands himself and his motivations and also can work well with his team to achieve the desired goal of using data to help the company fulfill its mission. You will say that perhaps data scientists are not leaders, but if emotionally-intelligent data scientists are tomorrow's superstars, then they will be C-suite executives the day after tomorrow. 

Yet, implementing this vision presents a number of challenges, starting with the fact that the people teaching tomorrow's data scientists may not have high EQ. This is in part because, in some disciplines, the trend to collaborate on certain projects among colleagues rather than with one's graduate students has emerged only recently. Also, the people who teach EQ-related subjects in college may not be superstars of Daniel Goleman's caliber and may, in fact, have underwhelming credibility. Successful leaders (often leaders with high EQ) may come and give a talk at a university here and there but this does not amount to teaching students how to develop high EQ. 

Emotional intelligence is also perhaps not a skill best learned in the classroom. Perhaps it is best to let students observe peers and managers from up close once they are working and try to identify what they think they make this person successful and that less so. To learn that in college, there would need to be debriefing courses or seminars that the students take after they return from summer internships.

The best book I have found on emotional intelligence in business is the poorly-titled Primal Leadership: The Power of Emotional Intelligence. Once you get past the bad title and the just-as-bad cover, you discover a very useful book that should be required reading of all college students, in particular all students in industrial & systems engineering or operations research programs. IE/OR students are already been taught the analytical skills in such high demand, and developing their emotional intelligence would undoubtedly position them for great success in the workforce. 

#INFORMS15 TutORial 2: How Analytics can Impact Promotion Pricing

How Analytics can Impact Promotion Pricing

This tutorial will be given by Professor Georgia Perakis, William F. Pounds Professor of Management Science, MIT in Room 108A at Convention Center.

Joint work with Lennart Baardman (ORC PhD student), Maxime Cohen, (ORC PhD student), Swati Gupta (ORC PhD student), Jeremy Kalas (EECS Undergraduate), Zachary Leung (recently graduated ORC PhD student), Danny Segev (Visiting Scholar ORC/MIT Sloan from U. Haifa) 

as well as Kiran Panchamgam (Oracle RGBU) and Anthony Smith (formerly from Oracle RGBU) 

Pricing has been a field that has seen a lot of exciting developments in the recent years. A particular area of pricing that has recently emerged is promotion pricing. In many important settings, promotions are a key instrument for driving sales and profits. Important examples include promotions in grocery retail among others. The Promotion Optimization Problem (POP) is a challenging problem as the retailer needs to decide which products to promote, what is the depth of price discounts, when to schedule the promotions and how to promote the product.

In this talk we will discuss how analytics can have a key impact on promotion pricing. This presentation will reflect our ongoing collaboration over the past few years with Oracle RGBU. We will describe the journey we took with them on introducing analytics tools for promotion planning in the grocery industry. We will describe optimization models we have built in order to determine which products to promote and when, as well as which vehicles to use to promote each product (e.g. flyer versus radio announcements among others) and how deeply to promote each product.

An important consumer behavior we will incorporate and which is a direct consequence of promotions in grocery retail is that consumers stockpile the products on promotion and then experience promotion fatigue after the promotion ends. Therefore, as a first step, we study general classes of demand functions that capture this effect and can be directly estimated from data. Using these demand functions, we model and study the promotion planning problem through an optimization formulation. Unfortunately, the underlying formulation even for a single product is NP-hard and highly nonlinear. We will first propose a linear approximation and show how to solve the problem efficiently as a linear programming (LP) problem. We will illustrate how this approximation idea has applications in many areas beyond pricing. We will discuss analytical bounds on the accuracy of this LP approximation relative to exact problem solution. We will also consider a graphical representation of the problem which will allow us to employ a Dynamic Programming (DP) solution approach as an alternative. We will discuss the tradeoffs between the two approaches (LP vs DP). Furthermore, we will incorporate in the optimization formulations, apart from the pricing aspect, how to decide which vehicle to use each time in order to promote which product. This further complicates the problem. We will introduce greedy and integer optimization ideas in order to solve the vehicle selection problem in a tractable way. These methods are computationally efficient and hence are easy to use in practice. We will also discuss some performance guarantees for these methods.

Finally, we will illustrate how our approach generalizes to consider multiple products within a category that are substitutes and/or complementary. We will discuss the tradeoffs when there are cross product effects.

Together with our industry collaborators from Oracle Retail, our framework allows us to develop a tool which can help supermarket managers to better understand promotions by testing various strategies and business constraints. We show that the formulation we propose solves fast in practice using actual data from a grocery retailer and that the accuracy is high. We calibrate our models using actual data and determine that they can improve profits by 3% just by optimizing the promotion schedule and up to 5% by slightly modifying some business requirements.


 Georgia Perakis is the William F. Pounds Professor at the Sloan School of Management at MIT since 1998. She received an M.S. degree and a PhD in Applied Mathematics from Brown University and a BA from the University of Athens in Greece.  

 Perakis' research studies the role of operations in many areas such as pricing, supply chain management, energy and transportation applications among others. She has widely published in journals such as Operations Research, Management Science, POM, Mathematics of Operations Research and Mathematical Programming among others. She has received the CAREER award from the National Science Foundation and subsequently, the PECASE award from the office of the President on Science and Technology awarded to the 50 top scientists and engineers in the nation. In 2007 she received an honorable mention in the TSL Best Paper Award, she also received the second prize in 2011, the first prize in 2012 and in 2014 in the Best Paper competition of the Informs Service Science Section for some of her papers. In 2015, her work on promotions received the Best Application of Theory Award from NEDSI (Northeast Decision Sciences Institute) Conference. She also received the Graduate Student Council Teaching Award as well as the Jamieson Prize for excellence in teaching and the Samuel M. Seegal prize for “inspiring students to pursue and achieve excellence”. Perakis was the recipient of the Sloan Career Development Chair and subsequently of the J. Spencer Standish Career Development Chair. In 2009, Perakis received the William F. Pounds chair that she currently holds. Her work on promotion pricing work was a Finalist at the Practice Award of the Revenue Management and Pricing Section of INFORMS in 2015. Perakis has passion supervising her students and builds lifelong relationships with them. So far she has graduated seventeen PhD and thirty Masters students.


 Perakis has served as the co-director from the MIT Sloan School side for the Leaders for Global Operations (former LFM) program. She is currently the group head of the Operations Management Group at MIT Sloan. She serves as an Associate Editor for the journals Management Science, Operations Research and Naval Logistics Research and a senior editor for POM. Perakis has served as a member of the INFORMS Council. She has also served as the chair of the Pricing and Revenue Management Section of INFORMS and in 2009-2010 as the VP for Meetings of the MSOM Society of INFORMS. She has co-organized the MSOM 2009 conference and served in the organizing committee of the 2010 MSOM conference. She has also been the co-chair and co-organizer of the Annual Conference of the INFORMS Section on Pricing and Revenue Management for several years and the chair of the cluster on the same topics for the annual INFORMS and ISMP conferences for several years.

INFORMS 2015 TutORials volume

IMG_2921I found the printed version of the INFORMS 2015 TutORials book in my mailbox today! (See picture on the left.) I co-edited it with Dionne Aleman of the University of Toronto. It is entitled "The Operations Research Revolution". We have some amazing tutorials lined up by experts in our field such as Laura McLay, Guzin Bayraksan, David Goldsman, Erick Delage, Dan Iancu, Art Chaowalitwongse, Josh Taylor, Meinolf Sellmann and more, and I can't wait to hear their presentations at the annual meeting. The papers will be available for free to INFORMS members. The link isn't up yet but I'll make sure to post an update as soon as I have more information. Dionne and I are looking forward to an informative set of presentations on cutting-edge O.R. topics in just a few weeks in Philadelphia!

#Analytics15 conference: Managing Risk track

Last week I attended the Analytics conference in Huntington Beach, CA. As a member of the organizing committee, my responsibilities were to co-organize the Early Career Connection program with Michelle Opp of SAS and help organize the new "Managing Risk" track alongside several other sub-committee members led by Freeman Marvin of Innovative Decisions. (I also presented a poster of my research with my PhD student Tengjiao Xiao, about our analysis of "inefficient" health plans on public health exchanges.) This post is on the new Managing Risk track. The purpose of the track is to provide a "home" for talks that do not fit neatly in more traditional categories such as supply chain or marketing yet capture important dimensions in risk management of benefit to analytics practitioners.

The track had five talks: Kenneth Fletcher of the Transportation Security Administration discussed how enterprise risk management can be used to identify threats to priorities and goals, Angela Fontes of NORC and the University of Chicago described a comprehensive model of consumer risk tolerance, Mike Dziecichowicz of Ernst & Young provided an overview of retail credit loss forecasting that I hope to convince him to present again at Lehigh to my financial optimization students in the spring (Mike is a former student of mine and it was great to hear about some of the projects he's been working on since Lehigh), Milind Tambe of USC delivered a fascinating presentation on the emerging science of security games and finally Sam Savage ended the day with a talk on how to apply risk model analysis to utility business decision-making.

The talk farthest from my own background was Tambe's and it is all to his credit that he was able to captivate a large audience of conference attendees with disparate interests and backgrounds with his examples of security games drawn from research projects from security at LAX (Los Angeles Airport) to wildlife protection in Africa. You can learn more about his research here.

Conference attendees can download presentation slides online (at least those from the speakers who share them) through the INFORMS Connect service. Unfortunately, no presentation from the Managing Risk track is currently available, but attendees can browse through 68 very informative presentations from other tracks.

Tom Davenport in praise of "light quants"

Here is an article I want my students to read in the fall for my "Operations Research Models & Applications" course. It is called "In praise of light quants and analytical translators" and was authored by Tom Davenport, the President's Distinguished Professor of Information Technology and Management at Babson Analytics and an independent senior advisor at Deloitte Analytics. Davenport deserves much of the credit for bringing the concept of "analytics" to the attention of the general public through his books Competing on Analytics, Keeping Up With the Quants, and more. (Full list here.)

In the article linked above, Davenport argues that we need more "light quants" and "analytical translators": people who understand analytics and can communicate effectively those analytical insights to, as well as interact with, the business side of the company. Davenport makes the distinction between "light quants" and "heavy quants", but it seems to me that more broadly the distinction is between BS in operations research related fields and PhDs. MS holders can fall in either category, depending on their training, expertise and job experience, and also obviously their communication skills. 

To me, the perfect example of a light quant as described by Davenport and for the US market would be a graduate with a BS in Industrial and Systems Engineering from Lehigh and (a nice plus) a MS also from our department, for instance in Management Science and Engineering, or a dual BS in Integrated Business and Engineering (our four-year honors program) with a concentration in ISE and BS in ISE. Those students receive advanced operations research training suitable for their level but also have exceptional communication skills that make them uniquely capable of playing the role of "light quant" described by Davenport. (They will also be the audience of my course in the fall; hence the choice of article.)

I think many Lehigh ISE undergraduates may not realize their skills are in high demand, because they have spent most of their time in the classroom and at one, perhaps two, internships, where they may or may not have made full use of their operations research and analytical skills. It's important for them to realize they fill a critical need in the market. The excellence of the training they have received was documented by the position of finalist (1 out of 3) that the department has received four years in a row in the international INFORMS UPS George D. Smith Prize Competition, which rewards the programs best preparing O.R. students to become future practitioners.

Data-driven insurance

The Economist had an interesting article in its March 14 issue on the way data and technology are starting to up-end the insurance business ("Risk and reward"). The author gives the example of drivers buying car insurance from Progressive in the U.S., who can install a device in their car that monitors their driving and "adjusts the rate they pay accordingly". He/she argues that "data mining and monitoring not only allow insurers to price policies more accurately, but also enable them to modify customers' behaviour."

Another interesting excerpt: "[a client's] database singles out unsafe drivers so that agents can visit them at home in an effort to persuade them to change their habits. Kaiser Permanente, an American health insurer, does something similar with its most at-risk policyholders." The article provides some evidence suggesting that such approaches do succeed in promoting 'good' behavior, although it may be that the people who sign up are naturally inclined to adjust their ways. ("Demand for Progressive's prying car insurance... now accounts for over $2 billion in premiums.")

Risk pools will become smaller, which should result in lower premiums in competitive environments where insurance companies will vie for the lowest risks. Left unspoken in the article but a natural consequence of those data-driven developments is the fact that enrollees could in effect decide how much they want to pay for car insurance (health insurance is trickier since people have less control on some of their health conditions) and adjust their behavior accordingly. They may also create their own models of insurance premiums to figure out how to pay as small a premium a possible while taking into account their driving style and what the Progressive-type device monitors.

The article also touches upon the possibility of new entrants with extensive experience in data analytics but no experience in insurance, such as Amazon or Google. "In response, insurers are busy trying to make themselves more like tech firms... Partnerships with non-insurers are another way for conventional insurers to smarten up their act." It looks like the insurance business is headed for a shake-up. 

OR: A Catalyst for Engineering Grand Challenges

I finally found the time to read Operations Research: A Catalyst for Engineering Grand Challenges, which has been making the rounds of the OR-related departments in the country. The report's authors advocate using the National Academy of Engineering's Engineering Grand Challenges "as a source of inspiration for the OR community", and in particular they recommend: "(1) an NSF announcement of "Grand Challenge Analytics" as a major EFRI topic, and (2) an NSF sponsored insitute for multidisciplinary OR and engineering." 

The authors make inspiring predictions about the potential OR could have in such fields ("these initiatives are likely to unleash a vast array of methodologies onto the engineering Grand Challenges of today"). This is an important report that makes a strong contribution toward turning OR and analytics into required staples of the engineering arsenal, and it provides important statistics in the introduction drawn from IBM's vision to build a "smarter planet." I do wish that the use of OR to transform (solve?) the engineering grand challenges resulted from a pull from the engineering community rather than a push from the OR folks - in other words, I would have preferred if the report also had made a case from the engineering community that not only do they need the current OR techniques available today but the problems they face, for instance with big data in geosciences, requires the design and analysis of cutting-edge algorithms. What are the collaborations between OR and engineering faculty members happening today (I'm sure there are plenty)? Which OR-trained engineering faculty member can talk about the need to develop new OR techniques because the ones he knows have reached the limits of their usefulness? I'm sure there are plenty too.

The report does make clear that "[its] goal is to view these challenges as an opportunity for the OR community to play the role of a catalyst - utilizing our ideas and tools to address some of the more pressing technological challenges facing humanity today. Because of this emphasis, the report will NOT [emphasis theirs] focus on challenges for OR; instead the focus is on "Catalysis"." It's good to see that the OR community is becoming better at marketing itself and choosing good buzzwords ("catalyst" definitely beats "science of better" - who doesn't want to be a catalyst?), and it's definitely getting as much mileage as it can from the "catalyst" word and its variants.

Now, let me be clear that leveraging the NAE Grand Challenges is a fantastic idea. The report makes valuable suggestions. It is a good read. I also do believe that, while there is a lot of talk about OR serving as a catalyst, OR runs the risk of being seen as a tool rather than an evolving field. You know how when the conversation veers to MOOCs, you'll always find a humanities professor to assert that MOOCs are very good for vocational training (insert sneer/smirk there) but those courses won't teach students the more fundamental skill of how to think? The report, which heavily portrays OR as a needed tool to solve engineering problems, made me think of this dichotomy. I think it cheapens OR to only consider the "tool" angle.

I looked at the section about "OR for Health Care" in most detail given my research in healthcare finance, and while the summary is fair given the space the authors had, I was a bit disappointed by what I read. For instance, the report (quickly) mentions hospital planning/scheduling, operating room scheduling, bed allocation, nurse staffing, but there is no mention whatsoever of the features that make those problems challenging every time so that  they will not fit a cookie-cutter mold. Admittedly, there is also a lot of low-hanging fruit in healthcare resource management and revenue optimization, so that undergraduates in their capstone project can indeed make enormous and quick contributions by applying OR techniques on short-term projects. Fair enough. But it would have been helpful to emphasize the scope of possible improvements OR can make in engineering fields, so that "OR as tool" would be perfect for undergraduate and Master's students doing a semester project or joining companies in entry-level positions, and "OR as way of thinking" (for lack of a better term) would fit the competencies of PhD students and PhD graduates.

Overall, this is an important report that will hopefully stimulate a healthy discussion, not only among OR professionals but also in the broader engineering community.

HBR on Lots of Little Data

ThumbnailThe December 2013 issue of Harvard Business Review had an article on Focused Leaders by Daniel Goleman and one on Analytics 3.0 by Thomas Davenport, both household names at the top of their fields, yet the article that stole the show, in my opinion, was "You may not need big data after all" by Jeanne Ross, Cynthia Beath and Anne Quaadgrass. The article is subtitled "Learn how lots of little data can inform everyday decision-making." I liked the article because there is so much talk of Big Data nowadays, with anyone wanting to show or fake knowledge in analytics throwing the term around, and it is good to take a step back and think: what do we like so much about Big Data? What can we get done with (lots of) little data? What are good steps to take in creating a data-driven or evidence-based decision-making culture? 

Here is an excerpt: "The biggest reason that investments in big data fail to pay off... is that most companies don't do a good job with the information they already have. They don't know how to manage it, analyze it in ways that enhance their understanding, and then make changes in response to new insights." The authors advocate a culture of evidence-based decision-making and provide the example of Seven-Eleven Japan, whose success can be attributed to empowering salesclerks to act on a lot of little data, or "betting your business success on the ability of good people to use good data to make good decisions."

I also found particularly illuminating the example of Aetna, whose head of operations (later CEO) found out that "all the divisional heads could show him a spreadsheet with performance data indicating that their divisions had been profitable the previous year - even though Aetna as a whole had recorded a loss of almost $300 million."

The authors recommend the use of scorecards (with the right metric) "to clarify individual accountability and provide consistent feedback", with the example of PepsiAmericas's shift from P&L to recurring monthly revenue (RMR). They also describe the companies most likely to benefit from big data as (a) companies with a tradition of fact-based decision-making, such as UPS, (b) engineering and research companies such as ExxonMobil, and (c) the best web-native companies - and there the examples are plentiful, such as Google, Amazon, Netflix and eBay.

The article ends with a sidebar on "Do you have an evidence-based culture?" with 11 questions such as: Do you rely on a single source for performance data? and: Do you create and revise business rules on the basis of business analytics? Very interesting read, highly recommended.