So Yeon Chun

About Me

So Yeon Chun

Associate Professor of Technology and Operations Management


Research Interests

My research interests are data-driven revenue and management (pricing and forecasting), and operations and marketing interface utilizing applied statistics and stochastic optimization.

  • Operations with loyalty programs and point currency
  • Data-driven revenue management and dynamic pricing with applications to transportation, retail, and hospitality industries
  • Consumer choice behavior and substitution patterns
  • Behavioral experiments and field studies
  • Statistics and stochastic optimization algorithms with large-scale datasets

More details can be found on the <Research> page.




Ph.D., Operations Research

Georgia Institute of Technology, Atlanta, Georgia
School of Industrial and Systems Engineering

  • Dissertation:
    “Hybrid is Good: Stochastic Optimization and Applied Statistics for OR”

M.S., Applied Statistics

Georgia Institute of Technology, Atlanta, Georgia

B.S., Industrial Engineering

Seoul National University, Seoul, Korea
Department of Industrial Engineering

  • Admitted with highest honors
  • Graduated with highest honors (summa cum laude)

Professional Experience

Operations Research Analyst

JDA Software Group, Inc., Atlanta, Georgia

  • Analyzed hotel and cruise line booking and sales data
  • Developed price elasticity estimation and demand forecasting systems for a price optimization solution for hospitality clients

Summer Researcher

IBM Thomas J. Watson Research Center, Yorktown, New York

  • Analyzed traffic time series data
  • Developed traffic decision support system (DSS) architecture and optimizer (control) algorithms in real-time traffic management centers

Summer Researcher

IBM Thomas J. Watson Research Center, Hawthorne, New York

  • Analyzed retail sales transaction data for promotion planning
  • Developed spatial-temporal data mining statistics techniques for real-time anomaly detection
    for road user (congestion) charging
  • Developed optimization models for a back-office enforcement process of vehicle records

Professional Service


INFORMS Revenue Management. and Pricing Section, 2015-2016

Revenue Management and Pricing Track Organizer

POMS Annual Conference, 2016 (13 sessions with 54 talks)

Revenue Management and Pricing Cluster Organizer

INFORMS Annual Conference, 2015 (48 Sessions with 185 talks)

Conference Session Chair

POMS Annual Conference, 2015
INFORMS Annual Conference, 2011-Present

Paper Competition Judge

MSOM Student Paper Competition, 2014-Present

Invited Journal Reviewer

Management Science, Operations Research, Manufacturing and Service Operations Management, Production and Operations Management, Transportation Science, IIE Transactions, Naval Research Logistics


INFORMS Future Academician Doctoral Colloquium

Austin, Texas, 2010

Thank a Teacher Recipient

Georgia Institute of Technology, 2010 (Honored at the Dean Griffin Day Luncheon)

Award for Top Performance on Ph.D. Comprehensive Exam

Georgia Institute of Technology, 2007 (Perfect Score)

Scholarship Recipient

Dongbu Group, 2006, 2007

Presidential Honor Recipient

Seoul National University, 2004 (Ranked First in the
Industrial Engineering Department, Class of 2004)

Best Student Scholarship

Seoul National University, 2001, 2002, 2003, 2004

Math Olympiad

Seoul, Korea, 1993 (Honorable Mention)



I enjoy building (stochastic) mathematical models and testing them with real data to examine the chaotic reality around us.

This endeavor is inherently interdisciplinary, requiring application-specific knowledge and simultaneous expertise in two fields: statistical analysis and optimization under uncertainty.

My research takes a multi-method approach that combines analytical and empirical methods and addresses challenging problems that lie at the intersections of Operations, Marketing, Economics, and Risk Management.


Refereed Publications

Dynamic Pricing with Point Redemption

(with Hakjin Chung and HyunSoo Ahn)
Forthcoming at Manufacturing and Service Operations Management


Dynamic Pricing with Point Redemption Program

Many sellers allow consumers to pay with reward points instead of cash or credit card. While the revenue implications of cash and credit card purchases are transparent, the implications of reward sales are not trivial, especially when a firm that issues points is not a seller. In this case, a seller receives a monetary compensation or reimbursement from the point issuer when a consumer purchases the good by redeeming points. In this paper, we examine how reward sales influence a seller’s pricing and inventory decisions. In particular, we consider a consumer who can choose to pay with cash or points based on her attributes — reservation price, point balance, and the perceived value of a point. Then, we incorporate this consumer choice model into a dynamic pricing model where a seller earns revenues from both cash and reward sales.

How Loyalty programs are Saving Airlines

(with Evert Boer)
Harvard Business Review (online), (2021)


Refocusing Loyalty Programs in the Era of Big Data: A Societal Lens Paradigm

(with Valeria Stourm, Scott A. Neslin, Eric T. Bradlow, Els Breugelmans, Pedro Gardete, P. K. Kannan, Praveen Kopalle, Young-Hoon Park, David Restrepo Amariles, Raphael Thomadsen, Yuping Liu-Thompkins, and Rajkumar Venkatesan)
Marketing Letters, Special Issue: 11th Triennial Invitational Choice Symposium, 31 (2020): 405–418


Risk-based Loan Pricing: Portfolio Optimization Approach with Marginal Risk Contribution

(with Miguel Lejeune)
Management Science, 66.8 (2020): 3735-3753


Risk-based Loan Pricing: Portfolio Optimization Approach with Marginal Risk Contribution

We consider a lender (bank) who determines the optimal loan price (interest rates) to offer to prospective borrowers under uncertain risk and borrower response. A borrower may or may not accept the loan at the price offered, and in the presence of default risk, both the principal loaned and the interest income become uncertain. We present a risk-based loan pricing optimization framework, which explicitly takes into account marginal risk contribution, portfolio risk, and borrower’s acceptance probability. Marginal risk assesses the amount a prospective loan would contribute to the bank’s loan portfolio risk by capturing the interrelationship between a prospective loan and the existing loans in the portfolio and is evaluated with respect to the Value-at-Risk and Conditional-Value-at-Risk risk measures.

Loyalty Program Liabilities and Point Values

(with Dan Iancu and Nikolaos Trichakis)
Manufacturing and Service Operations Management, 22.2 (2019): 223-428

Loyalty Program Liabilities and Point Values

Loyalty programs (LP) introduce a new currency, the points, through which customers transact with firms. Such points represent a promise for future service, and their monetary value thus counts as a liability on the issuing firms’ balance sheets. Consequently, adjusting the value of points has a first order effect on profitability and performance, and emerges as a core operating decision. We study the problem of optimally setting the points’ value in view of their associated liabilities.

Strategic Consumers, Revenue Management, and the Design of Loyalty Programs

(with Anton Ovchinnikov)
Management Science, 65.9 (2019): 3949-4450

Strategic Consumers, Revenue Management, and the Design of Loyalty Programs

We study the interaction between the design of a premium-status loyalty program, revenue management, and strategic consumer behavior. Specifically, we consider a contemporaneous change where firms across several industries switch their loyalty programs from quantity-based to spending-based designs. This change has been met with fierce opposition from the media and consumers. We present a model for strategic, forward- looking, and status-seeking consumers’ decisions on how much to purchase over a certain time period, and endogenously derive strategic consumer demand as a function of the firm’s prices and loyalty program decisions. We then incorporate such demand into the firm’s pricing and loyalty program design problem and compare different loyalty program designs. We identify the conditions under which, by coordinating pricing and loyalty functions, the firm can benefit from strategic consumer behavior, and show that by switching to a spending-based design, the firm can benefit from strategic behavior even more, under broader conditions, and in a Pareto-improving way. We also analyze combined designs, which utilize quantity and/or spending requirements, and provide additional insights on how the firm can better manage the transition toward spending-based designs, possibly minimizing negative consumer reactions.

Modified Distribution-free Goodness of Fit Test Statistic

(with Alex Shapiro and Michael Browne)
Psychometrika 83.1 (2018): 48-66

Modified Distribution-free Goodness-of-Fit Test Statistic

Covariance structure analysis and its structural equation modeling extensions have become one of the most widely used methodologies in social sciences such as psychology, education, and economics. An important issue in such analysis is to assess the goodness-of-fit of a model under analysis. One of the most popular test statistics is the asymptotically distribution free (ADF) test statistic introduced by Browne in 1984. The ADF statistic can be used to test models without any specific distribution assumption (e.g., multivariate normal distribution) of the observed data. Despite its advantage, it has been shown in various empirical studies that unless sample sizes are extremely large, this ADF statistic can perform very poorly in practice. In this paper, we provide a theoretical explanation for this phenomenon and further propose a modified test statistic that improves the performance in samples of realistic size. The proposed statistic deals with the possible ill conditioning of the involved large scale covariance matrices.

When Friends Become Competitors: The Design of Resource Exchange Alliances

(with Anton Kleywegt and Alex Shapiro)
Management Science, 63.7 (2016): 2127-2145 (Previous title: “Resource Exchange Seller Alliances”)

Resource Exchange Seller Alliances: Two-stage stochastic model with equilibrium constraints

Many carriers, such as airlines and ocean carriers, collaborate through the formation of alliances. The detailed alliance design is clearly important for both the stability of the alliance and profitability of the alliance members. This work is motivated by a real-life liner shipping “resource exchange alliance” agreement design. We provide an economic motivation for interest in resource exchange alliances and propose a model and method to design a resource exchange alliance. The model takes into account how the alliance members compete after a resource exchange by selling substitutable products and thus enables us to obtain insight into the effect of capacity and the intensity of competition on the extent to which an alliance can provide greater prot than when in the setting without an alliance. The problem of determining the optimal amounts of resources to exchange is formulated as a stochastic mathematical program with equilibrium constraints (SMPECs). We show how to determine whether there exists a unique equilibrium after resource exchange, how to compute the equilibrium, and how to compute the optimal resource exchange. SMPEC problem, which is generally very difficult to solve, is well-posed in the paper, and robust results can be obtained with a reasonable amount of computational effort.

Conditional Value-at-Risk and Average Value-at-Risk: Estimation and Asymptotics

(with Alex Shapiro and Stan Uryasev)
Operations Research, 60.4 (2012): 739-756

Conditional Value-at-Risk and Average Value-at-Risk : Estimation and Asymptotics

Value-at-Risk and Average Value-at-Risk (Conditional Value-at-Risk, Expected Shortfall) are widely used measures of financial risk. To estimate accurate risk measures taking into account the specific economic conditions, we considered two estimation procedures for each conditional risk measure; one is direct and the other is based on residual analysis of the standard least squares method. Large sample statistical inference of obtained estimators are derived and compared. In addition, finite sample properties of both estimators are investigated and compared in an extensive Monte Carlo study. Empirical results on real-data (different financial asset classes) are also provided to illustrate the performance of proposed estimators.

Construction of Covariance Matrices with a Specified Discrepancy Function Minimizer, with Application to Factor Analysis

(with Alex Shapiro)
SIAM Journal on Matrix Analysis and Applications, 31.4 (2010): 1570-1583

Construction of Covariance Matrices with a Specified Discrepancy Function Minimizer, with Application to Factor Analysis

Abstract: Covariance structure analysis, and its structural equation modeling extensions, had become one of the widely used methodologies in the social sciences such as psychology, education, economics, and sociology. An important issue in such analysis is to assess the goodness-of-fit of a considered model. For this purpose various test statistics were suggested and their behaviors have been studied in numerous publications. It is well understood that in real applications no model fits data exactly and at best a good model can be considered as an approximation of reality. This raises the question of properties of the respective test statistics for misspecified models. The main goal of this paper is to develop a numerical procedure for construction of covariance matrices such that for a given covariance structural model and a discrepancy function the corresponding minimizer of the discrepancy function has a specified value. Often construction of such matrices is a first step in Monte Carlo studies of statistical inferences of misspecified models. We analyze theoretical aspects of the problem and suggest a numerical procedure based on semidefinite programming techniques. As an example, we discuss in detail the factor analysis model.
Keywords: Model misspecification, covariance structure analysis, maximum likelihood, generalized least squares, discrepancy function, factor analysis, semi-definite programming.

Normal versus Noncentral Chi-square Asymptotics of Misspecified Models

(with Alex Shapiro)
Multivariate Behavioral Research, 44.6 (2009): 803-827

Normal versus Noncentral Chi-square Asymptotics of Misspecified Models

Abstract: The noncentral chi-square approximation of the distribution of the likelihood ratio (LR) test statistic is a critical part of the methodology in structural equation modeling. Recently, it was argued by some authors that in certain situations normal distributions may give a better approximation of the distribution of the LR test statistic. The main goal of this article is to evaluate the validity of employing these distributions in practice. Monte Carlo simulation results indicate that the noncentral chi-square distribution describes behavior of the LR test statistic well under small, moderate, and even severe misspecifications regardless of the sample size (as long as it is sufficiently large), whereas the normal distribution, with a bias correction, gives a slightly better approximation for extremely severe misspecifi- cations. However, neither the noncentral chi-square distribution nor the theoretical normal distributions give a reasonable approximation of the LR test statistics under extremely severe misspecifications. Of course, extremely misspecified models are not of much practical interest. We also use the Thurstone data (Thurstone & Thurstone, 1941) from a classic study of mental ability for our illustration.
Keywords: Model misspecification; covariance structure analysis; maximum likelihood; generalized least squares; discrepancy function; noncentral chi-square distribution; normal distribution; factor analysis

Papers under Review/Revision/Working Papers

Paying with Money or Paying with Points:
How the Stability, Numerosity and Favorability of the Exchange Rate Influence Loyalty Point Redemption

(with Rebecca Hamilton)
Under review

Paying with Money or Paying with Points:
How the Stability, Numerosity and Favorability of the Exchange Rate
Influence Loyalty Point Redemption

Many loyalty programs award members points for their purchases, which can be redeemed later for additional goods and services. In this research, we examine how the exchange rate between money and points that is set by loyalty program managers affects consumers’ choices to redeem points for a purchase by analyzing data from an airline loyalty program and conducting four experimental studies. Specifically, we investigate how the numerosity, stability, and favorability of the exchange rate affect the consumers’ payment choices (paying with money or points).

Monetization of Loyalty Digital Currency: Mental Accounting anad Income Effects on Spending

(with Freddy Lim and Ville Satopaa)
Working paper

Seasonality for New Product Diffusions

(with Evren Ozkaya and Pinar Keskinocak)
Working paper

Seasonality model for new product diffusions

In forecasting new product diffusions with short life cycles, seasonality plays a significant role. Seasonal data series have not been widely used in the diffusion modeling context as the majority of the studies focus on macro-level diffusion models that use annual data. In this paper, we analyze the impact of seasonality on new product diffusions with short life cycles and within-the-quarter seasonal demand patterns. We propose a method to estimate seasonality of monthly data series with diffusion trend. We show, under both simulated and real data, that the proposed approach can significantly improve seasonality factor estimates and therefore forecast accuracy, especially when data series is short with (nonlinear) diffusion trend and high random error variance.

In Preparation

Gamification of Loyalty Program and Consumer Behavior

Intrinsic vs. Extrinsic Value of Loyalty Points, Evidence from Hotel Reward Program Data

(with Rebecca Hamilton and Chekitan Dev)

Empirical Analysis of Loyalty Point Redemption Behavior for Experience Products

(with Felipe Walker)

Internet Service Pricing and Data Throttling: Evidence from Internet Service Sales Data

Dynamic Discrete Choice Modeling for the Optimal Stopping Problem, with Application to Auto Industry Sales and Promotion Data

(with Anton Kleywegt)

Dynamic Discrete Choice Modeling for the Optimal Stopping Problem

Sales promotion plays an important role in revenue generation in business. Understanding consumer behavior on promotion are crucial. Traditional discrete choice models assume that decision makers choose one alternative at a point in time, without the option to wait, collect more data, and revisit the decision later. We develop dynamic discrete choice models in which, at each decision time, decision makers have the option to postpone a decision until later. Alternatives and attributes may change over time, and decision makers attempt to forecast attribute values at the next decision time. Our model performs better than the default static discrete choice model and therefore our models appear to hold promise for building better business strategies.


Demand Analysis and Price Elasticity of Demand Estimation System for Revenue Management and Pricing

(with Ronald Menich)
Submitted for filing

Targeted Enforcement for Road User Charging

(with Todd Appel, Duncan Ashby, Milind Naphade, Richard Nash, Anshul Sheopuri, Anders Thorndivist, and Martin Vuyk)
U.S. Patent Application No. 12,752,578

Intelligent Decision Support System Optimizer for a Real-Time Command Center

(with Laura Wynter)
U.S. Patent No. 8,458,113. 4 Jun. 2013

Anomaly Detection for Road User Charging Systems

(with Todd Appel, Duncan Ashby, Milind Naphade, Richard Nash, Anshul Sheopuri, Anders Thorndivist, and Martin Vuyk)
U.S. Patent No. 9,261,375. 16 Feb. 2016












Teaching Experience

Foundations of Operations Management B (2020-Present)


ART: Analytics for Retail and Travel (2019-Present)

MBA Elective, INSEAD, France/Singapore

Instructor, OPIM 173 Business Statistics (2013-2019)

Georgetown University, Washington D.C.
McDonough School of Business

Instructor, ISYE 2028 Basic Statistical Methods (Spring 2010)

Georgia Institute of Technology, Atlanta, Georgia
School of Industrial and Systems Engineering

Teaching Assistant, ISYE 6402 Time Series Analysis (Spring 2007)

Georgia Institute of Technology, Atlanta, GA
School of Industrial and Systems Engineering

Teaching Assistant, ISYE 3770 Probability and Statistics (Fall 2006, Spring 2008, Fall 2009)

Georgia Institute of Technology, Atlanta, GA
School of Industrial and Systems Engineering

Highlights from Course Survey

  • “Professor Chun is one of the best and most inspirational professors I have had yet at Georgetown, and she will make you fall in love with statistics. Everything is clear cut and explained and is taught in an amazing fashion.”
  • “Professor Chun was probably the ideal professor to have – helpful, extremely organized, and very kind. She seemed to be genuinely interested in teaching students and if every professor was like her, Georgetown would be #1 ranking.”
  • “Professor Chun is an excellent teacher and is always willing to help and explain things that I have trouble with. Additionally, she is always very happy and brings a lot of energy to the class! “
  • “I believe Professor Chun really made the difference in making statistics an interesting and understandable subject. Her dedication and resourcefulness are unparalleled to that of any other professors I have had. She is extremely helpful and offered guidance every step along the way.”
  • “Prof. Chun was a really great professor! Her enthusiasm about business stats was evident in each class, which made her lectures interesting and engaging. She is sweet and incredibly approachable, and cares so much about each student’s success that she really goes out of her way to make sure people receive help when they need it. The exams in this course were challenging, which really required me to study hard, but I now have a solid foundation in stats that has already helped me in my other classes. I am so glad I took this course with Prof. Chun and will be sure to recommend her to other students!”
  • “Thanks for an enjoyable semester. I really liked the way the course was taught because you focused more on practical topics that we will one day use in our careers, rather than just dwelling on proofs and derivations.”
  • “I really really admire the professor! I know she was genuinely interested in the betterment of all of her
    students. Since the first day of classes, we were challenged and nicely encouraged to do our best. Right from the start, I knew that the professor had best intentions at heart. Honestly, couldn’t have asked for a better professor!”
  • “I had a lot of fun in this class. You had a very organized set up and your notes were very easy to follow. The tests were just right, challenging enough so students can differentiate themselves but not so difficult that the course becomes discouraging. You were always very helpful and straightforward. I really enjoyed this class and am thankful to have had you as my professor.”
  • “Great and effective instructor! I learned a lot from this course. The lectures were extremely well organized and the tests were fair. Hope to take another class with you in the future!”
  • “Professor Chun is the best teacher I’ve had in the MSB. She genuinely cares about her students, challenges us, and goes above and beyond in and out of the classroom for us.”
  • “Professor Chun was amazing! She is by far my favorite professor that I’ve had at Georgetown. She is
    helpful and engaging during class and during office hours. She makes the material fun to learn and she is
    very helpful during office hours!”
  • “Out of my two semesters at Georgetown, Professor Chun was by far the most enthusiastic about her course material and teaching her students.”
  • “Excellent instructor, she is really willing to help the student progress throughout the semester.”
  • “Thanks for being a great teacher. I am hoping that we will stay in touch for the future. You definitely know what you are talking about and I know that I could use your statistical guidance and advice. Thanks for a great semester!”
  • “Excellent job connecting with students. Status check helped focus attention on most important topics. Great job!”


+33 1 60 72 40 00

Fontainebleau, France

How Can I Help You?