Category Archives: Presentations

My INFORMS 2017 Presentations

Sunday, 8:00-9:30am, SA37, Linking Delay Announcements, Abandonment, and Staffing: A Behavioral Perspective, Room 352B

Monday, 8:00-9:30am, MA04, Mind The Gap: Coordinating Energy Efficiency and Demand Response, Room 320A

Tuesday, 4:35-6:05pm, TE04, Utility Ownership of Decentralized Combined Heat and Power, Room 320A

The two energy papers of my dissertation were invited. The call center paper was scheduled for presentation with the other finalists in the in the Service Section Best Student Paper Competition.

Talks I Attended at MSOM 2017

Documented mostly for my future reference.

-Peak Load Energy Management by Direct Load Control Contracts, by Ali Fattahi, Sriram Dasu, Reza Ahmadi
-Is electricity storage green? A study on the commercial sector, by Yangfang Zhou (Helen)
-Promotion Planning of Network Goods, by Saed Alizamir, Ningyuan Chen, Vahideh Manshadi

-Seeking to Belong, by Bradley Staats, Paul Green, Francesca Gino
-Familiarity in Creative Teams: The Effect of Task Nature, by Murat Unal, Karthik Ramachandran, Necati Tereyagoglu
-Designing Sustainable Products under Co-Production Technology, by Yen-Ting Lin, Shouqiang Wang, Haoying Sun

-That’s Not Fair – Tariff Structures for Electricity Markets with Rooftop Solar, by Siddharth Prakash Singh, Alan Scheller-Wolf
-Mind the Gap: Coordinating Energy Efficiency and Demand Response, by Eric Webb, Owen Wu, Kyle Cattani
-Using Transparency to Manage the Sourcing of Complex Non-routine Litigation, by Jacob Chestnut, Damian Bell

-Environmentally Friendly Contract in a Supply Chain: Stimulating Supplier’s Environmental Innovation for a Manufacturer under Emission Tax, by Kun Soo Park
-Multi-agent Mechanism Design without Money, by Santiago Balseiro, Huseyin Gurkan, Peng Sun
-Payment for Results: Funding Non-Profit Operations, by Sripad Devalkar, Milind Sohoni, Neha Sharma

-Designing Incentives for Startup Teams: Form and timing of Equity Contracting, by Evgeny Kagan, Stephen Leider, William Lovejoy
-Integrating Managerial Insight and Optimal Algorithms, by Blair Flicker, Elena Katok
-Modeling Newsvendor Behavior: A Prospect Theory Approach, by Bhavani Shanker Uppari, Sameer Hasija

-Last Place Aversion in Queue, by Ryan Buell, Michael Norton, Jay Chakraborty
-Learning Preferences and User Engagement Using Choice and Time Data, by Tauhid Zaman, Zhengli Wang
-Relative Performance Transparency: Effects on Sustainable Purchase and Consumption Behavior, by Ryan Buell, Shwetha Mariadassou, Yanchong Zheng

Talks I Attended at INFORMS 2016

Documented mostly for my future reference.

Sunday 8am, track SA36:
-An Empirical Investigation of Network Effects in Automobile Sales by Tianjun Feng, Fuqiuang Zhang, and Peiwen Yu
-The Operational Value of Social Media Information by Dennis Zhang, Antonio Moreno-Garcia, Ruomeng Cui, and Santiago Gallino
-When you work with a super man, will you also fly? An empirical study of the impact of coworkers on performance by Serguei Netessine and Fangyun Tang
-CEO overconfidence and inventory management by Fuqiang Zhang, Tianjun Feng, and Qing Zhang

Sunday Plenary: Cognitive Computing: From Breakthroughs in the lab to applications in the field by Guru Banavar of IBM Research

Sunday 11am, track SB09:
-Biomass Supply contract pricing and environmental policy analysis: an agent-based modeling approach by Shiyang Huang and Guiping Hu
-On the effectiveness of tax incentives to support biomass co-firing by Hadi Karimi and Sandra Eksioglu
-A game-theoretic model of biomass co-firing policies by Sandra Eksioglu and Armin Khademi
-Evaluation of a wind farm project by Metin Cakanyildirim

Sunday 1:30pm, track SC28:
-Dynamic optimization of multichannel advertising campaigns in an online advertising supply chain by Changseung Yoo, Anitesh Barua, and Genaro Gutierrez
-Variability in labor schedules: Effects on store performance and employee turnover by Hyun Seok Lee, Saravanan Kesavan, and Camelia Kuhnen

Sunday 1:30pm, track SC30
-Managerial Attention, Reminders, and the Energy Efficiency Gap by Enno Siemsen and Suvrat Dhanokar
-Does learning from inspections affect environmental performance? Evidence from unconventional oil and gas wells by Suresh Muthulingam and Vidya Mani

Sunday 4:30pm, track SD29:
-Valuing distributed energy resources in electricity system planning: locational benefits and economies of unit scale by Jesse Jenkins
-Combined heat and power production – valuing flexible operation in an uncertain environment by Chritoph Weber

Sunday 4:30pm, track SD28
-Robust Supply function equilibrium in renewable energy markets by Yuanzhang Xiao, Chaithanya Bandi, and Ermin Wei
-An analysis of demand response programs in the wholesale electricity market by Asligul Serasu Duran, Baris Ata, and Ozge Islegen

Monday 8am, track MA35:
-The use of technology to improve engagement through accountability by Gad Allon
-Innovations in teaching operations management at UCLA by Guillaume Roels
-Architecting new business models (in the classroom) by Karan Girotra
-Ideo: Human-centered service design – multimedia-enhanced teaching and learning by Ryan Buell

Monday Plenary: Public health preparedness: Answering (largely unanswerable) questions iwth operations research by Margaret Brandeau

Monday 1:30pm, track MC29:
-Operational response to climate change: Do profitable carbon abatement opportunities decrease over time? by Christian Blanco, Felipe Caro, and Charles Corbett
-Closing a supplier’s energy efficiency gap: The role of assessment assistance and procurement commitment by Quang Dang Nguyen, Karen Donohue, and Mili Mehrotra
-Mind the Gap: Coordinating Energy Efficiency and Demand Response by Eric Webb, Owen Wu, and Kyle Cattani

Tuesday 8am, track TA29:
-Energy efficiency contracting in supply chains under asymmetric bargaining power by Ali Shantia, Sam Aflaki, and Andrea Masini
-An analysis of time-based pricing in electricity supply chains by Asligul Serasu Duran, Baris Ata, and Ozge Islegen
-Investments in renewable and conventional energy: The role of operational flexibility by Kevin Shang, Gurhan Kok, and Safak Yucel
-Explaining the variation in progress in the US nuclear industry by Christian Blanco, Felipe Caro, and Charles Corbett

Tuesday 11am, track TB29:
-Green sourcing – the role of premium sharing and consulting services by Xi Chen
-Inducing prompt disclosure in the presence of evasive effort by Shouqiang Wang, Peng Sun, and Francis De Vericourt
-The adoption of smart home appliance form energy shifting by Wenbin Wang and Yannan Jin
-Incentives for joint product and process improvement under collective extended producer responsibility by Luyi Gui

Tuesday 1:30pm, track TC34:
-Do mandatory overtime laws improve quality? Staffing decisions and operational flexibility of nursing homes by Lauren Lu and Susan Lu
-Predicting Nurse Turnover And Its Impact on Staffing Decisions by Eric Webb and Kurt Bretthauer
-Hospital readmissions reduction program: An economic and operational analysis by Dennis Zhang

Tuesday Keynote: Optimizing the future – supply chain at Amazon by Jason Murray

Tuesday 4:30pm, track TD37:
-The impact of carbon pricing on improving supply chain energy efficiency by Quang Dang Nguyen, Karen Donohue, and Mili Mehrotra
-Quantifying the impact of intermittent renewable generation on German electricity market by Shadi Goodarzi, Derek Bunn, and Syed Basher
-Designing hydro supply chains for water, food, energy, and flood nexus by Kwon Gi Mun, Raza Ali Rafique, and Yao Zhao
-Reversing the death spiral: A new business model for utility firms under social network effects by Safak Yucel, Gurhan Kok, and Kevin Shang

Wednesday Keynote: The goals of analysis are understanding, decisions, and influencing policy by Gerald Brown

Wednesday 11am, track WB31:
-Ethics, Bounded Rationality, and IP sharing in knowledge-based outsourcing by Manu Goyal and Krishnan Anand
-Accurate estimation of retail store traffic from people counters to achieve better conversion by Anup Hanamant
-Mitigating digital discrimination with reviews in the sharing economy: Field evidence from Airbnb by Dennis Zhang, Jun Li, and Ruomeng Cui
-Impact of operational risks in financial organizations by Yuqian Xu, Fangyun Tan, and Sergeuei Netessine

Wednesday 12:45pm, track WC31:
-Rational abandonment from observable priority queues by Philipp Afeche and Vahid Sarhangian
-Design of discretionary service lines: An operational driver of variety by Laurens Debo and Cuihong Li
-Linking Customer Behavior and Delay Announcements: Are Customers Really Rational? by Eric Webb, Qiuping Yu, and Kurt Bretthauer

Wednesday 2:45pm, track WD27:
-The effect of discrete workshifts on non-terminating queues by Robert Batt, Diwas KC, Bradley Staats, and Brian Patterson
-A near-term mortality indicator for terminal cancer patients using high frequency medical data by Donald Lee and Edieal Pinker
-A machine learning approach for personalized health care outcome analysis by Guihua Wang, Jun Li, and Wallace Hopp
-Are patients patient? The effect of universal healthcare on emergency department visits by Diwas KC

Wednesday 4:30pm, track WE32:
-An analysis of world baseball softball confederation premier 12 schedule by Seong Kim and JC Kim
-The role of offensive system in the NBA draft by Ryan Chen, Eli Shayer, Travis Chen, and Nicholas Canova
-Using Past Scores and Regularization to Create a Winning NFL Betting Model by Eric Webb and Wayne Winston
-An optimal pacing strategy for ultramarathons by Kristoper Pruitt and Justin Hill

Qualifying a Worthy Problem

Great talk by Professor Gerald Brown on the last day of INFORMS. If you ever have a chance to hear him speak, take it. Slides from his talk, which started with some motivating examples of military OR, are available here.

5 steps to qualifying a worthy problem to study/solve:
1. What is the problem?
If you can’t describe the problem, how do you know there is one? How would you ever solve it? The client never gives an unambiguous problem description, so work to get to the heart of the matter. If you can describe the problem, move to step 2.

2. Why is this problem important?
Don’t waste your time on trivialities. If the problem is important, move to step 3.

3. How is this problem now solved?
Few problems are entirely ignored, so be sure to understand how the problem is currently solved to ensure you are providing adequate improvement. If you can do significantly better, move to step 4.

4. How will you solve this problem?
Up until now, “solving the problem” has been agnostic toward the type of analysis. Now, choose an appropriate methodology and ensure the problem is tractable.

5. How will you know when you have succeeded?
Answer this before you start solving. It’s difficult/impossible to succeed if the goal is constantly moving, so hammer out what success looks like for this problem.

INFORMS 2016 Presentations

Come show me some love in Nashville. Here are my presentations:

1. Session MC29 – Issues in Energy Efficiency and Renewable Energy
November 14, 2016, 1:30 – 3:00 PM, 202A-MCC
3rd Presentation (of 3): Mind The Gap: Coordinating Energy Efficiency And Demand Response
Authors: Eric Webb, Owen Wu, Kyle D. Cattani, Kelley School of Business, Indiana University
Abstract: Traditionally, energy demand-side management techniques, such as energy efficiency (EE) and demand response (DR), are evaluated in isolation. We examine the interactions between long-term EE upgrades and daily DR participation at an industrial firm. We find that EE and DR act as substitutes in terms of reduction of peak electricity demand, and the long-studied energy efficiency gap between firm-optimal and societal-optimal levels of EE is smaller when DR is considered. We suggest three approaches to reducing the energy efficiency gap, including an original suggestion that relies upon the interactions between EE and DR.

2. Session TC34 – Public Policy and Healthcare Operations
November 15, 2016, 1:30 – 3:00 PM, 204-MCC
2nd Presentation (of 3): Predicting Nurse Turnover And Its Impact On Staffing Decisions
Authors: Eric Webb, Kurt Bretthauer, Kelley School of Business, Indiana University
Abstract: Nurse turnover remains a significant problem in skilled nursing facilities across the United States. High turnover leads to two important questions: (1) Hiring decisions – What applicant attributes should be valued when hiring nurses, in order to hire nurses that are effective at their jobs and likely to stay for a long duration? (2) Staffing decisions – How should nurse workload be managed in order to prevent burnout and decrease turnover? Based on a large dataset from skilled nursing facilities in the United States, we first use a survival model to predict nurse turnover. For this talk we then focus on staffing and incorporate these empirical results into analytical models for nurse staffing decisions.

3. Session WC31 – Consumer Behavior in Services
November 16, 2016, 12:45 – 2:15 PM, 202C-MCC
4th Presentation (of 4): Linking Customer Behavior And Delay Announcements: Are Customers Really Rational?
Authors: Eric Webb, Qiuping Yu, Kurt M. Bretthauer, Kelley School of Business, Indiana University
Abstract: We empirically explore customer abandonment behavior in the presence of delay information using data from a call center. Previous work has assumed that customers are at least partially rational in responding to announcements. In contrast, we relax all rationality assumptions. Our findings indicate that customers exhibit loss aversion behavior. In addition, customers may update their announcement-induced reference point as they hear subsequent announcements. Our results also indicate that customers may fall for the sunk cost fallacy while waiting in the queue. We show the impact of these effects on staffing decisions using a classic staffing model.

4. Session WE32 – Sports and Entertainment
November 16, 2016, 4:30 – 6:00 PM, 203A-MCC
3rd Presentation (of 4): Using Past Scores And Regularization To Create A Winning NFL Betting Model
Authors: Eric Webb, Kelley School of Business, Indiana University, and Wayne L. Winston, Bauer College of Business, University of Houston
Abstract: Is the National Football League betting market efficient? We have devised a profitable betting model that would win 52.7% of the 7,705 bets against the spread it would have made over 34 seasons. Scores from previous weeks are used to estimate the point value of each team’s offense and defense. These values predict next week’s scores, and a bet is placed against the advertised spread. The sum of squares of offensive/defensive point values are constrained to be less than a regularization constant.

What is your presentation’s throughline?

When you start designing your presentation, take a step back and ask yourself: What’s my main point? What do I want the audience to remember, to takeaway? Build your talk around this concept. From the TED Talks book:

There’s a helpful word used to analyze plays, movies, and novels; it applies to talks too. It is throughline, the connecting theme that ties together each narrative element. Every talk should have one.

Since your goal is to construct something wondrous inside your listeners’ minds, you can think of the throughline as a strong cord or rope, onto which you will attach all the elements that are part of the idea you’re building.

This doesn’t mean every talk can only cover one topic, tell a single story, or just proceed in one direction without diversions. Not at all. It just means that all the pieces need to connect.

Basically, make everything in your talk connect to the main point. Minimize the distractions and superfluous details.

Handling Questions Gracefully

1. Accept questions from the audience after finishing thoughts, not mid-thought. Professors do this better than students.

2. Listen. Listen to the question to ensure you understand it.

3. Answer quickly if possible. If a quick answer is not possible and the answer will be revealed in upcoming segments, ask the questioner to wait until then. If a quick answer is not possible and the answer will not be revealed in the rest of the presentation, tell the questioner that you will speak to him offline. Offline refers to after the presentation has ended and you are no longer speaking to the entire audience.

4. If the answer does not satisfy the questioner or he has a follow-up, determine what to do. You don’t want your presentation derailed by a single line of questioning, but you don’t want to look evasive either. If the question wasn’t clear to you (even after asking for the questioner to repeat), offer to answer the question offline. In small audiences, it’s better to delay and answer correctly (or explain why you don’t know) than to guess at the questioners meaning and misinterpret.

Slide Do’s and Don’ts (2016 edition)

Most people use slides when giving a presentation. Unfortunately, most slides are awful. Here are some tips to help.

This is a post I will re-visit each year. Here are the 2016 suggestions:

Do’s
1. Use images and graphs. As much as can reasonably be allowed. Visuals trump words.

2. Make your slides readable. Use at most three font sizes. Large size is for titles/headlines, medium is for your main ideas, small is for supporting ideas. You may use bolding, but avoid italics and underlining as they are hard to read. Ensure contrast between the text and background.

3. Think about going to a blank slide when you want to talk and need the audience’s full attention. Most presentation clickers have a button that can do this. From the TED Talks book: “Just go to a blank, black slide and then the audience will get a vacation from images and pay more attention to your words. Then, when you go back to slides, they will be ready to go back to work.”

Don’ts
1. Don’t use slides made in LaTeX. LaTeX is great for making papers. But it makes boring presentations that look like paper subsets. Presentations aren’t for reading, they’re for listening and seeing. Every presentation I’ve seen where the slides were made in LaTeX Beamer has been boring. And, perhaps worse, each presentation is the same sort of boring, as there seems to be little customization. I’m sure it’s possible to give a good presentation with this tool, but I haven’t seen it.

2. Don’t create slides that work as a stand-alone document. You are there to give your presentation. If everything you want to say is already up on the slides, why are you there? Your job is not to make all-encompassing slides. It is to make slides that support your presentation.

3. Don’t put items on your slides that you don’t want to discuss. If the information is not important enough to mention, it shouldn’t be on your slides. Generally, slides shouldn’t speak for themselves. If you think someone might ask about it, move it to your backup slides.

Memorize Your Talk?

Should you memorize your presentation? There are two correct answers to this question:
1. No, you shouldn’t memorize your talk, but you should be familiar with the flow and have practiced enough that it sounds natural.
2. Yes, you should memorize your talk so well that it sounds natural, not robotic.

Personally, I don’t memorize wording. But I know memorizing makes certain people feel more relaxed and secure. The important thing is that you practice enough so that your talk sounds natural.

If you don’t memorize and don’t practice, your talk will be littered with “uhs” and “umms” while you try to work out what to say next. That sucks.

If you memorize, but not well enough, you’ll fall into the uncanny valley. From the TED Talks book: “[The Uncanny Valley] is a term borrowed from a phenomenon in computer animation where the technology of animating humanlike characters is super-close to seeming real but is not quite there. The effect is creepy: worse than if the animator had steered clear of realism altogether.” Speakers that try to memorize, but don’t do it well enough, fall here. Their talk will sound robotic and rehearsed. But by persisting with practice, you can fight through the uncanny valley. You need to know the talk so well that recalling it is a snap, no matter where you are interrupted. “Then you can use your conscious attention to focus on the meaning of the words once again.”

uncanny-valley
(image via Slate)

So what is the conclusion? Practice. Then practice again. By practicing and preparing, you can give a talk that values the audience’s time and wins you support.

Planning the Length of Your Talk

For every one minute you exceed your allotted time, 10% more of the audience wants to kill you.

I am a stickler for timing, and I am typically incredibly annoyed when presenters cannot plan well and run overtime. When you run overtime, you are implicitly telling the audience (and any subsequent presenters) that their time is not valuable to you. So my first advice this week is to plan an appropriate presentation length.

Your prepared talk is only one of four things that will happen in your allotted time. The other three things are introductions/setup, questions during the talk, and questions after the talk. When you first take the stage, any introduction and preparation of slides must be considered in your timing. If you are in academia, your talk may be frequently interrupted with clarifying questions or interjections from the crowd. Such questions may be held until the end of the talk for certain kinds of talks. And then, at the end, there will be audience questions about the topic and talk.

If you use all your time on your talk and get no useful questions, your talk has been a failure. Questions can alert you to holes in your current research. Questions can illustrate future avenues of fruitful research. A lack of questions signifies that you either did not interest the audience or you did not leave enough time for questions.

So how much talk should you plan for? Let’s assume you are giving a 22 minute academic presentation. Leave about 5 minutes at the end for questions. Assume 1 minute setup and 3 minutes of interruptions during the talk. That leaves 13 minutes. Plan a talk that you can comfortably give in 13 minutes. Without being rushed. That’s not a long time. You won’t be showing all the details of your research. But at a conference where attendees can hear 20 talks a day, they won’t remember all the details anyway. Your goal is to present motivation, basic approach, and interesting results in order to get other people interested in your research.

If no one interrupts you (saving three minutes), have some backup material on the next most interesting aspect of your work to show.

If no one asks questions, start the Q&A session by asking for advice on a specific aspect of the research. Don’t just let the five minute Q&A time wither unused. If you get too many questions and are out of time, offer to take more questions offline after the presentation.

For a talk that won’t be interrupted (questions at the end, if at all, like a TED talk), here is the timing suggestion from the TED book: “Your finish line is your time times 0.9. Write and rehearse a talk that is nine-tenths the time you were given: 1 hour = 54 minutes, 10 minutes = 9, 18 minutes = 16:12. Then get on stage and ignore the clock. You’ll have breathing room to pace yourself, to pause, to screw up a little, to milk the audience’s response. Plus your writing will be tighter and you’ll stand out from the other speakers who are dancing to the rhythms of the same time limit.”