John K. Thompson
Keynote Speaker – Thursday, December 1
Analytics Teams: How to Design, Build, Manage and Grow a World Class Analytics Organization
Analytics teams are not like other teams. They are not software development teams. They are not data engineering teams. They are not project management or outsourcing management teams. Analytics teams need to be designed for creativity, innovation, failure/learning, invention, and to drive change. Analytics teams are more akin to a collective of artists than a software development team. We will talk about all these aspects of analytics teams and more.
Attendees will learn:
- How to architect an analytics team, 3 options
- How to ensure value delivery, and
- How to evolve an analytics team over time
Edward Malthouse
Integrated Marketing Communications and Industrial Engineering and Management Science
Northwestern University
Keynote Speaker – Friday, December 2
Saving Local News With Big Data, Lifetime Value Models and Causal Inference
We have worked with Google, Mather Economics and midwestern news organizations to develop the Subscriber Engagement Index (SEI) system that helps news organizations find sustainable revenue models. SEI implements reporting, segmentation, churn, and lifetime value forecasting and optimization models using clickstream and payment data from publishers. Causal inference approaches and communication theories allow for us to imagine and prescribe new touch points to create value for readers and improve subscriber retention.
Rhiannon Clifton
SHIELD Illinois
Len Musielak
SHIELD Illinois
Best Practices When There are No Practices: What To Do When the Data Doesn’t Exist
We live in a data driven world. We were taught, and have taught, that the best business decisions are based on data. What happens when there is no data? When you are doing something that has never been done before? When you need to make decisions in line with an organism the world didn’t even knew existed the year before? SHIELD Illinois operations leaders Rhiannon Clifton and Len Musielak tell the informative, inspiring and entertaining story of how a small group from the University of Illinois rolled out testing to 1,700 sites in three months in 2022.
Attendees will learn:
-
Be creative: even when you don’t think you have data, data exists and can be useful
-
Continually check your assumptions as your data grows
-
Rethink your assumptions and look for new ways to model
Polly Mitchell-Guthrie
Kinaxis
How More Useful Forecasts Can Improve Supply Chains
If supply chains had perfect information, perfect forecasts would allow us to make exactly the right amount to balance demand and supply. But forecasts are not perfect predictions of the future, and never will be. George Box was right when he said “All models are wrong but some are useful,” so what does it take to make a forecasting model useful? Wrong” forecasts are rarely about the math and more often about information failures before and after the model is built. This talk will explain how intelligent agility can mitigate these tradeoffs, even in the face of the tremendous disruptions we’ve seen.
Gary Cokins
Analytics-Based Enterprise Performance Management (EPM)
Many organizations are far from where they want and need to be with improving their performance. They apply intuition, rather than fact-based data and analytics, when making decisions. Enterprise performance management (EPM) is now viewed as the seamless integration of managerial methods such as strategy execution with a strategy map and its companion balanced scorecard (KPIs) and operational dashboards (PIs); enterprise risk management (ERM); capacity-sensitive driver-based budgets and rolling financial forecasts; product/service/channel/customer profitability analysis (using activity-based costing [ABC] principles); supply chain management; lean and Six Sigma quality management for operational improvement; and resource capacity spending planning. Each method should be embedded with business analytics of all flavors, such as correlation, segmentation, regression, and clustering analysis; and especially predictive analytics as a bridge to prescriptive analytics to yield the best (ideally optimal) decisions. This presentation will describe how to complete the full vision of analytics-based enterprise performance management.
This presentation will cover:
- Why business analytics, with emphasis on predictive analytics and pro-active decision making, is becoming a competitive advantage differentiator and an enabler for trade-off analysis.
- How strategy maps and their companion balanced scorecards communicate strategic objectives with target-setting to help cross-functional employee teams align their behavior to the strategy and better collaborate.
- Why measures of channel and customer profitability and customer value are now superseding profit and service-line measures – and shifting from product to customer-focused organizations including future potential value – customer lifetime value.
- How activity-based cost management (ABC/M) provides not only accurately traced calculated costs (relative to arbitrary broad-averaged cost allocations), but more importantly provides cost transparency back to the work processes and consumed resources, and to what drivers cause work activities.
- Reforming the broken annual budgeting process with performance based budgeting that links strategy to operations and is process volume sensitive rather than simply incremental at each cost center.
- How EPM/CPM also applies to public sector government to understand their “output costs” and better serve citizens.
- How all levels of management can quickly see and assess how they are doing on what is important – typically with only a maximum of three key performance indicators (KPIs).
- How to integrate performance measurement scorecards and ABC/M data with:
- Strategy formulation.
- Process-based thinking and operational productivity improvement.
- Channel/customer profitability and value analysis and CRM.
- Supply chain management.
- Quality and lean management (Six Sigma, cost of quality).
Melanie Shanks
Morgan Templar
Beverly Wright, CAP
Burtch Works
Panel Discussion – Analytics for Social Good
Join our panelists in a discussion on how analytics and operations research experts can contribute to society via pro-bono and social good efforts. Learn about our panelist’s personal experiences and ways for you to use your skills and knowledge to change the world for the better!
Impactful O.R./Analytics Prize Presentations
INFORMS Chicago Chapter is sponsoring an award competition to encourage students as well as scholars from academia and industry to have an impact on their community, infused with O.R./Analytics methods. The winners of this year’s competition will present their impactful work in two categories: (1) Academic/Researcher of the year, and (2) Student of the year.
Pengyi Shi
Krannert School of Management, Purdue University
Jonathan Helm
Researcher of the Year: Pengyi Shi
This will be a joint presentation led by the two researchers from the academic team. The second presenter is Jonathan Helm.
Delta Coverage: The Analytics Journey to Implement a Novel Nurse Deployment Solution
During and after the pandemic, hospitals suffered from excess demand and diminishing supply of nurses. In partnership with the largest health system in Indiana with 16 hospitals statewide, we embarked on a journey to co-develop and implement a suite of data and decision analytics to support a novel internal travel nursing program. Our tool integrates state-of-art machine learning based time-series prediction with a nested, multi-stage stochastic optimization to produce nurse transshipment decisions. Our practice-driven methodological innovations were implemented in October 2021 through a portable PowerBI analytics dashboard that has led to an estimated annualized reduction of 4% understaffing and 3% misallocation of resource nurses with a saving of over $400K. We continue to work with our partner for broader dissemination of our work to other hospitals through the health systems internal innovation program.
Bio
Pengyi Shi is an associate professor at the Krannert School of Management, Purdue University. She received her PhD degree in industrial engineering from Georgia Institute of Technology before joining Purdue in 2014. Her research interests include data-driven modeling and decision-making in healthcare and service operations. She has collaborated with practitioners and faculty members from different healthcare organizations, including major hospitals in the U.S., Singapore, and China. Her research has won the first place of MSOM Responsible Research in OM Award in 2021, the first place of INFORMS Pierskalla Best Paper Award in 2018, and the second place of POMS CHOM Best Paper Award in 2019 and 2020.
Jonathan E. Helm is an associate professor, Grainger Fellow, and Co-Director of the Center for the Business of Life Sciences at Indiana University’s Kelley School of Business. He received his PhD in industrial and operations engineering from the University of Michigan before joining Indiana University in 2012. He has been a part of analytics implementations in hospitals in the Midwest and East Coast of the United States and internationally. His research has won awards including MSOM Responsible Research, the Pierskalla Best Paper, and POMS CHOM Best Paper.
Zhiqiang Zhang
University of Chicago Booth School of Business
Student of the Year
When Machine Learning Classifications Impact Resource Allocation Decisions: Combining ML and Human-in-the-Loop Design to Improve Incarceration Diversion Decisions
Machine learning (ML) tools are widely used to assist judges to make incarceration diversion decisions (probation, community corrections, etc). However, prediction errors could lead to suboptimal recommendations of diverting people to these programs. We develop decision support tools that combines queueing theory and ML algorithms for routing individuals that have different risks of recidivating to these programs. The resulting ML-based routing decisions account for tradeoffs among limited capacity of the treatment programs, treatment effect, and recidivism and incarceration costs. We show that ML-based routing decisions lead to a simple prioritization scheme, which is optimal when we can assess the recidivism risk with a high degree of confidence, but would be suboptimal otherwise. This leads us to propose guidelines for when to follow the ML-based routing recommendation and when to also involve human judgement. We are collaborating with Adult Redeploy Illinois (ARI), a state-wide community-based diversion program, to calibrate and validate our model.
Bio
Zhiqiang Zhang is a first year PhD student at the University of Chicago Booth School of Business, in the dissertation area of management science and operations management. Prior to his PhD study, he received his bachelor’s degree in mathematics at the University of Chinese Academy of Sciences and his master’s degree in computational and applied mathematics at the University of Chicago. His research interests lie in combining structured stochastic models with data science. Involved methodologies include queueing theory, machine learning, optimization, statistics, and game theory.
