SPECIAL FEATURE: Pitfalls in Forecast Evaluation
To lead this issue’s special feature, Christoph Bergmeir explores forecast evaluation hazards that data scientists need to be aware of, and offers better practices to avoid them. In commentaries, Zeynep Erkin Baz notes that these same risks apply generally to any forecasting practitioner, arguing that understanding the business problem before starting the solution is a key to success. Then, Shari De Baets takes a decision science perspective to focus on one pitfall, that a forecast needs to be a realistic scenario.
Cross-Learning Huijing Chen, John Boylan, and Ivan Svetunkov propose an enhanced taxonomy, based on vector exponential smoothing models, that can aid in cross-learning from similar series with very short histories. Their approach addresses a challenge common in retail forecasting. In commentaries, Stefan de Kok proposes a further enhancement to the new taxonomy, and Malvina Marchese notes the continuing value of exponential smoothing, especially in this new cross-learning-enhanced version.
Commentary on 2023:Q1 Special Feature: Does Forecast Accuracy Even Matter? Foresight’s issue 68 special feature continues to generate discussion, with Integrated Business Planning Editor Niels van Hove exploring the limitations of forecasts and plans on decision making.
Government and Public Policy In 2005 (Foresight issue 2), Elaine Deschamps organized a special feature on the “Organizational Politics of Forecasting.” As our Government and Public Policy Editor, Elaine now revisits the topic and her earlier guidance, adding key learnings from the disruption brought by the COVID pandemic.
- In his review of Nouriel Roubini’s Megathreats, Ira Sohn warns us that the future is going to be a bumpy ride.
- Stephan Kolassa recommends Nicolas Vandeput’s Demand Forecasting Best Practices as an alternative to painful experience, especially for anyone just starting out in demand forecasting.
- New contributor Sara Barradas finds a valuable read in two complementary Hans Levenbach books, Change & Chance Embraced and Four P’s in a Pod.
Practitioner’s Corner Patrick Bower, joined by Chip Gretok, share their list of 10 lies told in consensus meetings, gleaned from the over 50 years of supply chain experience between them.
Preview of 2023:Q4 Special Feature: The Current State and Future Direction of Forecasting Software Excel is usually the starting point when organizations begin forecasting, and many organizations fall back to Excel after failed implementation of more advanced demand planning software. So why does this continue to happen, and what can be done? These are just two of the questions posed by Igor Gusakov in the search for what is wrong with demand planning software. Foresight’s Q4 issue will have a special feature on “The Current State and Future Direction of Forecasting Software,” with contributions from Anne-Flore Elard, Johann Robette, and other vendors and practitioners in the field.
Spotlights Mark Chockalingam is founder and president of Valtitude/Demand Planning LLC and Malvina Marchese is Foresight’s Financial Forecasting Editor and a Senior Lecturer at Bayes Business School in London.
- Thanos Goltsos and Aris Syntetos on forecasting for a circular economy.
- Stephan Kolassa on flatline forecasts.
- Malte Tichy on the technological limits to forecasting.
Minitutorial Lawrence Vanston encapsulates his many years of long-range forecasting experience into a two-page infographic on using S-curves for forecasting new product adoption.
To receive this issue, start your membership to the IIF today! Join Now