Forecasting the future
Download as PPTX, PDF0 likes367 views
This document discusses challenges in forecasting for the travel industry and lessons learned from forecasting research. It summarizes that: 1. Both statistical models and human judgment have been unable to fully capture future uncertainty and accurately predict black swan events. 2. Simple statistical models often predict the future better than more complex models, despite fitting past data less accurately. 3. Averaging predictions from multiple individuals generally improves forecasting accuracy over relying on expert judgment alone. Managing complexity through signal detection, contingency planning, and managing black swans is important.
Forecasting the future
- 1. Forecasting - Does it still work? EyeForTravel November 29, Prague Kate Varini, Oxford Brookes Larissa Koupriouchina, Hotelschool The Hague
- 4. Uncertainty is different
- 5. Uncertainty is different
- 6. Black swans are on the increase See the video clip here http://www.youtube.com/watch?v=BDbuJtAiABA
- 8. ❖ Mobile web access from smartphones will surpass traditional PCs by 2013 (Gartner Research) ❖ 1,200 % increase in 2011 of travel category searches via mobile device 2011(Google) ❖ 3,000 % increase of hotel searches on Google Maps (Google) ❖ Mobile bookings are expected to exceed $160 million during 2011 (PhoCusWright) Speed of transactions
- 9. Flash sales are moving from niche to mainstream distribution channel
- 11. Budget hotels in top 6 Chinese cities
- 13. Climate change
- 14. Traditional segmentation doesn’t work
- 15. Traditional segmentation doesn’t work
- 16. Managing complexity • Signal detection • Contingency planning • Managing black swans?
- 17. 5 Top Lessons from forecasting research 1. “Simple” models do not necessarily fit past data well but predict the future better than complex or sophisticated statistical models. 2. Both statistical models and human judgment have been unable to capture the full extent of future uncertainty.
- 18. 5 Top Lessons from forecasting research 3. Expert judgment is typically inferior to simple statistical models. 4. Forecasts made by experts are no more accurate than those of knowledgeable individuals. 5. Averaging the predictions of several individuals usually improves forecasting
- 19. Judgmental forecasting Conventional wisdom Empirical findings The amount of information does The more information we have, not improve the accuracy, the more accurate the decision instead it increases the confidence Monetary rewards and Human behavior is too complex punishments contribute to better to be motivated by monetary performance factors alone In many repetitive, routine Experience and /or expertise decisions, experience and/or improve accuracy of decisions expertise do not contribute to future-oriented decisions
- 20. Forecasting by experts
- 21. How to deal with