The Lancaster Centre for Forecasting is studying the practice of data sharing between firms and the firms they supply their product to with the aim of improving forecast accuracy. Here is a description of the research project and a call for participants to complete a questionnaire:
Collaborative forecasting research study
The increased volatility and competitiveness of today’s market has led firms to engage in collaborative forecasting and information-sharing practices in order to improve forecast accuracy. However, knowing how and when to collaborate with downstream partners remains a challenge. It seems that the benefits are widely touted though little substantive research exists about how to best use the mass of downstream data available to forecast more accurately.
At the Lancaster Centre for Forecasting we are conducting a research study to address some of the open questions in this domain, initially through a web-based survey of forecasters and demand planners. The survey can be found at www.collaborative-forecasting.co.uk/survey/cf_survey.html and your participation is encouraged.
All survey respondents will be entitled to a copy of the results and their data will remain private with no details of specific cases made available. The survey should take between 15 to 20 minutes to complete and it is possible to use the save & continue option to break up the task. Subsequent follow-up interviews and case studies may be conducted with interested companies to obtain an even richer picture of the current state of the art.
Please feel free to extend this request to any colleagues or members of your network who would be suitable respondents (forecasters/demand planners at manufacturing companies), however we are only looking for one response per business unit.
To be more specific, we are focusing on the following research questions:
· What forms of collaboration are companies participating in and what data is being shared by their downstream partners?
· How are firms using this information, if at all, in their forecasting process? Through statistical methods, analytics or judgement?
· How do firms cope with differing data conditions and forecasting requirements of different customers?
Our theory is that there is a huge amount of data available, but except for a few idealistic cases published, companies are either not able to leverage the data through their forecasting systems and rely on either a judgemental or exception management approach rather than the advanced statistical techniques available.
In a later phase, we look to test alternative methods of integrating downstream data into the statistical forecasting process automatically and compare accuracy with established benchmark methods.