Successful LCF Workshop on New Product Forecasting
12 May 2016
Lancaster Centre for Forecasting (LCF) hosted a very stimulating half-day workshop on New Product Forecasting on 22 April in London, with an attendance of about forty practitioners from a wide variety of organisations.
The first talk was by Paul Goodwin, Professor Emeritus at Bath University. He outlined the problems and biases associated with unaided judgement. He showed how structured assumption modelling can help by decomposing the problem into its constituent elements. Paul shared a case-study of forecasting the sales of a new drug for hay fever in the USA, by considering the number of sufferers, the age demographic served by the drug and its penetration through a variety of outlets.
The second talk was by Daniel Barrett, Head of the Demand Planning Centre of Excellence at LEGO. He emphasized the importance of having a structured process throughout the company. This involves close collaboration between Demand Planning and Marketing in the stages of development pre-launch and between Demand Planning and Sales in the early stages after launch. LEGO also uses measures of Forecast Value Added to monitor judgmental changes to any statistical forecasts.
Sandra Maor, a Director of the European Forecasting Team at Nielsen Innovation Practice, spoke about their BASES approach. This approach is, to some degree, dependent on intentions data gathered from consumer surveys (for which Nielsen are very well known). Evidence from many surveys, conducted in different countries and products, has enabled Nielsen Innovation Practice to refine their models for New Product Forecasting and to adjust for the differences between stated intentions and behaviour.
The final talk was given by Dr Sven F. Crone, Co-Director of LCF. He presented a case-study of New Product Forecasting at a company specialising in fabrics and bedding. Such products are not truly ‘new to the world’ and are usually new variations on existing themes. This gives the opportunity of matching data on new launches, from the very first observation, with previous patterns of launches, to help predict the growth of sales of the new product.
The day concluded with a panel discussion, chaired by Professor Robert Fildes, Founding Director of the Lancaster Centre for Forecasting. The workshop provoked much discussion amongst the audience, which comprised around forty practitioners from a wide variety of organisations, and elicited very positive feedback from participants.
Presentation downloads:
- Sven F. Crone - Predictive Analyticsfor New Product Forecasting. More Statistics and More Analytics and less Judgment: a Case Study
- Paul Goodwin - Making the best use of management judgment in new product forecasting