LCF presents at ICT and Innovation Forecasting workshop in Paris
26 June 2015
26 June 2015
Professor Robert Fildes and his doctoral student Oliver Schaer attended the 2nd Workshop on ICT and Innovation Forecasting, held Paris during May 2015.
This international workshop series, organised by the International Institute for Forecasters (IIF), and supported by ESSEC, aims to bridge the gap between theory and practice by bringing together leading academics and practitioners.
Fast-paced technological development drives consumer behaviour. Dealing with this challenge is extremely difficult and in particular for ICT companies, where large infrastructure costs are involved to provide the desired service level to the customers. Forecasting the adoption and impact of technology, therefore, is an important task for exposed companies if they are to remain in the market, but yet it is a difficult one. The highly stimulating workshop included talks ranging from the latest development in the ICT forecasting literature to measuring the impact of regulation and becoming a fore-acting company (that is to say proactive in determining your technological future). In addition, talks from practitioners clearly demonstrated the value of good forecasts.
The presentation from Robert Fildes focused on the boundaries of ICT forecasting and the many diverse problems arising from the blurring of boundaries between mobile computing, entertainment and leisure activities and the consequential changes placed on mobile design and platform functionality. Such changes not only affect user behaviour, but challenge the providers of ICT infrastructure and influence demand for equipment and components. The presentation discussed the reliance of the providers of the different platforms and their business models on forecasts of such changing consumer behaviour. Examples of this diverse set of problems included the new voice-call function included in WhatsApp or Google Glass. Robert Fildes demonstrated insights into possible approaches while the lack of research evidence on the resulting accuracy suggests a multi-method approach focussing on the range of probabilistic outcomes.
Oliver Schaer presented his latest research findings on forecasting product sales with search traffic data as an explanatory variable. Intuitively one may expect SE data to be leading product sales, since many customers nowadays research products online prior to purchase. But one might question this direction of causality, as well as whether SE information is leading, contemporaneous or lagging. Moreover, is this lead/lag-order consistent over time? First results from the case study with product sales from the video games industry and Google Trends data observed leading information to present in many cases, but with frequent changes between lead and lag orders. Moreover, including search traffic information for forecasting Christmas sales did not improve accuracy substantially. This contradicts findings from the literature and rises questions on the usefulness of such data.