Kantar Worldpanel -

Demand Forecasting

At the unique propositioning of consumer behaviour data, Worldpanel analytics team using the live behavioural data combined with the most advanced set of Machine Learning algorithms in the marketplace to facilitate FMCG players identifying future growth potential for channel, category or brand performance which consists of three stages as below, 

Data inputs: Trackable data at monthly/quarterly level from a consistent source combining Government, Kantar or client data sets. It is used to identify the influencing factors that may impact category sales. 

Machine learning Process: Testing more than 10 AI/ML approaches for the best result. Building a model that looks at the relationship between those variables and category sales. 

Simulator outputs: Forward forecast the most important variables and feed these back into the model to get category sales predictions in the simulator. 

Our approach, based on consumer behaviour centric forecasting, enables confident decision making and effective prioritization amongst a rapidly changing society and shifting consumer behaviour.

Get in touch

Piradej Boonyavanit
Analytics Lead - Kantar, Worldpanel Division Thailand