Forecasting Sales of Durable Goods – Does Search Data Help?

Search data can be used to forecast macroeconomic measures. The
present study extends this research direction by drawing on real sales data from a
household panel over two years. Specifically, the study analyzes whether search
data improves forecasts for seven products groups of durable goods. The forecast
model also includes the average weekly price and a dummy for the Christmas
season. Forecast accuracy is indeed improved when search data is included even
for product groups that have a short information and search phase. The product
groups, however, need to be chosen carefully, because some durable goods show
no lag between online search and purchase.