MARKET RESEARCH AND FORECASTING

Often conventional research and forecasting methods do not work and new ones have to be invented.

For one project I had to forecast the sales of television sets in the different countries of South America. The client wanted to promote a local factory for electronic TV components. Thus, a forecast over at least a decade was required. This was at a time when there were very little historical data available on that continent. So I looked at historical data in North America and Western Europe. I adjusted for the differences in economic level by looking at the sales of new television sets as a fraction of the country's gross national product (GNP). The GNP and forecasts of it are available for most countries. Plotting the percentage of money spent on television sets on the vertical axis and the saturation of television sets (percentage of households with at least one set) on the horizontal axis, I found that these curves are almost identical in every country. So I used these curves to forecast the markets for new television sets in the South American countries. These forecasts turned out to be very good for many years. The method is described in my article of August 1982 in the Journal of Marketing Research, and is mentioned in market research textbooks.

For another project I could not find statistics about industrial activities in the different regions of Italy. I did find statistics of the industrial use of electricity and I used that as an approximation. First I adjusted for (deducted) the use of electricity by large consumers such as aluminum smelters.

Bertil Lindberg

List of deliverables. 

Ask a question 
Form for requests and comments.

Top of this page

Last revised July 16, 2003.
ELN