I chose the following excerpt as an introduction to the article:
Following the growing number of Internet users and the increasing popularity of Google as a search engine for obtaining information about cars, Google search data become a leading indicator for the long-term forecasting of product sales. Since 2004, it has offered a tool called Google Trends, which provides information on the relative interest of users in a particular search query, at a given geographic region and at a given time (the data are available on a weekly or even daily basis). Moreover, Google Trends can attribute queries to different search categories (Autos, Computers, Finance, Health and others). In recent years, researchers worldwide have begun to use online search data to produce real-time forecasts where information from official sources is released with a lag (such as ‘nowcasting’), or simply as an additional variable for forecasting purposes (from Fantazzini and Toktamyasova “Forecasting German car sales using Google data and multivariate models”, December 2015)
Consumer interest to a product comes first. Then goes the sales. Google Trends allows for monitoring a consumer interest to a certain product by tracking frequency of the product related search queries like ‘buy tv’. Researchers can refine the queries to compare interest in the product categories over time and geographic location.
The raises and falls in the keyword search frequency curve follow actual sale pattern, predicting it with a lag from week to months! You can see how the spikes indicate raise of consumer interest to the TVs before the end of a year, clearly signaling the expected sales peaks during Christmas sales. Comparing product categories, the researchers can forecast product penetration curves for new TV sizes.
Note that the interest in 32 inch diagonal TVs over-performed the interest to other, bigger sizes. That was a surprise for TV makers, although, the reality is that 32-36 inch TVs still sell better than TVs with larger screens. Why? That was counter-intuitive at the beginning until the surveys showed that people like mid-size TV sets for its fit to a standard living room size, since most of the population live in the city apartments with a smaller living space than that in a house.