Google Analytics data can help predicting tourist numbers, research shows


Google Analytics data could help the tourism industry gain a better understanding of tourist dynamics and trends when traveling, according to recent research MODUL University Vienna conducted analyzing Google Analytics data, according to a press statement.

Data extracted by Google Analytics from travel information sites can contribute to better and more accurate predictions of tourist numbers for large cities, especially for time periods within the next three to twelve months. This is the recently published result of a research project conducted at MODUL University Vienna that analyzed the use of selected Google Analytics data sets in estimating future tourist numbers for large cities. The project formed part of the university’s research focus on the use of new media for modern management.

Travel information sites are widely used by people interested in traveling in the near future. Google Analytics, a software designed to track and report website traffic, collects data on visitors’ usage patterns and provides anonymous and averaged statistics. Currently, this information is mainly used by IT departments in order to optimize web design. Dr. Ulrich Gunter and Dr. Irem Önder at the Department of Tourism and Service Management at MODUL University Vienna in Austria have now proven that there is more to these data than has previously been shown. They evaluated the power of these data for forecasting future tourist numbers for large cities, an important factor for resource management in tourism.

“Our results clearly demonstrate that complementing certain forecast models with Google Analytics data can make the models quite powerful for predicting future numbers of tourists to a given destination,” Dr. Gunter explains, commenting on the work published in Annals of Tourism Research.

The Google Analytics variables used by Dr. Gunter and Dr. Önder included Average Session Duration, Average Time on Page, Bounce Rate, New Sessions, Page Views, Returning Visitors, Social Network Referrals, Total Sessions, Unique Page Views, and Users. All of these data were extracted during a period of time beginning in August 2008 and ending in October 2014.

The team used the data for so-called vector autoregression (VAR) models. These are econometric forecast models suitable for cases where several mutually causal variables need to be integrated. “Overall we found that models that included the Google Analytics data predicted future tourist numbers better for time periods of three to twelve months ahead than models without these additional data,” explains Dr. Önder. “Whereas for shorter time frames, models without them performed better.” In order to assess the performance of the respective models, the team compared predicted tourist numbers with actual values provided by the TourMIS database, a leading European database for tourism information developed by research associates of MODUL University Vienna.