Analyzing and forecasting online tour bookings using Google Analytics metrics

An essential part of business operation for tourism industry is revenue management, i.e., how to sell the right tour package, to the right customers, at the right time, at the right price through the most appropriate and cost-effective channels. In today's world, the internet has revolutionized many business operations in the tourism industry which plays an important role in Thailand's GDP. Most tour operators utilize websites as the main channel to build relationships with customers. Thus, website performance measurement is an important strategic factor for online marketing. The objectives of this research were to identify factors contributing from Google Analytics metrics to online bookings and to forecast online bookings using those impactful factors. Several machine learning models namely artificial neural network (ANN), support vector regression, and random forest, were proposed to forecast online bookings using the mean absolute percentage error (MAPE) as the criterion for comparison. It was found that there were three Google Analytics metrics that contributed to online bookings, which were the sessions from referral, unique returning users, and the average session duration. In addition, the ANN model provided the highest accuracy result with a MAPE of 11.39%. The framework from this research can be applied to other online companies to forecast their online bookings, which is an important part of revenue management since accurate forecasts can help companies to achieve their goals.