Making a successful sales prediction or forecasting in retail markets remains challenging despite years of practice and efforts. In this study, we attempt to address this challenge by incorporating the Google Trends search data into traditional time series models that feature geodemographic and industrial-level variables for the purpose of predicting Hass avocado sales in different regions of the United States. The results imply that, for conventional Hass avocados, the use of Google Trends search data can produce better predictions than the models without Google Trends search data. Moreover, using categorized Google Trends search data can improve predictive results even more. However, the models with Google Trends search data fail to improve the predictive power for the consumption of organic Hass avocados. The results suggest that categorized Google Trends search data can be helpful in improving prediction and forecasting for various business stakeholders in general.

尽管经过多年的实践和努力，在零售市场做出成功的销售预测或预测仍然具有挑战性。在这项研究中，我们试图通过将谷歌趋势搜索数据整合到传统时间序列模型中来应对这一挑战，这些模型以地理人口和工业水平变量为特征，以预测美国不同地区的哈斯鳄梨销量。结果表明，对于传统的哈斯鳄梨，使用谷歌趋势搜索数据可以比没有谷歌趋势搜索数据的模型产生更好的预测。此外，使用分类的谷歌趋势搜索数据可以进一步改善预测结果。然而，带有谷歌趋势搜索数据的模型未能提高对有机哈斯鳄梨消费的预测能力。