Forecasting the importance of product attributes using online customer reviews and Google Trends

During the early stage of product design, product manufacturers seek to identify the most relevant product features that will meet the demands and needs of consumers. Conventionally, several surveys have to be undertaken during the time interval between product design and the launch of anew product, to understand any changes on the importance of the product attributes. However, the process is time-consuming and costly. Recently, online customer reviews have been generated on many websites and can be used to analyse the change of the importance of the product attributes. Also, Google Trends has been adopted in previous studies to understand consumers interests in certain products over a period of time and can be considered in analysing the change in product attributes importance. However, no such kinds of studies have been reported. This study aims to present an empirical approach that uses online big data, to identify and predict product design attributes of products that will be relevant to consumers in the future. To achieve this aim, we propose a methodology for forecasting the future importance of product attributes based on online customer reviews and Google Trends. A case study on an electric hairdryer is presented to illustrate the proposed methodology. Validation tests on the proposed fuzzy rough set time series method were conducted. The test results indicate that the proposed method outperforms the fuzzy time series, the fuzzy k medioid clustering time series and the ANFIS method in terms of forecasting accuracy. Our results contribute to the processes of new product development and can potentially assist R&D managers to establish methodologies and processes for product designs capable of generating higher returns.