Forecasting new product diffusion using both patent citation and web search traffic.

Area of Science:

• Business and Economics
• Technology Management
• Market Research

Background:

• Accurate demand forecasting for new technology products is crucial for business success.
• Existing forecasting methods may not fully capture the dynamics of novel product adoption.
• Understanding diffusion patterns is key to predicting market penetration.

Purpose of the Study:

• To propose and validate a novel forecasting method for new technology products.
• To assess the distinct roles of technology diffusion and interest diffusion in sales prediction.
• To evaluate the model's applicability across different product types (consumer vs. industrial).

Main Methods:

• Developed a forecasting model integrating technology diffusion (measured by patent citations) and interest diffusion (measured by web search traffic).
• Applied the model to forecast sales data for hybrid cars and industrial robots in the US market.
• Analyzed the predictive power of each diffusion measure for different product categories and time horizons.

Main Results:

• Technology diffusion (patent citations) effectively explains long-term sales for both hybrid cars and industrial robots.
• Interest diffusion (web search traffic) improves short-term sales predictability for hybrid cars.
• Interest diffusion showed limited explanatory power for industrial robot sales due to market-specific characteristics.

Conclusions:

• The proposed model, combining technology and interest diffusion, offers a robust approach to forecasting new technology product sales.
• The relative importance of diffusion measures varies depending on product type and market characteristics.
• The model demonstrates particular effectiveness in explaining the diffusion patterns of consumer goods.