Objective  With the acceleration of population aging, the market demand for footwear for older people continues to grow. However, there is a disconnect between the consumer experience of existing products and technical optimization. To accurately capture the real needs of elderly consumers, this paper proposes a research method based on sentiment analysis of shopping review texts on e-commerce platforms. It aims to explore consumers’ demands for footwear for older people, convert them into actionable design indicators, and enhance the consumer orientation and market adaptability of product development.

Methods  Taking the reviews of the top 10 best-selling footwear for older people on three major e-commerce platforms (i.e., JD.com, Taobao, and Pinduoduo) as the research data source, a total of 47,371 valid reviews were collected. In the topic extraction stage, the Bidirectional Encoder Representations-based Topic (BERTopic) model was used for topic clustering of the review texts, identifying 8 topics of high concern to consumers, including fabric comfort, practicality and daily matching, appearance and details, breathability, size accuracy, elasticity and softness, stylish and versatile design, as well as logistics services and shopping experience. In the sentiment tendency analysis stage, a hybrid Bidirectional Encoder Representations from Transformers-Text Convolutional Neural Network (BERT-TextCNN) sentiment analysis model was constructed to conduct quantitative sentiment analysis on each topic, resulting in sentiment tendency scores for each. Based on the negative sentiment intensity and topic weights derived from the sentiment analysis, 6 core needs were selected. The House of Quality (HoQ) in Quality Function Deployment (QFD) was established to convert consumers’ sentiment tendency data into the priority ranking of design requirements and match corresponding technical solutions.

Results  The results of sentiment tendency analysis show that the proportion of negative sentiment among consumers for 6 topics exceeds 60% for each, including breathability, appearance, stylish and versatile design, size accuracy, elasticity and softness, and fabric comfort. Among these, the negative sentiment ratios for breathability and stylish and versatile design are as high as 93.28% and 93.90%, respectively, which are selected as the key topics for in-depth analysis. Through the transformation mechanism of QFD, the priority ranking of demands is determined as follows: breathability, appearance, stylish and versatile design, size accuracy, elasticity and softness, and fabric comfort. On this basis, combined with existing research literature and feedback from practitioners, the corresponding technical solutions are identified. Calculations of the priority of technical solutions indicate that the heat-zone breathable design, directional air exhaust technology, blended fabric processing technology, and antibacterial fabric composite technology achieve the highest scores (1.93 and 1.85, respectively), which are confirmed as the core improvement directions. Next, the integrated fly-knit molding process, double-layer shock-absorbing sole structure, memory foam shoe collar design, last design, and intelligent size recommendation technology are designated as balanced development technologies. Meanwhile, basic shoe style design and replaceable accessory schemes are classified as observation and reserve-level technologies. Additionally, to address design conflicts, such as those between breathability and fabric comfort, and between elasticity and softness and size accuracy, solutions including zoned structure design and gradient shock absorption are proposed.

Conclusion  By integrating BERTopic modeling, BERT-TextCNN sentiment analysis, and the House of Quality (HoQ) in Quality Function Deployment (QFD), a complete transformation path from consumers’ shopping reviews to design decisions is constructed. This path can identify the high demands of elderly consumers for attributes such as breathability and stylish design. The established framework of "topic clustering-sentiment quantification-demand mapping-solution ranking" features good structuration, reproducibility and extensibility. It not only provides clear optimization directions for the design of footwear for older people but also offers methodological support for the design optimization of other functional footwear and even wearable products. In the future, in-depth research can be conducted on the engineering trial production of technical solutions and user experience validation.