Main results and implications
| Research aspect | Method | Result | Implication |
|---|---|---|---|
| Consumer-driven product insights | Content analysis (key-words, topic mining) Sentiment and emotions analysis Combination of text-mining and standard statistical analysis | Weak positive correlation between star rating and sentiment in the text Distinction between high and lower involvement products (more expensive) Negative relationship between the price and numerical rating Helpful reviews are those longer and those of more extreme ranking | Need to study text reviews also not only rely on star rating This result is especially important when a consumer is making a choice between different (similarly star rated) products, because in this case the likelihood of reading the reviews increases If price is taken as indication of involvement level, higher involvement evaluations are more sentimental and more negative The relationship is much weaker than in the case of tourism. This indicates that on average consumers are much less involved in appliances purchases To decide, people consider the extremes, average and longer reviews. The consumer on average reads up to 10 reviews. It is more likely that they will read those longer and those either with a high or low star rating |
| Price sensitivity and strategic pricing | Sentiment analysis, standard statistical methods | Price does not relate directly to quality (as evaluated using different methods) Reviews for more expensive products more critical | Emotional analysis complements the sentiment analysis well and is especially relevant in comparative analysis at both producer and product level. It can help identify brand trust or provide additional information (add explanation to otherwise numerical sentiment) about the quality of the product Due to the increasing reliance of consumer decision-making on online reviews, the relationship between the sentiment in the text and the price is an indication of producer and product competitiveness |
| Boosting corporate performance | Content analysis (key-words, topic mining) Sentiment and emotions analysis Combination of text-mining and standard statistical analysis | Consumers primarily evaluate functionality of the product The most common words are either nouns, describing the main features of the product, use or handling or (primarily) positive adjectives Two topics dominate: technical aspect and functionality | Evaluations are beneficial to producers, since they point to the products’ characteristics consumers care most about Firms should focus on enhancing the qualities of the functionality and should also stress these qualities in commercials, since consumers primarily care about those |
| Research aspect | Method | Result | Implication |
|---|---|---|---|
| Consumer-driven product insights | Content analysis (key-words, topic mining) | Weak positive correlation between star rating and sentiment in the text | Need to study text reviews also not only rely on star rating |
| Price sensitivity and strategic pricing | Sentiment analysis, standard statistical methods | Price does not relate directly to quality (as evaluated using different methods) | Emotional analysis complements the sentiment analysis well and is especially relevant in comparative analysis at both producer and product level. It can help identify brand trust or provide additional information (add explanation to otherwise numerical sentiment) about the quality of the product |
| Boosting corporate performance | Content analysis (key-words, topic mining) | Consumers primarily evaluate functionality of the product | Evaluations are beneficial to producers, since they point to the products’ characteristics consumers care most about |
Source(s): Authors’ own