When lacking consumer sentiment data, organizations resort to blanket promotional schemes, thereby not matching the true consumer expectations. Using a corporate blog as an eCRM tool, organizations can deliver promotional outbound messages to the consumers, and analyze the consumer responses on the blogs for effective campaign management.
Blogs are knowledge repositories which can be used as decision support tools. Every time an organization posts a message on its blog, the consumer receives some benefit through the posts—be it details about an organizational achievement, a promotional message about a product, request for feedback or response to a controversy. As these benefits increase, consumers increase their interaction with the organization, thereby offering more information about themselves. The consumer thought process as reflected in the comments under the posts is knowledge capital which can be mined to extract explicit information which can be leveraged from the organizational perspective for diffusion and exploitation and subsequent competitive advantage.
In A Troubled General Motors Blogs to Connect With Customers, I discussed how organizations could categorize content (blog post) as “Promotional,” which in this case would be equivalent to a promotional campaign or “Relational,” that aimed at building a relationship with the consumer.
The content categorization was taken further, by performing a factor analysis (based on responses from a set of consumers) on types of comments made by consumers on several blog posts. Studies were primarily conducted using the Corporate Blog of Southwest airlines as a base. This helped identify 24 different comment typologies, which loaded onto the factors of Consumer “Liking,” “Satisfaction” and “Involvement.” The consumer comments were further subjected to sentiment mining to extract a sentiment score for each consumer and then relate it to the consumer’s relationship state with the organization.
Tracking of consumer sentiment through a blog
Considering comments as sets of opinionated text, with the assumption that the text (each set of comments on a single post) is related to a single issue or item , it was interesting to see that the consumer opinion was either positive or negative or featured somewhere on the continuum between these two polarities. This was done by converting each comment into a feature vector by using a text processing tool and then identifying the sentiment bearing features.
By using a sentiment mining tool, where each opinionated word had been allocated a sentiment score on the basis of its wordnet synset, (Esuli, Sebastiani), a sentiment score was calculated for each individual comment. In this context, term occurrence was used as an indicator and not term frequency because in traditional sentiment classification, increased term occurrence does not emphasize/change the sentiment polarity. Further, considering the algebraic sum of the term orientations as representative of the sentiment behind the comment, the scores were calculated. It was important here to correlate each term to the correct wordnet synset it belonged to, as that held the key to the score.
By tracking the consumer sentiment for individual consumers and tabulating the individual sentiment scores, the following observations were made:
- From the set of consumers studied, 17 percent of consumers displayed negative sentiment and the remaining 83 percent displayed positive sentiment. Out of these, 57 percent displayed a sentiment in the range of 0 and .5. The mean of the entire population hovered around a sentiment score of 0.23.
- While consumer sentiment for individual “campaigns” (posts) was a function of the respective campaigns, no correlation was observed between no. of words per post and the sentiment score of the campaign.
- It was interesting to note that there existed a positive relationship between the sentiment score of a relational post hosted by the organization and the mean consumer sentiment in response to these posts.
- A further correlation was observed between the sentiment score depicted by the comments on relational posts and no. of posts per month.
Organizations can use their blogs for better campaign management. Consumer segmentation can take place on the basis of a consumer’s position (by virtue of his sentiment score) on the continuum between liking and involvement and strategies for consumer targeting can be formulated accordingly. Consumers displaying negative sentiment can be handled appropriately and redressal organized. Mean of individual consumer sentiment scores on an individual post can give the organization “some” input into the success of a particular campaign (as represented by a blog post).
Further, in times like these, maintaining positive thought by the organization through “relational” posts can help consumers maintain a positive outlook, too. Frequent communication with a consumer through higher postage frequency can further contribute to the consumer relationship.
Interpretations of CRM terminology and thoughts over utilization of the web for maintaining better consumer relationships may vary. The final objective remains the same: aid marketing in delivering higher profits through better consumer targeting and help retain existing consumers. Maybe blogs can help you peek into the thought process of your consumers and identify clearer strategies for acquisition and retention!