Business Intelligence Strategies for the CIOSocial media analytics is the practice of gathering data from blogs and social media websites, such as Twitter, Facebook, Digg and Delicious, and analysing that data to inform business decisions. The most common use of social media analytics is gauging customer opinion to support marketing and customer service activities.
The first step in social media analytics is determining your purposes for social media involvement in the first place. Typical objectives include increasing revenues, reducing customer service costs, crowdsourcing, getting feedback on products and services, and improving public opinion of your company or products. Most enterprises have some combination of such objectives.
Once your social media goals have been defined, you should identify key performance indicators (KPIs) -- business metrics used to objectively evaluate specified factors. You could evaluate customer engagement, for example, through numbers of followers of your corporate Twitter account and numbers of retweets and mentions of your company name.
There are a number of types of tools for various functions in the social media analytics process. These tools include applications to identify the best social media sites to serve your purposes, applications to harvest the data, a storage product or service, and data analytics software. However, text analysis and sentiment analysis technologies are the foundational components of social media analytics. According to Pradeep Kumar, vice president and customer intelligence director at advertising firm DraftFCB, they make it easier to analyze unstructured data such as tweets and Facebook posts.
Although social media analytics has a lot of promise, the technology is far from mature. Sentiment analysis software is designed to target emotionally-charged or significant words in unstructured data as a means of evaluating, for example, customer reaction to a new product or service. The complexity of human conversation makes accuracy difficult, however. Here’s an example: Taken out of context, the word “sick” can’t be evaluated with any level of assurance. Although it means “unwell” in the conventional sense, it is also current slang meaning “cool.” One commenter could say “Sick new interface – big improvement,” while another might say “That new interface almost made me sick – so flashy I thought I was going to have a seizure.”
To be fully effective, a sentiment analysis tool would have to be sophisticated enough to include context in evaluations. The product would also have to be able to detect sarcasm, which subverts the meaning of an apparently straightforward message. Given that detecting sarcasm is difficult for many humans, software is unlikely to develop that capacity in the near future. As a result, Kumar and others say that users must interpret and validate the findings generated by social media analytics tools.
See also: business analytics (BA), social media marketing, social CRM, social networking, predictive analytics
Join the conversationComment
Share
Comments
Results
Contribute to the conversation