Services: Contact Center – Social Media
Industry: B2C, Retail
01 | Opportunity
A large discount retailer was looking for assistance in monitoring, analyzing and reporting on the voice of its millions of customers mentioning their experiences with the brand on social media. With more than 300 million social sites, including Facebook, Twitter, and various blogs and forums, there is an abundance of information to sift through. An even bigger hurdle is that the client’s brand name is a commonly used word, causing even more mass amounts of excess, non-relevant information to sift through.
02 | Solution
To more accurately analyze the voice of the customer, the unnecessary noise of irrelevant posts needed to be filtered out first. The large discount retailer was introduced to Social Lift, a proprietary social media monitoring and engagement tool that uses predictive analytics to score data attributes and determine which social media posts are actionable and relevant. It then routes the data to the customer service agent who is best suited to engage based on post topic and type.
After gathering three months of human categorized data, the Social Lift tool was able to build a predictive model specific to the client’s needs. The predictive model analyzes the text within the new posts and compares it to the past text and then creates a categorization to make a prediction if the post is relevant or irrelevant. The predictive model built was able to sort social posts by relevancy, date and client-requested categories, routing the most critical posts to the social team first. The model was trained that certain types of posts, even though they contain the brand name of the client, were not relevant conversation based on surrounding text.
03 | Results
Dialog Direct was able to provide more accurate insights to the retailer with the irrelevant noise being filtered out. Dialog Direct’s predictive model was able to move from a 53% relevancy rate to a 94% relevancy rate in 12 months. This has increased operational efficiency by 41%.
