Using Kamanja to Monitor Social Media

Social media has changed the way companies communicate with their customers. In the bad old days, customer contact strategies were largely one-sided. Marketers would package their key messages and distribute them via mass media ads and direct mail. However, when a customer had a point to make, they were limited to some fairly unappealing options. Have a question? Call our offshore call center and wait in a queue, or send in an e-mail and hope for the best.

Social media platforms allow customers to speak to companies directly, immediately, and publicly in real-time. By creating a platform to monitor real-time social media sentiment we can:

    • Actively engage in key conversations
    • Identify events requiring rapid reaction before they get reported through traditional channels
    • Shorten the time to resolve customer complaints

    One major retail bank has taken this to heart and built a solution with Kamanja. They started with unstructured data from Twitter, Facebook, and other sources. Then, using natural language processing and inference models, they identified the nature of customers’ public comments. In some cases, they could also identify the severity, based on the frequency and intensity of similar references. Kamanja processes all of that data as it comes in, makes a decision based on the model, and then routes the intelligence to a management dashboard. When appropriate, information also flows through to rapid response units within the relevant operations departments.

    At a high level, the solution looks like this:

    They are already using this implementation to detect geographically-specific problems such as ATM network outages. Acting on customer complaints more quickly is just a start, though. In the future, we see this company using the same technology to evaluate the effectiveness of viral media campaigns, and even to assess the ROI on some of the sporting events that they sponsor, making adjustments to their media strategy on the fly.