Please note: we are currently updating this post to provide you the latest Salesforce Marketing Cloud capabilities (May 2013)
Lymbix specializes in sentiment analysis that mixes Natural Language Processing (NLP) with adaptive learning gathered from human-powered techniques, creating a power house system that provides a different layer of sentiment analysis than traditional sentiment.
Here are the different options Lymbix offers within the Radian6 Analysis Dashboard.
Similar to the way that most of us are used to seeing sentiment displayed, the General Sentiment insight will give you the breakdown of neutral, positive, or negative sentiment as it pertains to the social posts within your results. If you would like to understand more about the way that Lymbix sentiment stacks up, take a look at their blog post answering some customer FAQs.
Here’s the view of how General Sentiment will look in the Analysis Dashboard:
Dominant Emotion will give you the breakdown of what emotion is really driving the conversation, including those conversations that are coming up neutral. This is the first spot where we see a noticeable difference in Lymbix sentiment analysis. Instead of just classifying posts as positive or negative, as with General Sentiment, Dominant Emotion gives you an emotional breakdown of the posts.
How does Lymbix determine what emotion is dominant? Their system gives every post a score for the emotions you’ll see below; the emotion with the top score is the Dominate Emotion. A post might have another underlying emotion, but the Dominate Emotion is determined to be the driving the emotion behind the post. If you would like to get an idea of how this data is pushed out in to the LymbixAPI check out their documentation.
Here’s how the Dominant Emotion view breaks down in the Analysis Dashboard:
You might not always just want to see the Dominant Emotion. For example, there may be times when your product management team wants to pull intel around those people excited about your new features and you want a very specific emotional response to review. That is why the Lymbix Insights package allows you to break your conversations down by emotion, ranking the response for a particular emotion as High, Medium or Low.
Here’s an example of how that Emotion view looks in the Analysis Dashboard:
The emotions available for analysis are as follows:
Later this week we will take a look at a use case example of how Lymbix insights can be put to use in a real-life situation.
What type of sentiment analysis offered by Lymbix interests you the most? What emotions are you interested in using? How would using a different type of sentiment analysis improve your brand reputation reporting?