Sentiment Analysis Using Web Analytics for CPG by TEG Analytics


Can Social Media Scores Predict Sales in CPG Effectively?
Business Question
  • Can consumer sentiment across various brand attributes such as quality, value, innovation be elicited through social listening?
  • Impact of sentiment scores on consumer purchase decisions
  • How can social listening guide tactical and strategic decisions?
pic2
Solution
  • Specific brand attributes were identified for each brand based on brand identity and category characteristics
  • Online conversations were scored for each brand attribute based on various lexical strings and relevance and intensity measures
  • Advanced machine learning techniques were used to quantify the relationships between sentiment scores and sales volume to capture whether sentiment is a leading indicator of sales
  • Impact of increase in sentiment and decrease in sentiment were checked for non-symmetric impact on direction and magnitude of sales
  • Factors such as price, promotion and seasonality were accounted for in the machine learning models

Robust Scores + Happy Tweets = High Sales
pic4
Result
Effective predictability of sales volume depends on the robustness of social listening, and sophistication of the scoring mechanism.

Leave a Reply

Hide dock Show dock Back to top
Loading