AI-based driver analysis with quantified impacts on brand perception
After a data centralization campaign, Teranalytics worked with the customer to identify several quantified brand perception metrics that could be predicted using machine learning. The rest of the data were used to build AI models that provided the best balance between accuracy, sensitivity, and specificity (measure by the area under the ROC curve). Simpler models gave slightly worse predictions with the advantage of being interpretable, and where used to derive a driver matrix. Each driver was related to a specific customer sentiment, with a precisely quantified impact on the outcome variable. The marketing group from the customer company only had to pick the strongest drivers and determine which marketing campaign to design to activate these levers and improve brand perception.
1. AI models ingesting both survey and transactional customer data
2. Quantified impact of marketing drivers lead to tactical recommendations
3. 360-degree analysis of customer behavior for more personalized marketing campaigns