Case Study: Leading US PC Manufacturing Company achieves automated multi-channel customer feedback classification and better segmentation with Blueocean Market Intelligence

Large technology vendor improved customer feedback capture from text data across multiple communication channels

Leading US PC Manufacturing Company faced a costly, manual process for classifying customer issues posted across technology forums and call-center transcripts and sought to automate text classification to approach the accuracy of human labeling. Blueocean Market Intelligence was engaged to apply its data-integration and text-classification services to consolidate and analyze the company’s disparate communication channels.

Blueocean Market Intelligence merged and cleaned 130,000 rows of forum and transcript data, built a comprehensive dictionary with synonym sets and business rules, and developed a multi-level classification model (six levels, with Level 5 containing 4,277 nodes). The first iteration achieved close to 60% accuracy, delivered a roadmap to extend classification to general blogs and forums, and enabled measurable business impact including better segmentation strategy, automated issue classification, and the ability to target marketing efforts using data science.