Case Study: Zignal Labs achieves 30% higher sentiment accuracy and 90% lower costs with Amazon Web Services

Zignal Labs Performs Next-Level Sentiment Analysis Using Amazon SageMaker and Amazon EC2

Zignal Labs, a San Francisco–based media-intelligence company, helps Fortune 1000 customers measure brand impact across the entire digital landscape. Facing the limitations and poor scalability of third-party sentiment tools, Zignal needed a more nuanced, reputation-focused sentiment solution that could analyze billions of stories and social posts in real time.

Zignal built a custom pipeline on AWS using Amazon SageMaker and Amazon EC2 C5 instances (Intel Xeon Skylake), with a distributed streaming stack and LSTM models retrained daily using human-labeled data. The new system processes more than 3 billion documents per month, improved sentiment precision by 30%, cut development and operations costs by 90%, and boosted customer acquisition and retention.

Zignal Labs

Andras Benke

Manager of Data Science and Innovation Labs