Case Study: Leading Social Media Company achieves large-scale, high-quality training data and on-time product release with Appen

Our ability to deliver training data across a diverse set of users in a tight frame while maintaining a high level of quality was a key success factor for this project

Leading Social Media Company needed large volumes of labeled training data to improve a natural-language tool that identifies user intent, sentiment, and entities, but its own user-generated content was insufficient and the project faced a tight internal deadline. The company partnered with Appen for crowd-sourced training data collection using Appen’s rater pool and internal data-collection tools to capture both clear examples and ambiguous/noisy phrases required for robust model training.

Appen recruited hundreds of participants within days and, in less than two months, collected more than one million diverse samples across categories such as transportation, events, movies, and sports. The data — rich in regional, demographic, and linguistic variation — enabled the Leading Social Media Company to improve its help center, ads, videos, and ML models, release the product on time, and control project costs compared with other data-collection methods.