Case Study: Meredith achieves rapid, human-level content classification and data-driven personalization with Google Cloud Platform

Meredith Digital Entering a new era of data-driven publishing

Meredith Corporation, the 115‑year‑old media company behind brands such as PEOPLE, Allrecipes, and Better Homes & Gardens, needed to scale a detailed content taxonomy and trend-detection process across more than 40 properties to deliver more relevant, personalized experiences. Manual classification at Allrecipes proved accurate but took years and couldn’t be replicated efficiently across the portfolio, limiting Meredith’s ability to spot emerging trends and tailor content at scale.

Meredith implemented Google Cloud AutoML Natural Language and the Cloud Natural Language API, training custom models on high-quality labeled data (including ~10,000 recipes) to automate taxonomy-driven classification and sentiment/entity analysis. The models matched human reviewers on accuracy, precision, and recall, reduced rollout time from years to months, and improved trend awareness and personalization—supporting stronger audience engagement and loyalty (Allrecipes has previously seen return visits +10%, homepage views +20%, and daily homepage visits +10%).

Meredith

Grace Preyapongpisan

Vice President Business Intelligence