With more than 15 million monthly unique visitors exploring and adding recipes to Yummly each month, Yummly’s head of data, Sean McGlinchey, is responsible for ensuring recipe and ingredient data is captured consistently and accurately. “We apply proprietary technology to understand food and make recommendations built upon mapping the food genome,” explains Sean. Doing so requires Sean and his team to understand everything about a recipe and the properties of everything in a recipe: nutrition, diet, allergy, flavor and taste, techniques, prep and cooking time and more.
With thousands of recipes being yummed by users every day across the web, it’s critical that the Yummly team is able to identify new ingredients and patterns in recipes to continue to grow and improve the food genome. With Logentries, Yummly found a solution that enabled non-development staff to monitor and configure alerts for the recipe datastream to ensure continued recipe digestion, pipeline up-time, and new ingredient identification.
After evaluating several options, the Yummly team chose Logentries to monitor their recipe ingestion pipeline. At last, the Yummly data team could identify things like ingredient validation errors and pipeline downtime without tying up precious development resources. Custom tagging through Logentries enabled Sean to easily spot errors as they occurred. Identifying data sources using incompatible recipe and ingredient formats became automatic, enabling Yummly to scale with less friction as developers spent more of their time developing new features.
Download Case Study