New Way Now: Roundel™ Media designed by Target helps ads hit the bullseye with Google’s Data Cloud

*Featured in this video:* Guthrie Collin, VP of Product Management at Roundel™ *Executive summary:* Roundel™, Target’s retail media network, is using Google Cloud to turn insight into action for brand advertisers while the shopping day is still live. Using tools like Google Cloud Lakehouse, Cloud Storage, Managed Service for Apache Spark, and BigQuery, Roundel modernized the Target Product Ads technology stack from the ground up, rebuilding its entire data pipeline. The new Lakehouse architecture brings together Target’s enterprise data in one place, cutting the time it takes to deliver ad performance data from two days to just two hours. Now, advertisers can optimize their campaigns while insights are still fresh. Delivering this enhanced performance for brands has fueled double-digit ad revenue growth for Target. This shift isn’t just improving ad performance but also establishing a strong foundation to power sophisticated AI — including conversational analytics and agentic workflows — grounded in the real-world context of the Target experience. *Challenge:* Advertisers need performance data within hours of a shopping event to optimize campaigns and ensure ads feel relevant and useful for Target guests. Roundel’s previous stack had to collect, prepare, and process terabytes of data fragmented across multiple systems. *Solution:* Using Google Cloud Lakehouse, Roundel created a unified lakehouse architecture, rebuilding its data pipeline to move from static reporting to real-time action. Cloud Storage and Managed Service Apache Spark ingest and process signals from external and internal systems nearly instantly, while BigQuery serves as a powerful analytics engine, allowing queries and AI models to run on the latest, most accurate data — without having to move or prepare it before activation. In addition, consolidating data in one place also simplifies governance, making it easier to implement policies and maintain the oversight needed to meet Target’s privacy and compliance standards. *Results:* With the new Lakehouse architecture, Roundel has cut ad reporting latency from two days to just two hours. By giving advertisers the immediate feedback they need to take action during the shopping day, Target realized double-digit growth in ad revenue generated by its new data capabilities. Even more crucially, this unified, governed data foundation is opening the door to new ways to interact with data and power even more sophisticated AI — from conversational analytics to AI-powered workflows. *By the numbers* → 2 hours of reduced ad reporting latency (down from 2 days) → Terabytes of unified data with Google Cloud’s next-generation Lakehouse → Double-digit growth in Target’s ad revenue → Same-day visibility achieved into guest shopping behavior *Key takeaways and highlights from our interview with Guthrie Collin, VP of Product Management at Roundel:* → “We actually rebuilt our entire data pipeline native to the Google Cloud. The most direct impact, our data freshness. We’re going from two days to two hours. That is an incredible improvement. Because advertisers are seeing more results, we’re seeing double-digit growth in the actual revenue we're collecting from the new data capabilities. And the best part is advertisers use that data to delight our guests, so everyone’s happy.” → “When data is generated, it lands in Cloud Storage almost instantaneously. Then, we use a managed Spark instance in order to process all the data. And the last part, BigQuery is our analytical compute engine. It allows folks to do ad hoc queries in near real time to solve special client needs and power our AI algorithms faster than we were before. This architecture shortens the distance between signal, insight, and action, and it's a really powerful driver to grow both the sales and the advertising results.” → “Before moving to Google Cloud, we couldn't even consider conversational analytics because our data was too fragmented. Having all the data in a single place allows us to make the artificial intelligence better. When you take generic models and you ground them in real contextual data, they drive much bigger results.” *Google Cloud products used:* Lakehouse, Cloud Storage, BigQuery, Managed Service for Apache Spark *Learn more:* → Roundel™ Media designed by Target: roundel.com