Demand Manager’s New Era of Automated Yield Management
Publishers are confronted with an endless number of yield management choices daily that can materially impact revenue. The guesswork of wrapper optimization is one of them. The intricacies of managing Prebid configurations can cause publishers to hemorrhage time and money and publishers struggle to understand what data is meaningful and which wrapper settings warrant adjustments. Moreover, trialing wrapper settings while balancing page performance poses its own set of challenges.
Enter Machine Learning Wrapper Optimization, Demand Manager’s new feature that gives publishers the tools to better optimize wrapper settings. Engineered to streamline wrapper optimization, this next-generation A/B testing tool offers Recommendations to improve performance and drive revenue outcomes.
While Demand Manager’s A/B testing feature laid the groundwork for wrapper optimization, this opens the doorway to greater innovation with a focus on efficiency and control for publishers. The goal is for publishers to be able to optimize wrapper performance and generate more revenue without having to manually “do” anything. The suite’s enhanced automation tools signal a new era in yield management – powered by machine learning.
Machine Learning Wrapper Optimization drives automated improvements that serve publishers in numerous ways:
Revenue gains: Initial testing showed that 80% of wrappers that ran a machine-generated experiment saw an immediate increase in revenue compared to the existing setting.
Seamless integration: A user-friendly interface allows publishers to activate Recommendations within their existing setup, and the toolkit is designed for ease of use.
Publisher control: Publishers are empowered to leverage the full power of Demand Manager’s Analytics Suite, including its in-depth Core Web Vitals information.
Here’s a deep dive into Demand Manager’s breakout tool. Key features include:
- Best-in-Class A/B Testing: Publishers can test and measure the impact of their Prebid integration against the entirety of their ad stack. In real-time, publishers can experiment to see expected revenue increase based on Prebid and GAM auction data and session data, surfacing opportunities for new revenue. Publishers may create up to five experiment wrappers and compare them.
- User-Friendly Interface and Comprehensive Workflow: Publishers never have to leave the Demand Manager UI to implement Recommendations; the tool is already integrated into the interface’s established workflow. With traditional experiments, it takes significant time to surface new revenue opportunities. Now, publishers using a single click within the Demand Manager UI, can implement experiments and translate insights into projected improved revenue outcomes. A key component of the tool is that it’s designed to have minimal to no impact on the user experience. Page performance is preserved by recommending optimal settings to increase revenue without adding to latency.
- Customized Yield Optimization: By default, publishers can target wrappers by device type, browser, location, and domain. If publishers want extra criteria like page type, site section, or visitor data, they can define it in the Demand Manager UI. They can also configure their page to send this data via the wrapper request for custom Key Values. This toolset combines sophisticated machine learning algorithms with existing industry-leading capabilities for publishers to manage their Prebid stack.
The Big Impact of Small Changes
By infusing A/B testing with machine learning, Demand Manager’s new Machine Learning Wrapper Optimization makes it easy for publishers to improve revenue and maintain page performance. At Magnite, we look forward to finding more ways to scale this technology to bring published revenue-driving tools that also save time and resources.