Roku Patent Balances Ad Output as Streaming Industry Plateaus

Roku wants to get the most bang for its buck. 

The entertainment tech company filed a patent application for a “reinforcement learning model for optimizing long term revenue.” Roku’s system essentially aims to get as much revenue as it can without ruining the user experience. 

“By increasing the number of advertisements, as one example, the revenue can be increased,” Roku said in the filing. “However, the increase in the number of advertisements can result in poor user experience for the users.” Roku’s tech relies on AI to help it strike the right balance. 

This system monitors the “user state” — the content users are currently watching, as well as profile information such as demographics, tenure, and activeness on the platform — to “personalize user actions based on different tolerance of the advertisement.” Along with tracking this information, Roku’s tech also determines the “revenue value” associated with the user. 

The information is then fed to a reinforcement learning model, which gradually and subtly adjusts advertising output to avoid a sudden increase in ad activity that may cause user churn. The changes could relate to the type, frequency, and timing of ads. 

This method is better for “capturing delayed rewards,” such as long-term revenue from increased watch time and user retention, Roku said. 

This isn’t the first time Roku has taken an interest in making ads that don’t turn off viewers. The company has previously sought to patent a system for “emotional evaluation” of content as a way to guide ad placement without violating users’ privacy. It’s also researched ways to use generative AI to personalize content that could better tailor ads to user preferences.

Roku’s recent earnings report highlights its hot streak with ads: The company’s platform revenue unit, which includes advertising, rose 11% year-over-year to $824.3 million. The company noted in its earnings release that its ad performance outperformed the overall market. 

However, the streaming industry is having some growing pains. While Netflix remains the king of the market as it continues to gain subscribers, tons of streamers are cutting costs. 

Using AI to tailor ad content and delivery to the individual could help keep users engaged. Though the market for ad-supported streaming is getting increasingly crowded, recent surveys from the Interactive Advertising Bureau and the Video Advertising Bureau both indicate that the majority of Americans would rather watch ads than pay for subscriptions or premium tiers.