Whole-Page Optimization: The Next Frontier of Personalization for Streaming Home Pages
By Manoj Yerrasani, Vice President of Data & Personalization, Global Platform Engineering, NBCUniversal Media
Introduction
For years, personalization in streaming platforms meant recommending the right titles to the right users. Today, that definition is expanding. My teams earlier at Warner Bros. Discovery (WBD) and now at NBCUniversal (NBCU) have pioneered a broader approach called Whole-Page Optimization (WPO), an evolution that transforms how streaming services like HBO Max and Peacock design, assemble, and adapt their home pages in real time for every viewer.
Instead of focusing solely on which shows to recommend, WPO personalizes how the home page itself is composed: from module placement to artwork hierarchy, creating a living, dynamic interface that adjusts to each user’s intent and behavior.
The Challenge: Personalization at Scale Across Global Brands
One of the defining experiences of my entire career was helping merge two major streaming organizations: Discovery and Warner Bros. Each had its own personalization and data stack, engineering culture, and global audience. The goal for the launch of Max (Now HBO Max) was clear yet technically daunting: move to a unified tech stack including a personalization framework that could deliver tailored home pages for tens of millions of users while maintaining speed, resilience, and creative control.
After Max’s successful rollout and global expansions, I continued the vision at NBCUniversal, where the challenge grew larger: unify data and personalization for Peacock and international streaming brands under Sky Showtime, Now/Wow, and Show Max. The result was an enterprise-scale personalization engine capable of serving hundreds of millions of viewers across multiple regions, languages, and content catalogs.
Whole-Page Optimization represents a paradigm shift, turning the home page from a static shelf into a personalized, living experience.
What Whole-Page Optimization Is
Whole-Page Optimization redefines personalization by expanding it beyond content ranking to layout decisioning. Key components include:
- Dynamic Module Orchestration: Instead of fixed carousels (“Continue Watching,” “Trending Now”), the system selects which modules appear, what content they offer after a required level of de-duplication and in what order, based on each viewer’s engagement signals, history, and current intent.
- Adaptive Layouts: The visual hierarchy, row placement, and even thumbnail styles are dynamically generated to optimize for engagement and discovery.
- Unified Feedback Loops: Real-time data from playback, clickstream, and content metadata continuously retrain machine learning models that decide homepage assembly.
- Cross-Functional Signals: WPO integrates editorial, marketing, and recommendation inputs, balancing business priorities with viewer relevance.
The entire home page essentially turns into a customized “surface,” constantly learning and adjusting to user activity and real-time events.
Engineering the Framework
The implementation required a deep fusion of data platforms, machine learning, and high-availability engineering.
At WBD, we built a shared personalization layer on a scalable microservices framework that unified real-time streaming data with recommendation models. The architecture supported A/B and multivariate testing at scale, allowing us to quantify engagement lift from new homepage variants.
At NBCU, we extended this with a Streaming Data Platform that powers analytics and other platform operations, including personalization for all NBCU streaming brands. The new WPO system operates in less than a quarter-second response times and re-ranks or reconfigures home pages dynamically. This is critical during tent-pole events like major live sports streaming or popular movie launches on the streaming platform.
During pilot rollouts, users exposed to WPO-optimized recommendations showed a significant increase in watch time, play rate, click-through rate and notable improvements in session duration, which are clear indicators of enhanced user engagement.
Key Learnings for the Industry
- Treat Layout as a Personalization Surface: Don’t stop at content recommendations; layout and sequencing are equally powerful engagement levers.
- Unify Data Early: Consolidating streaming data across legacy systems is essential to drive consistent personalization logic.
- Design for Global Scale: Plan for multilingual, regionalized, and brand-specific variations from day one.
- Integrate Editorial & Algorithmic Judgment: Machine learning should complement, not replace, human creativity and content strategy.
- Measure Holistically: Beyond watch time, play rate, CTR, track diversity of consumption, cross-genre discovery, and retention impact.
The Future: From Predictive to Generative Personalization
The next evolution of WPO lies in Generative AI. Future personalization engines will not only select and arrange content modules but will also generate adaptive text, thumbnails, and summaries on the fly. As viewers move between mobile, TV, and in-car interfaces, home pages will reconfigure in real time, powered by multimodal AI that understands both content semantics and individual preferences.
The boundaries between recommendation, UX design, and storytelling are blurring. Streaming platforms that embrace this convergence will define the next era of audience engagement.
Conclusion
Whole-Page Optimization represents a paradigm shift, turning the home page from a static shelf into a personalized, living experience. My journey from launching WPO for Max at Warner Bros. Discovery to scaling it across Peacock and Sky brands at NBCUniversal has shown that personalization’s true potential lies not just in recommending what to watch, but in intelligently designing how audiences discover it.
In a streaming world defined by choice overload, WPO offers clarity, relevance, and delight, all driven by data, guided by design, and powered by AI.
About the Author
Manoj Yerrasani is Vice President of Engineering, Data & Personalization at NBCUniversal Media. He leads data and personalization technologies across NBCU’s streaming portfolio, including Peacock and Sky brands. Previously, he led search and personalization service engineering for Warner Bros. Discovery’s HBO Max launch. With over twenty years of industry experience, Manoj’s work focuses on innovation at the intersection of AI, data platforms, and media personalization.