Quickplay – Streaming TV’s new power couple: Generative AI and CMS
Naveen Narayanan, Vice President, Product Management, Quickplay
The longtime challenge of content discoverability in streaming is impacting revenue generation and viewer retention more than ever as new services and/or new bundles entice users to stray. One way to discourage them from shopping for a new platform is by ensuring you’re serving up what they want to watch, most importantly when they themselves may not know.
Today’s content search experience
Search today is often prone to delivering inaccurate results, limited personalization, keyword reliance, time wastage, and navigation complexity. The challenge of content discovery for viewers has become greater in the age of abundant streaming platforms and vast content libraries. As the number of available options has grown, users often find themselves overwhelmed and may abandon the search before finding something suitable.
In fact, Accenture finds that 72% of consumers “report frustration at finding something to watch,” up six percentage points from the previous year. If viewers are watching less or, worse, watching on other platforms, revenue will be impacted. The time is now to invest in more effective ways to enable viewers’ consumption of relevant, available content.
The biggest challenges for direct-to-consumer streamers
Three trends in the streaming marketplace point to the importance of discoverability for streaming services:
- Competition for Eyes: The competition for viewers’ attention is hotter than ever as new streaming services come online. Growth is particularly strong in the free ad-supported TV (FAST) tier: Kantar reports that in Q3 2023, “adoption of FAST services outpaced video on demand streaming, two-fold.”
- Subscription Churn: Churn has nearly tripled over the past four years. Poor content discoverability is often a factor: Variety finds that services with the highest share of households that watch only one of their top 50 programs are more likely to see them stop watching after one month. And if a household isn’t using a service, they’re unlikely to keep paying for it.
- Content spend: In today’s high-interest rate environment, companies are tightening their belts — and content is on the chopping block. While Disney has gotten the most buzz over plans to cut content spending, other legacy media companies are also expected to trim spending, according to
By making it easier for people to find content that already exists, streamers are positioned for efficient revenue growth – capturing and retaining viewers without having to spend more money on content.
The industry needs advanced content discovery search tools and AI to expedite search and discovery and enhance consumer experiences. Large Language Models (LLMs) have emerged as a solution, offering natural language interpretation, contextual understanding, semantic grasp, conversational interfaces, reduced search time, personalized recommendations, and cross-platform discovery.
However, standalone LLMs face challenges such as lack of personalization and difficulty in adapting to streaming platform constraints. This is where the clever coupling of Generative AI and CMS really shows its symbiotic power.
Generative AI + CMS = content discovery advancements
Only 10% of consumers report that streaming services recommend content that is “very consistent with my past viewing and potential interest.” Moreover, with millions of titles available, the 2023 State of Play report from Nielsen’s Gracenote noted that streaming viewers are spending a record 10.5 minutes/session deciding what to watch.
Substituting viewing time for search time increases subscriber engagement, loyalty and monetization. This is where Generative AI combined with a powerful CMS can make a world of difference.
Generative AI can connect viewers with content they love, which is pivotal to the entertainment experience. It is proven to offer remarkable capabilities in generating diverse forms of content, including images, videos, text, and even whole television channels. It involves advanced algorithms that autonomously create new content based on patterns and data fed into the system.
The CMS serves as a streamer’s central repository, managing the catalog of content assets. It encompasses comprehensive descriptive metadata, such as titles, descriptions, tags, and moods, alongside business rules and any relevant content restrictions. It seems logical that these tools be integrated and cross-leveraged, however it is a massive technology undertaking.
Combining Generative AI capabilities with a CMS drives the creation of spot-on storefront rails and delivery of pinpoint-accurate search results, making it faster, and more accurate than ever before for viewers to find appealing content to watch.
With this combined approach, more informed, data-driven decisions are possible across the entire content management lifecycle such as:
- Augmenting content metadata with additional insights, including micro-genre descriptors, emotional tags, and embeddings to strengthen recommendations
- Determining natural content breakpoints to identify binge markers and ad-breaks that can improve usability and expand monetization
- Auto-generating content carousels based on themes extracted from the catalog or trending content across a variety of engagement metrics
- Personalizing content carousels based on individual taste preferences, content analysis, and consumption patterns
- Providing a natural language interface for assisted discovery, including catalog browsing, filtering, and cross-referencing assets.
Ultimately, the marriage of Generative AI and CMS simplifies the actions needed for both programmers and viewers. For programmers, it offers the peace of mind that their technology solutions are continuously learning user interactions and then adapt recommendations over time, contributing to a continuously evolving and personalized content discovery experience. For viewers, it dramatically elevates the entertainment experience by enabling them to more quickly discover content that meets their personal desires.
Why now?
As we approach 2025, the evolution of the video streaming marketplace will be amplified. More AI advancements and implementations will make competition not just stronger, but will require faster innovation to retain subscribers. The combination of Generative AI and a powerful CMS will equip the most future-thinking tier one service providers with the solutions needed to ultimately align with the evolving expectations of savvier users and viewers.
