Recently, Dutch streaming service NLZIET examined the impact of Media Distillery’s content discovery solution on the user experience during a beta test. The results demonstrate overwhelming support for the expanded title descriptions, shorter chapters, topic-based upgrade of the search catalogue, and other newly added features.
NLZIET is a fast-growing Dutch OTT streaming service owned and operated by the country’s three major broadcasting groups (NPO, RTL, and Talpa), offering the biggest online collection of content from Dutch broadcasters. It provides live TV, catch-up (replay), and on-demand content.Recently, NLZIET and Media Distillery started a collaboration to explore the AI potential in the video industry. Shared values and goals make the companies “natural partners”, as explains NLZIET CTO Sander Kowenhoven.
“We both believe that innovation is the key factor for taking the quality of video entertainment to the next level.” - shared Sander.
The collaboration tackles the lack of descriptive metadata - a notorious industry pain point, particularly in non-scripted live content like news, talk shows, or reality shows. Listing only the large chunks of content such as economy or weather, these shows do not give an actual topic overview, for instance,. the World Cup or Formula 1. As a result, content loses its appeal to users.
Due to the prohibitive costs and time constraints involved in live broadcasts, TV service providers cannot create the descriptive metadata manually. As a result, users frequently cannot discover the content, so its full potential is not being realised..
Further on, Media Distillery and NLZIET focus on making the chapters visible in the play bar. This feature could help users to quickly see what the program is about and watch only the parts of the content they are interested in. With the successful implementation, TV operators can prevent churn among users with a preference for snackable content, popularised by TikTok and Youtube Shorts.
The third objective was easy and fast content discovery, so users can follow their topic of interest in the catalogue, and watch new content as it appears. Moreover, it comes in handy for those users who do not know the exact title of the show and can now find it with keywords.
“One day, our CEO said - ‘Hey, I’m watching this news programme on replay; why can’t I skip to the topic I’m interested in?’ - remembers NLZIET CTO Sander Kouwenhoven.“We decided that this was worth exploring as a great way to elevate the user experience: it could give our subscribers new innovative ways to interact with content, and help them find value in a broader array of content”.
This latest solution from Media Distillery generates granular content metadata, so viewers can easily discover a video on particular topics, and skip to the interesting parts through automated content chaptering and chapter title generation.
“We did consider other providers and technology solutions, but no one we spoke to could do what Topic Distillery does.”
The AI-powered Topic Distillery™ technology, automatically extracts metadata out of video, audio, teletext, and EPG, utilising Computer Vision, Speech-to-Text, and other AI/ML technologies. The real-time processing identifies topics and splits the content into logical segments, adding, on average, 21 topic tags per program.
For up to 82% of programs, Topic Distillery™ can add topics not present in the EPG data. After only 15 minutes from the end of the show, the results are delivered to customers, which can be used to boost search and discovery, recommendations, and enable chapter-based navigation. All this combines to a great increase in the utility of the content, the EPG, and improves the user’s experience of, and attitude toward, the overall service.
Unlike the current solutions in the market for creating video highlights, Topic DistilleryTM can automatically produce new assets and update its topic models. This helps users to discover content better, even the emerging topics that are not yet included in the classification system.
The first beta test evaluated the viewing experience of NLZIET users with the following features:
The end survey has revealed some interesting trends. For instance, while 85% of respondents said they manually scroll through the video timeline to find interesting parts, 90% would be interested in jumping straight to the segment of choice.
Furthermore, chaptering has been found to provide users with more control over the desired topics (72% of responses) and give a better idea about what the content is about (74%). More than 82% of respondents agreed that chapters help to access the parts of interest faster.
80% of the participants indicated an interest in finding topics within episodes (f.e. when searching for 'World Cup' you can find the corresponding chapters among all programs in the catalogue). 74% of responses supported a similar search functionality with personas. Skipping a series intro was a popular choice among 80% of the participants.
A staggering 92% of respondents said they would use the solution when it becomes available.
“Super handy, it saves flushing and often accidentally flushing too far.”; “I quickly find the subject I'm looking for and I can get to it quickly without having to rinse and search for a long time.”, “Easy scanning for correct fragments” - shared participants.
In addition to the successful beta test, Topic DistilleryTM already started gaining industry-wide recognition. The recent Video Tech Innovation Awards brought Media Distillery and NLZIET two wins in the categories of UX and Content Discovery and Advanced TV Service of the Year. cable, In addition, Topic DistilleryTM won at the renowned Cable and Satellite International Awards in the category Best use of AI or Machine Learning in video category.
With such a promising start, Topic DistilleryTM has all chances to help TV service operators to elevate user experience, nurture customer loyalty, prevent churn and provide subscribers with an ultimate smart viewing experience.
“With Topic Distillery, this content suddenly, just 15 minutes after live broadcast, has long-lasting value – it’s a new way of consuming TV.” - concludes Kouwenhoven.