- AI and ML help with the “paradox of choice” of content
- Aim to increase value of video assets and remain relevant
- “Understanding the content is the key,” says Grey Media boss
Artificial intelligence (AI) and machine learning (ML) are key tools being used to create meaningful content experiences across a multitude of platforms to benefit user experiences (UX) whilst improving and streamlining the user interface (UI) and advertising models.
In an IBC365 webinar entitled Engaging users and boosting advertising with AI speakers looked at the cutting edge applications of AI and ML and how the technology is creating new ways to get more value from video content and engage the highest range of new audiences.
Amsterdam-based startup Media Distillery’ speaker - product owner Martin Prins - specialises in using AI to interpret what’s inside video.
Also taking part was Liberty Global senior manager discovery and personalisation Susanne Rakels, responsible for strategy and development of content discovery and personalisation across all markets at the media and telco giant.
They were joined by Grey Media founder and director Ruhel Ali who is currently working with the Bundesliga Digital Sports unit to help leverage the use of the cloud, video and statistics to improve the experience for football fans.
The way cutting edge AI and ML applications can transform the viewers content discovery and interact with content, increase value of video assets and enable operators to remain relevant were among key topics discussed by the panel.
Rakels explained the interest in AI for Liberty Global is the “paradox of choice” for consumers with too much content across its services the user experience and content discovery is fragmented. AI is being used as a solution to help personal content discovery across its 150 broadcast and linear channels, catch-up services and content from TV apps like Netflix and Videoland and short-form content from YouTube.
Citing research from Tivo, Ericsson and Digital Smith, Rakels said: “One in ten consumers can’t find something to watch anymore and spend one hour per day searching for that content.”
To mitigate the frustration from consumers, she added: “We have to invest in a personal content discovery and according to market research a proper content discovery experience can result in three times less churn and an improved customer retention of 140%.”
Liberty Global’s content discovery strategy is focussed on three points:
- Smart and simple UI with fewer layers and decision points
- Personal UI which is adaptive to individuals
- More relevant recommendations across all aggregated content
Liberty Global is working on a proof of concept to investigate the various approaches to offer personalised content and combining initiatives in a “hyper personalised” experience using the content data whilst ensuring privacy is the top of the agenda.
Enriching metadata
Media Distillery’s Prins explained that the company’s mission is to understand what “meaningful content” is about using AI to analyse real-time video streams including broadcast television and video on-demand.
Prins says: “We use algorithms from our deep-learning and ML and more traditional databases to extract speech to determine content including logos, recognise famous people, extract subtitles and with all that information we try to extract topics.”
The system has the ability to establish the start and end times of a programme and the advertising breaks based on the electronic programme guide (EPG) data to truly dissect the programme.
In practise this is to help inform broadcasters and content owners to help enrich the viewing experiences for consumers.
He adds: “This enables us to annotate per frame what is happening and using this information to allow people to search inside the content and search inside the content for certain topics ultimately to improve the user experience and boost revenues.”
Pointing to YouTube’s collection of user data, he said that broadcast and advertising agencies are asking how they can achieve this level of interaction by enriching user profiles to better target and offer more value to the viewer and advertisement for more value
Prins explained Media Distillery is currently testing this analysis with several operators to try and determine the high level topics in localised levels across content genres.
He says: “We are really excited to see what value this can bring.”
AI impact on UX
“Understanding the content is the key,” explained Grey Media’s Ali.
With the content landscape becoming more overcrowded with content from broadcasters, OTT, audio recordings, interactive streams
Echoing Rakels, he explained the level of confusion for users around the sheer volume of content discovery and the variety around various interfaces for content delivery.
He added: “Broadcasters and platforms now have TV guides and general EPGs that includes editorial metadata and editorial driven imagery,” however the recommendations have been driven by editorial feeds and UI apps, which are two “very different areas.”
Data points that operators can work with have traditionally been around awareness of audience usage, understanding where consumers go and what they do, “however that data has not been well used,” he said.
Looking at Netflix, Ali explained how they are setting the standard for personalisation and experiences based on mood with not many other operators competing in this level of personalisation.
He concluded the main concerns include tracking of users data in a secure environments and awareness of usage with privacy issues, but noted it does provide a way to give better recommendations for viewers away from tradition EPGs.
During the webinar, moderator High Green Media Rob Ambrose also addressed privacy issues, negative perception, countering the risk of overwhelming users and ways to advance personalisation with the speakers.
No comments yet