When Lifetime premiered the series You in September 2018, the response to the stalker-thriller looked far from favorable. The average viewership for each episode stood at a mere 650,000 – a performance dismal enough for the network to announce that the show would not return for a second season.
But when Netflix picked the series up in late December, You saw a sharp reversal of fortune. The New York Times reported that the show’s viewership over its first four weeks on the streaming service was on pace to reach 40 million. The show’s newfound popularity led Netflix to renew the once-doomed series for a second season – a dramatic turnaround that in many ways illustrates the ascendance of online video and the decline of traditional cable TV.
Indeed, more video content goes online in the span of 30 days than was produced by the three major U.S. broadcast networks in the past 30 years. With this colossal shift in how video is ‘broadcast’ naturally comes new patterns in how people connect with the content they consume. By 2022, online video will account for more than 82% of consumer Internet traffic, according to research from Cisco.
And if online video is revolutionising content consumption, Artificial Intelligence (AI) is revolutionising online video itself. How? AI is enabling more personalisation, better engagement, and unprecedented audience insights, with major implications for both how users consume content and the development of the content itself.
In today’s personalised ‘over-the-top’ (OTT) world, in which users can access the content that truly captivates them more easily than ever, AI is critical for helping publishers find the most impactful ways to engage consumers. By providing hyper-personalisation, AI can help drive up user engagement – and thereby boost revenue for publishers.
Netflix’s use of AI to curate content is a prime example. Leveraging user data – including what content users have watched in the past – the company’s machine learning algorithms generate personalised recommendations designed to direct users to content they’ll actually want to watch in the future. And it works: More than 80% of Netflix programming is discovered via the company’s recommendations.
AI can also be used to create and display content in engaging ways – for example, transforming once-static homepages into dynamic, interactive portals. Rather than scrolling through a flat page, Netflix users are able to view automatic live previews when they hover their mouse over the thumbnail for a particular show or movie. Instagram’s ‘Explore’ page – which aggregates content from across the platform for users based on posts they’ve liked and engaged with – has a similar feature. Countless other sites – from sports to news to weather – use captivating AI-generated teasers to draw visitors into watching their video content.
The bottom line: at a time when users are looking for more interactivity and new ways to explore content, AI is playing a key role in satisfying both demands.
For publishers, AI’s ability to derive actionable insights from big data is a tremendous asset in understanding what viewers want and thereby helping them develop content that will resonate. Every day, users generate copious amounts of data on their online video viewing patterns. By making use of this treasure trove of information and insight, publishers are better able to know what moves their audiences and to connect with them.
But AI can do far more than crunch user data – it can also analyse video content itself – and when its capabilities become even more sophisticated, the impact on content production will be immeasurable. At a very basic level, AI can analyze what is happening in a video: in a soccer game, for example, AI can detect whether a particular action is a goal being scored, an injury, a botched pass, or an egregious penalty, and utilise that information to create captivating teasers and highlight reels.
The true test for AI will be in its capacity to recognise the more subjective elements of content. Take the soccer game, for instance: AI can ascertain whether sports fans are more interested in viewing player injuries rather than goals during soccer highlights. With such insights, publishers can produce higher-performing, more compelling sneak previews. When these faculties are fully developed, publishers will have more insight than ever before into what is most interesting for viewers.
As the online video revolution proceeds apace, it will be fueled by more than just changing consumer preferences. It will also be propelled by the far-reaching capabilities of AI, which is allowing publishers to develop deeper understanding of user behavior and helping guide the development of content that will keep viewers watching. Our viewing experience will never be the same.
