Over the past years, video consumption has increased across the board. Consumers have quickly gotten accustomed to having multiple SVOD subscriptions, in addition to a pay-tv subscription that offers live, catch-up & replay content. This also means that traditional broadcasters, TV operators and OTT entrants are competing more than ever for viewer attention.
We believe that this battle for eye-balls will not be won by the companies that provide the most content, but by the ones that offer the most engaging experience and help consumers quickly discover and watch the content that appeals to them at that time.
That begs a few questions: What is appealing content? And how do you make sure the viewer discovers it in as little time as possible? As we all know, there’s no one shoe that fits all: binging a new series, getting a quick news update, swiping clips on Instagram when you have a few minutes, or watching that 4 hours extended cut of your favourite superhero franchise. Some people want to see everything about their favourite sports team, others just want to be surprised with amazing content. Often it will be a mix of the above, which makes us think that successful video services are the ones that can cater to these dynamic viewer needs.
At Media Distillery we use an understanding of video to help our customers increase engagement and viewing time while minimizing the time looking for something of interest.
We believe video is a powerful but still underutilized source of actionable and valuable information. And with the speed AI tech is progressing, new opportunities arise each day that can help you increase customer engagement and boost viewership. Let’s dive into some examples.
Inherently related to non-scripted live content such as news, sports and entertainment: you don’t know in advance what the actual content is. As such, the program metadata lacks descriptiveness. For instance,a news program listing states “the latest international and national news and weather”, instead of what it is actually about: the Mars Perseverance landing, the latest updates on Corona, Formula 1 highlights, Champions League result, or Hollywood controversy.
This means viewers interested in specific topics, sports, athletes, etc, will not find the content these topics appear in, unless someone tells them about it or watches the content.
We built powerful tech to not only identify topics, people and (club)names of interest, but also spot the actual logical start and stop of a topic, to make sure you will only serve content if the topic duration is long enough. When we evaluated this on many news and sports channels, we saw that for 75% of the live programs we could identify topics not present in the program metadata, including breaking news stories: Tiger Woods’ car crash, Megan Markle, the rise and fall of the European Super League. Imagine how access to this topic data in your video platform opens up a lot of new opportunities for the discoverability of appealing content you already have.
Almost all video services provide trending content as a recommendation, based on viewing behavior on the platform. But what if you could also tie into trends outside your platform, such as social media? There is often lot’s of buzz on topics that relate to (live) TV or items that will also be discussed on TV. We can automatically match social media trends with appearances in your content, such that you can automatically point the viewer to this content. A great new way to automatically expose content people talk about (actors, sports athletes, sports results, matches) with the content in your platform you didn’t even know you had!
People talking about the Premier League? We can detect related content on all tv-channels, such that you can present it to all the fans. News about Meghan Markle or Nasa Perseverance? Automatically offer the latest updates as the story develops. And because we detect in real-time you can engage the viewer immediately.
Your audience may only have a few minutes to spare or may be interested in parts of a program. YouTube content creators can manually add chaptering, giving viewers a quick overview of the contents and let them jump to specific parts. Netflix is experimenting with fast-laughs, a way to preview byte-sized snacks.
So, you may have great content, but not always the right format to serve it to the viewer. That’s also where understanding of video comes in: we detect the start and stop of the different topics in programs, whether it’s the news, a talk show, or sports program. We can provide these tags as chapters such that users can jump to the parts they want to know about. Additionally, you can offer individual topics as snacks directly to the viewer. Both are great ways to increase engagement, utilizing content you already have.
The UI plays a pivotal role in helping viewers pick the right content. And we all know a picture says more than a thousand words. Attractive, descriptive images help the viewer quicker select the right content and lead to more (longer) viewing. For VoD assets episode-specific images are sometimes available, but for live content this is not the case, leading many platforms to resort to the same dull stock image, even for their replay windows. This can be a thing of the past as we are now able to create appealing high-quality episodic images from live content and provide them almost real-time, showing key elements from the show such as the main actors, while removing elements such as logos and subtitles. This change allows you to increase engagement with up to 15% over stock images.
With the great number of high-quality video services, the battle for more eyeballs is on. Offering seamless viewing experiences is therefore becoming a true differentiator in the market. A better understanding of content will help you utilize your content library to the fullest potential and engage your audience in the best way possible. In other words: less time browsing and more time viewing.
May 10, 2021
