The evolution of Pay TV operators into super-aggregators, and the expansion of BVOD services beyond catch-up TV into digital entertainment libraries with vast box-set collections, make it even more important that television companies master content discovery. During a recent Videonet webcast featuring Liberty Global, Swisscom, ITV and Gracenote, it became clear that content discovery has a direct impact on business transformation and the opportunities to differentiate services and monetise them fully.
Susanne Rakels, Senior Manager Innovation, Liberty Global, pointed out that her company sees itself as a super-aggregator “and wants to make sure consumers have access to all the content they want in one place, available at any time, across linear TV, on-demand, Replay and OTT including apps like Netflix, Amazon Prime Video, Disney+ and local video apps.”
“We are expanding our library further,” she added, “looking at podcasts, music and games. This could be an overwhelming amount of content unless it is made accessible easily, with content discovery done properly.”
The danger for Liberty Global and other operators acting as super-aggregators, as articulated by Rakels, is that you are constantly asking viewers to make choices – and there is simply too much entertainment to pack into the user interface and its sub-menus. That is why advanced discovery techniques become more important.
“Search and recommendation are not enough anymore – the recommendations need to be personal and contain surprises to avoid people always being presented with the same suggestions. Search has to be universal across all content sources,” she offers as examples of what ‘advanced’ means in practice.
Liberty Global is working on a new feature that will allow people to recommend content to friends and family using a QR code within the programming information that you scan with a smartphone. A link to the content is created and shared via social messaging. The recipient of the link can read about the series or movie and either watch it or add it directly to their watch list.
This function is being tested and the results are very positive. More than 60% of the test consumers across all ages intend to use it, rising to 75% for the younger demographic.
Another advanced technique is the use of voice commands to move around the UI, skipping the usual steps in a ‘menu tree’. A simple example is where you say ‘Netflix’ to launch the app. Liberty Global already offers a voice interface (“It is used a lot,” Rakels reports) and is working on the next important step: support for the natural language processing that will enable semantic search, allowing a more conversational and casual approach from viewers. This will make it possible for consumers to combine multiple search terms into a sentence, like ‘Romantic movies with Brad Pitt that are on tonight’.
As Rakels points out, this requires that the voice UI understands timing (‘tonight’), genre and actor within a single request – “and that is quite challenging”. Another step is to to add topics, like ‘dogs’ as a simple example, to the metadata about content, so that these can also be weaved into searches. Rakels would like to see an understanding of topics applied in this context, and to enable content with similar topics to be easily grouped together.
Swisscom is another service provider looking to expand its entertainment horizons as viewers watch content from more streaming services alongside ‘traditional’ TV and VOD. Fabio Marti, Senior Data Scientist at Swisscom TV, says: “We think of ourselves as a super-aggregator and try to collate as much content as possible on our platform, but as the content catalogue starts to grow, discovery of that content may get confusing.
“If we do content discovery the right way, we can give our customers the orientation and guidance they need for a simple and satisfying entertainment experience. A growing volume of content means it is more and more important that you have powerful content discovery.”
According to Marti, good content discovery boils down to good UI design, which means the user interface is as simple and accessible as possible, not showing too much of the now vast content offering, nor too little. Analysis by Swisscom confirms that no single discovery method suits all users, resulting in a choice of ‘entry points’ including channel zapping (which is still very popular among Swisscom TV customers), targeted content promotions and personal recommendations based on behaviour.
Some customers use search as their primary discovery method, going back into ‘Search’ every week to find the next episode in a series they have been watching for weeks already. A ‘continue watching’ content ‘lane’ is aimed at these users, and new episodes of a series are also shown prominently to them. For users who still search for the next episode, recently searched keywords appear so they do not have to type the programme name.
“Thematic orientation is important” Marti added, returning to content discovery in general. “A user might not know what to watch, but they may have a genre preference, so we have sections based on genre.” Like Rakels at Liberty Global, Marti is a voice fan. Swisscom TV offers voice search as a subset of its wider voice assistant functionality that was developed in-house. “Voice is getting more and more significant,” he says of this discovery method.
Steve Forde, Director of Digital Products at ITV, the UK’s largest commercial broadcaster, pointed out that viewing decisions in the box-set world are a high stakes game for media owners, since you could lose a customer for five, 20 or hundreds of hours if they get absorbed in somebody else’s programming.
Speaking as a streaming service provider that operates the ITV Hub ad-supported streaming service and the ad-free subscription version called ITV Hub+, Forde said it is ITV’s job to move Hub visitors as quickly as possible to the content they originally came for, and then find the right moments to nudge them towards new content – the latter opportunities being “few and far between”. All of this requires personalisation (“finding the right content to put in front of viewers and knowing what you should not put in front of them”), and good timing, all backed by data analytics.
According to Forde, “We see that most viewers coming into our services know what they want to watch, the vast majority of the time, so the content discovery piece is about getting them into their programme as soon as possible. Personalisation is a big part of that – we are second-guessing what they have come in for.”
This personalisation is partly about sidelining shows that someone may not like, as well as highlighting those they probably will. “We are a mainstream, populist broadcaster, and within our vast range of content, ‘One man’s meat is another man’s poison’,” Forde explains. Thus, soap operas will not be made prominent for users who are, statistically, far less likely to enjoy them. The ‘second-guessing’ of what people like is based on algorithms, which themselves are driven by an understanding of user demographics (which is known from the original user registration) and behaviour.
“If we know that someone likes soaps and they tend to watch Coronation Street on Tuesday morning, then when they arrive on Tuesday morning, we can place that show front-and-centre for that viewer,” Forde explains.
“There is a myriad of really simple things we can do, but they rely on the more complex stuff, like getting the right data in the right place and making sure it is usable across everything we do.”
Turning to the theme of nudging people towards new content once they have seen what they arrived for, Forde emphasises that you should not overwhelm users on their way in with content choices when they already know what they want to find. “Maybe drive some awareness of other programming on their way through – programming they did not know you had,” he advises.
“So, if they are coming in (to the ITV Hub services) for Love Island, make that show as prominent as possible on the user interface, but maybe the big ‘hero slot’ is not Love Island but a similar programme. That is how we can drive awareness that we now have a bigger library of content.”
This ‘bigger library of content’ is a reference to the fact that ITV is evolving its Hub services from a catch-up TV viewing experience into a digital VOD destination in its own right. That means more box-sets and potentially more viewing hours on the service, and that means that ITV must change the underlying content discovery architecture. “When you are a catch-up service, the whole system is orientated towards the latest episode of a show, with the viewer landing on Episode 6. But when dealing with box-sets, you need to start people at Episode 1.”
Meanwhile, catch-up TV is about episodic images that orientate a viewer to a specific ‘airing’, whereas moving forwards, ITV will be looking at programme-level imagery, Forde reveals. “There are technology changes, and also organisational changes relating to the content we show and what metadata and images accompany it.”
Forde and his team are very focused on understanding when someone is next ‘biddable’ with content suggestions and where the best places are to nudge them to their next viewing choice. “In some scenarios, the home page is not as valuable as a post-viewing recommendation,” he reveals.
In a world of increased box-set viewing, the pressure has intensified to ‘win’ that next viewing decision and avoid losing a viewer for what could be weeks. Viewer dwell time on the home page is one important data signal that someone may be ready to move on to new content. And there are obvious flags – like someone reaching the penultimate episode in a series.
“At that point, alarm bells should be sounding, and all your cannons should start firing: hero slots on the home page for content they might like, trailers, push notifications, social media marketing – all to say, ‘Here is something very similar to watch’.”
Simon Miller, Managing Director, International at Gracenote, pointed out that his company provides the tools that underpin many of the requirements for both standard and advanced content discovery – enabling content presentation within a visual UI or a sophisticated voice UI, or both, including through search, universal search (across multiple apps/streaming services in a super-aggregator scenario) and recommendation.
Gracenote has a huge database that gives every piece of content a unique ID that applies to the show or movie regardless of which service it appears on, or its format (linear, recording, on-demand, etc.). Miller claims the Gracenote ID is the de facto industry standard for identifying and linking content within the same platform and across platforms. Celebrities also have unique and persistent IDs, ensuring the actors or actresses are never confused with anyone else and always accurately linked to content they appeared in.
Gracenote provides the descriptions of content, and information about actors, genre, etc., and more recently it started to offer deeper video insights – what Gracenote calls ‘Video Descriptors’ that reveal the essence of a show or film. As Miller explained, this additional insight includes the mood of content (like ‘suspenseful’ or ‘electrifying’) and the theme (like ‘camaraderie’ or ‘a duel’) and identifies the kinds of characters involved, which could be anything from ‘superheroes’ to ‘sidekicks’ to ‘Mums’. The time and place where the show or movie are set is also revealed.
In a simple example of how this can be harnessed in practice, viewers can search for ‘suspenseful’ films in addition to films containing a particular actress.
One of the challenges today is ensuring the rapidly growing universe of streaming content, including content that is only found in streaming services, is mapped – provided with an ID and metadata attributes that link it to related content and make it easy to integrate into the content discovery processes of a super-aggregating platform operator.
“We have to keep a very close eye on the content ecosystem, as there is tremendous investment in streaming content today. We are proactive; we know the catalogues,” Miller explains. “We also listen to our distribution platform clients, who are striking [app or content onboarding] deals with streaming providers. We have to be reactive to a certain extent. This is a very fast-moving space, so you have to be agile and quickly add streaming video to the [content ID and metadata] database.”
Like other panelists, Miller emphasised the need for more personalisation in the quest to solve the ‘paradox of choice’ in a world of expanding content services. He defined the task as getting the right content to the right person at the right time – and timing is important in this decision making. “The same person may come in with half an hour or with three hours to spare. And on Saturday morning they want something different to what they would enjoy on Sunday evening.”
Advanced content discovery can be used as a selling point in consumer marketing, just as super-aggregation (which is underpinned by advanced discovery) is increasingly presented as providing all or more of the content that consumers want, in one place. Swisscom has featured voice search in consumer marketing, for example.
But what is the ultimate prize for platform owners and broadcasters who provide great content discovery? According to Marti at Swisscom, “We want to retain our customers and possibly make them promoters of our service, and good content discovery is a key building block for achieving this. The big opportunity is increased customer engagement and customer satisfaction – they are the key metrics in a saturated market like ours.”
Listen to the whole webcast
This webcast, ‘Advanced content discovery and engagement: How everyone wins’, also includes:
- A look at how programme artwork can be used more effectively to gain people’s attention, whether this UI imagery is updated to reflect what has happened in a reality TV show or the imagery is personalised to each user.
- The reasons why Liberty Global and Swisscom are so interested in the concept of topic-based search (with Liberty Global already deep into tests).
- What topic-based search is, and how it works.
- The content discovery KPIs that matter to ITV and Swisscom.
- The results of an audience poll that asks whether, in the next two years, advanced content discovery will be mission critical or not very important (or something in between).