There is the possibility that TV operators could be squeezed out of the search game in the long-term, with management deciding that they do not need to worry about search, as this is a function that Google can do for them in a partnership. But they should be wary, as “Google is not a neutral player in the entertainment market.”
This was one of the observations made when Videonet hosted ‘The Content Discovery Leadership Breakfast’ in Amsterdam during September, supported by Gracenote. The event was conducted under the Chatham House Rule, attended by a selection of platform operators and content owners, and the role of on-platform and off-platform search revealed differing attitudes, depending on what role someone played in the value chain.
One Pay TV provider declared that a platform operator must be the single source of truth to viewers when it comes to content discovery on the platform. This means that viewers should not need an online search engine [e.g., Google] because they know they will find the content they want via the TV platform’s user interface. But a special interest streaming provider was all-in on search, including paid search, and this company translates the programme descriptions given by suppliers (which may assume that viewers already have reasonable knowledge about the topic) into SEO-optimised text that can be discovered more effectively via online search engines.
Search Engine Optimisation was separately cited as a huge factor for sports discovery – with an observation that pirate streaming services have excellent SEO. And a different Pay TV provider admitted that nobody can find content on their platform via search engines unless the operator buys paid search advertising – something they are now looking to remedy!
The psychology of content discovery, the critical role of unified search, what you do when promotions are persistently ignored, and the need to improve EPG and VOD imagery were among the subjects discussed. There were also insights on the importance of recommendation serendipity, coping with sports rights fragmentation, and dealing with missing or out-of-date EPG schedules.
What follows is a summary of the discussion. The attendees were asked to state what would help them achieve more viewing time (for their content or platform) over the next two years, and explore opportunities to improve content discovery or use content discovery to competitive advantage in the same timeframe. They were also asked to reveal their sharpest pain points in content discovery today and what challenges lay ahead, and consider where the content discovery value-chain can collaborate more effectively.
The points below are observations from at least one breakfast attendee that were tacitly accepted by the group – meaning that nobody argued against the view. Sometimes multiple people contributed to these points.
Helping consumers to avoid buying content they can get free
Given the cost-of-living crisis, this observation gets an early mention: good content discovery is not just about getting people to programming faster and ensuring a better match of what someone likes to what they see – it could also save people money (and no, not because they stay at home instead of going to the pub!). If consumers have better visibility of all the content they can watch free, it may reduce the need to subscribe to a new service or inadvertently rent a movie that they already had free access to.
The fear of being a television failure on Friday night
The psychology of content discovery was also addressed and there was much support for the assertion that viewers have a fear of failure when making content choices, especially when watching with someone else. Investing your time into content that you decide you don’t like is the worst thing that can happen on Friday night, one attendee declared. Linear TV still has value because it partially removes the pressure to make a decision, requires less effort (in discovery) and has an easier exit if you get it wrong.
Don’t bore people with promotions they keep ignoring
There was an acknowledgement that commercial imperatives may dictate some of the content that is promoted (where a platform or streamer has a new blockbuster Original that they own, for example). But one Pay TV provider was adamant that you should not “bore people” for weeks with promotions if they are not working.
After you have shown a title twice to a cohort, if there was a mere 2% pick-up rate (for example), continued promotion would create long-term fatigue and reduce trust in your recommendations, they argued. Try some optimisations, for sure (like changing the background image), supported by A/B testing, but if the response rate remains low, you should change the promotion, they advised.
Unified search is a key building block for super-aggregation
It was clear that unified search is a big priority for content discovery. An operator should not care where the content comes from, as long as someone is watching on their platform. This unified search needs to include VOD, SVOD, AVOD and FAST if you have them onboarded. Search is incomplete on many platforms today. Unified search is a building block for super-aggregation, as there may be dozens of places where consumers might find content they want to see. Good voice support boosts both endeavours.
Improving VOD/EPG imagery – including with personalisation
There was widespread agreement on the power of EPG and VOD imagery to improve content discovery and content take-up rates, with personalisation of imagery a goal worth setting. One speaker noted that empirical evidence shows that when you use images that suit the tastes of a viewer, there is an uplift in the number of minutes and different titles watched.
A piece of content like Game of Thrones can be characterised as both an ‘action adventure’ or a ‘love interest’ story and, depending on what someone has aligned with previously (based on what they have watched) the EPG imagery could reflect one of these characteristics. If someone has shown an interest in a particular actor, you can feature that actor in your imagery rather than a more generic image focused on the storyline.
This is summed up by the idea of ‘using the right picture language’. There was a suggestion that to help the moment of consumer decision making, simple phrases could be added to the content imagery. For example, the movie image includes the statement: ‘Critically acclaimed’.
Make it more likely that recommendations offer something new
On recommendations, there was an assertion that recommendations today narrow down the suggestions based on what the viewer has seen previously, and there should be more emphasis on offering something new. Filtering based on the attributes of the show itself will help and make it more likely that if someone wants to be frightened, or made to smile or cry, they will be offered programming that matches this need.
Granular programme descriptors will help in this endeavour, as these go well beyond genre and the cast list to flag factors like mood, scenario and character (so some films could be characterised as ‘heists gone wrong’, for example). This benefits both content owners and platforms, it was suggested, and if they use the same language [to help describe the content in greater detail] it helps surface content and connect similar content.
One speaker summarised how this could be applied in practice along with voice discovery, with a consumer saying to the TV: ‘Cheer me up’ and being given a list or programming with humour as a characteristic.
Keeping recommendations simple for service providers
One operator revealed that you can serve 90% of the customer base with acceptable recommendations based on dividing the population into just eight different profiles/personas. The other 10% of the customer base would require super-individual attention to meet their needs via a long-tail of minority use-cases. Another operator has been investigating how customer demographics can be used in the recommendation decision-making process (as part of a wider exercise that explores the use of demographics for advanced advertising). The challenge is how to get customers to opt-in with their demographic details.
Don’t be afraid to ask viewers: ‘What are you in the mood for?’
An attendee reported that when you start a healthy eating plan, the provider asks you what you want to eat, and what you cannot eat. “Maybe we should be just as explicit on television. We could ask: ‘How was your day today’, ‘What are you in the mood for’, ‘What would you hate right now?’”
High linear TV viewing may suggest your VOD discovery is weak
When one operator launched its multiscreen app, management expected VOD to dominate, but today 70% of viewing is either live/linear or catch-up, despite having a huge library that includes newer VOD titles. The data shows that linear viewers are also the most loyal. This operator decided these figures point to a weakness in VOD discovery and is looking to solve this.
Give content partners the brand prominence they deserve
One panelist reminded everyone to give prominence to the brands behind the content, as there is still enormous consumer awareness for what these brands represent. So, HBO and Discovery, for example, should be given strong brand prominence because people know what these media owners will deliver them. But brand prominence can get lost in the UI today.
Helping fans find live sports as rights fragment
The difficulty finding live sports is one of the biggest challenges for the content industry today, because for consumers it is one of their sharpest pain points, and the reason is the fragmentation of sports rights across more broadcast and streaming services. One round-table attendee, who is a rugby fan in the U.S., said it is a nightmare trying to find a rugby match you want to watch – to know what service it is on, and at what time.
There is inadequate metadata for live sport, and this is especially noticeable on streaming services. One major streamer could not confirm the start time for their football matches to a distribution / metadata partner.
There is also a need for better sports imagery for the user interface. It was suggested that we are due a redesign of how sports appear in the UI, moving from an action-based focus (as seen at one operator who proposed this) via carousels to a “more DAZN-like approach” focused on teams and players. Separately, it was suggested that the UI should offer more information about leagues, teams and players.
Filtering search to remove content in the wrong languages
There are opportunities to make content discovery more localised, even for operators already in-market. One operator said it has to respect multiple languages and ensure names of shows are displayed correctly in all languages. They are now working towards a search filter that will allow users to remove content that is not available in their preferred language, or which does not contain subtitles in their preferred language.
Handling EPG gaps and inadequate metadata
EPG gaps for live/linear television are a fact of life, and this will not change. Sometimes broadcasters do not know what they are going to fill the schedule with and sometimes they are in flux, as seen during the Russian invasion of Ukraine or the death of Queen Elizabeth II. An example was given of a channel owner that sends an operator a large EPG slot described only as ‘Children’s programming’, with no further detail on the programming within it, so the content cannot not be characterised or made discoverable. Sports metadata can be completely inadequate, and another example given was tennis content simply described as ‘tennis’.
Some improvement can be made by a resourceful distributor. One operator uses its own programmatic decisioning to predict what will go into an EPG position that is not detailed. If it knows that the Friday 7pm slot carries a drama, and it is halfway through the season, it makes a bet that the currently empty Friday 7pm EPG spot will be filled with this drama, for example. But this is not possible for all EPG gaps.
Informing out-of-market distributors about EPG disruptions
A Pay TV provider explained how a tennis match may be scheduled for 2 hours and 40 minutes, yet take 5 hours, disrupting the EPG. In the local market of the rights holding broadcaster, distribution partners are told about the schedule adjustments. But distributors outside the local market are not. This (out of market) operator only becomes aware of the EPG disruptions by watching the content feed and making EPG corrections. “It would be useful if they had a network to communicate the EPG changes to all distributors. Those changes could be communicated to everyone within minutes,” the Pay TV operator declared.
Limit image format customisations for the UI
There can be a lot of image formatting work when content owners launch onto new platforms and there was a call for standardisation to make life easier – or at least more flexibility on the part of distributors. One company was told by an OTT provider that it only accepted images in one particular format, otherwise they would not display correctly in their UI. A month from launch, every image in the supplier’s database had to be recast.
Very customised platforms are a barrier to agility. For a small content provider with limited resources, customisations limit the number of distribution end-point opportunities they can pursue.