Power to the people – personalising the TV experience

Pay TV operators are using personalisation as a key weapon in the fight to attract and retain customers. Graham Pomphrey reports on the latest technologies and business models available to them.

The average pay TV viewer is faced with a bewildering choice of content. Three hundred linear channels is not uncommon. Neither is a VOD catalogue running into thousands of hours of content. Add a few broadcaster catch-up platforms and some OTT video content and the options available to the consumer are staggering.

With this in mind, how can operators ensure that their customers are getting the content and services they want? It’s becoming a key issue in the battleground for customers. Those operators that can best tailor the service towards individual customers are most likely to keep hold of existing subscribers and attract new ones.

Starting point

Personalising TV services is not something new – operators have for years offered packages of channels, usually based around general areas of interest like sport, music and entertainment. Another basic step taken by some operators is to allow customers to create a list of favourite channels, saving them from having to scroll through the EPG to find relevant channels. But the advent of broadband-connected networks means the options available to service providers have suddenly increased.

In the face of competition from newer entrants, especially innovative IPTV operators, more established players have begun rolling out advanced platforms with IP connectivity to offer greater interactivity, using the internet to not only deliver content and ads, but also to glean data about how people are using the services. In the UK, Virgin Media recently rolled out an advanced OTT DVR platform powered by TiVo.

“Content online, on TVs, on mobile and on-demand is hugely exciting but consumers are facing increasingly complex choices and it can be difficult to find what you’re looking for,” says Virgin Media’s executive director, commercial, TV and online, Alex Green. Which is perhaps why the top five channels on the cabler’s EPG – BBC1, BBC2, ITV, Channel 4 and Five – take 56% of viewing share. But with TiVo, Virgin Media is likely to see this figure change as consumers become accustomed to the platform’s search and recommendation engines. Alongside recommending content based on viewing preferences, viewers can search across TV listings, seven-day catch-up, on-demand, favourite actors, future shows and online content all with one simple search.

“With rich programme details our TV customers can discover great new content based on their favourite actors, directors and more, and we’re able to deliver all of this through a seamless, easy-to-use interface,” says Green. “The service also features editorial collections, allowing customers to browse themed or seasonal programming according to their mood.”

Virgin Media TiVo can recommend content based on individual viewing habits. The system works by enabling users to give programmes one, two or three ‘thumbs up’, or a ‘thumbs down’ if they don’t like it. Over time, the set-top box will begin to understand the type of shows the user enjoys the most. “It’s this ability to understand individual preferences and personalise the viewing experience that will enable viewers to make the most of the new digital world,” says Green.

Twitter doesn’t just bring up the most obvious shows so it tends to pull people further down the EPG.
Tom Weiss, TV Genius

It’s still early days to ascertain just how customers will embrace these types of services, but there is some recent evidence to indicate that they are being well received. Virgin Media revealed that one in every four views of a TV channel on TiVo comes from a source other than the EPG – a sure sign that the search and recommendation engines are proving popular. Customers have also been embracing apps, with almost 80% of households having used the platform to access an app, with each box launching an average of 4.5 apps each week.

While Virgin Media and TiVo is a good example of an established pay TV operator using advanced technologies, and most importantly an IP return path, to enable customers to personalise the service to a greater extent, it is newer entrants that are not shackled by legacy networks and set-top boxes who are using that advantage to launch these kinds of services.

IPTV operators in particular have been keen to embrace advanced user interfaces and that makes personalising the service for customers much more achievable. And for newer operators, being able to differentiate the service can be crucial if they are to have a serious go at taking on much larger and more established providers. In the UK, IP Vision’s DTT/IP hybrid service Fetch TV launched in 2009. As well as offering the 50 channels available on the country’s Freeview platform, it aggregates in excess of 3,000 hours of on-demand content, as well as offering the BBC’s iPlayer catch-up service and pay TV operator BSkyB’s Sky Go multiscreen service. Commercial director David Bloom says that while IP Vision might not have the marketing budget or customer base of established pay TV operators, it has always had more flexibility in terms of being able to offer its customers as much choice as possible.

“To tell someone that you can choose an individual channel, or a package of channels for 30 days and then drop it, like we did from the start, was quite advanced. It is an early example of how we personalised the service,” he says. “Also, with the user interface effectively being delivered from the browser side of the service, we were able to see exactly what our customers were watching, which meant we could serve up our own promotions in the EPG or in the mini menus that could be targeted around the time of day, for example.”

Bloom says the company is starting to experiment with targeting content towards more defined segments like serving up pre- and post-roll adverts for videos based on the genre of the selcted content. The company is aiming to strike a balance between handing a degree of control to the user “in an intelligent and subtle way” whilst also building profiles around users to better target content to them. “Its about getting close to the customer but not too close,” he says. “We’ve been brought up on the idea that we are scheduled to so we need an evolution of that, which means giving our customers a bit more flexibility but within a structure. I think that’s where the newer players in the market have an advantage because traditional pay TV operators have less flexibility to really try things out. They feel under pressure to get it right first time and they don’t want disgruntled customers if they don’t get it right.”

Technology choices

While technology vendors have developed a range of solutions to help operators to deliver more personalised experiences, especially in the field of advanced search and recommendation platforms, operators are in many cases still wrestling with business models. “It’s about money at the end of the day,” says Matthew Huntington, vice-president, product marketing for TV technology company Nagravision. “Operators are trying to personalise services to retain customers or increase APRU and margin. When a piece of content is early in its release window, studios will generally take a 60-80% cut but that will reduce to around 40% over time. Through better recommendation and personalisation, operators can drive consumers to longer tail content further down the release window and therefore make better margin out of it. People are only going to find that content if it is promoted to them either through consumers searching for it or service providers making recommendations to them.”

TV is also unavoidably affected by the impact the internet is having on the way we consume media. In a world where we can pick and choose who to follow on Twitter, who to be friends with on Facebook and which music and videos we have on our iPods, TV viewers are likely to become more discerning in what they see on the screen and how it is displayed to them.

“We have an expectation now of being able to personalise our interactions with our electronic devices and internet services, even if it’s something as simple as being able to change a colour scheme,” says John Griffiths, solutions marketing manager at NDS. “With the internet, you get to see exactly what you want. Why must the TV be any different? The technology and power inside set-top boxes is there and we are increasingly living a connected TV world that means we are capable of fully personalising the service.”

Griffiths sounds a note of caution about using the internet as the basis of attempts to personalise the TV experience, however. “In reality, is there any value in being able to change the colour of the UI? Not really. The TV is a different experience to the web. It isn’t instinctively as interactive for a start.” But, he says, get the user experience right and operators could find ARPU increasing along with brand loyalty.

According to SeaChange’s senior vice-president of advanced technology Steve Davi, in order to get the most of the content assets available to an operator, they need to enhance the way they promote that content, improve the advertising models around it and offer customers the chance to better personalise the service.

[icitspot id=”16508″ template=”box-story”]In terms of personalisation, SeaChange has developed Nitro, a user experience product based on HTML5 that be used across multiple platforms. As well as integrating with recommendation engines, it also allows subscribers to develop favourite channels made up of personally selected content and then share it with their friends via social media networks. At IBC, the company plans to demonstrate its ‘virtual party’ – the ability for a user to set up an ‘event’ on Facebook based around an on-demand title from their pay TV operator’s platform, invite friends to the event and control when the content starts and stops.

Davi believes the future of personalisation will lie in companion devices, the main benefit being the ability to personalise down to the individual user rather than a household:  “Tablet devices and smartphones offer many capabilities for operators – they are personal devices that are designed to be interacted with. It means operators can target more effectively and get better feedback from their customers. They can also target advertising more effectively. A tablet will change the way we watch TV and I think that this is how people will browse what to watch on the main screen.”

According to David Allred, chief marketing officer of Sezmi, a US-based hybrid TV technology company, operators that are offering TV Everywhere services need to ensure a consistent user experience on each device and for each content source: “Advanced personalisation is more than set of isolated recommendations; it drives the entire user experience. Key elements for personalisation that need to be present in next generation platforms include providing a consistent experience across multiple devices and across multiple content sources.” And Allred said this means using a cloud-based architecture. Consumers, he says also want to be able to use the same user interface to search and browse all of their content regardless of source. “A return path is also a critical element of advanced personalisation. With no feedback loop, the operator can’t deliver individualised recommendations,” he says.

Search & recommendation

Virgin Media’s statistics are clear evidence of the growing role search and recommendation engines will play in enabling customers to discover relevant content.

From a technology point of view, basic search functionality is fairly undemanding to implement, says NDS’s Griffiths, especially when it is confined to the operator’s EPG and VOD catalogues. “Search is pretty straightforward. There are clients within our middleware that means we can do it in the set-top, and using the headend we can search across the whole linear and on-demand catalogue, and potentially to external sources in order to provide search results.”

TV Genius, a company that specialises in search and recommendation, counts companies including BSkyB, ITV and Freeview amongst its customers. According to CEO Tom Weiss, search is becoming much more complex, with the advent of HD channels, time-shifted channels and catch-up services. When someone launches a search it is becoming increasingly complex to establish which content they are actually looking for. “People are searching more for multichannel content than they were previously and it is hard to predict the shows people are looking for. We have to deal with the same show being available on multiple channels, and broadcasters airing repeats. If someone searches for a soap, for example, there is the catch-up from the previous night, tomorrow’s episode, not to mention re-runs on other channels in the afternoon and so on. We have to be able to bring the right results to the top.”

Weiss says TV Genius’s algorithm can take between 20 to 30 different factors into account, including how many people have watched a piece of content, when it was broadcast and on which channel it was originally aired.

The company was recently acquired by broadcast services provider Red Bee Media, which has a rich heritage in providing metadata services to broadcasters and pay TV operators. The deal will see Red Bee combine its data services with TV Genius’s search and recommendation.

“We see a lot of content from lots of different sources trying to compete with each other. We know from experience that people tend to like some form of curated viewing experience. Using technology to take the strain from the end user to find something interesting amongst all this content is important for service providers,” says Red Bee Media’s director, technology and innovation Steve Plunkett.

The company provides broadcast representation, meaning it writes descriptions that appear in EPGs and print publications. “It’s a means of promotion and signposting users to content they might be interested in, and that is an early precursor to what we’re seeing now from an algorithmic perspective in terms of recommendation. As technology plays a bigger part in that process we felt we needed to have a recommendation engine and a TV-centric search facility and something that understands where newer influences are coming from, such as social networks,” says Plunkett.

TV Genius is playing an active role in using the influence of social networks to enhance its recommendation capabilities. It recently launched a TV guide that integrates with social network site Facebook. The new solution uses the company’s content discovery platform to personalise TV guides – when a user logs into the EPG with Facebook Connect, all the shows their friends like are highlighted in the grid. The cloud-based solution means the Facebook-integrated EPGs can de deployed on the web, connected TVs, smartphones and tablets.

“The average person on Facebook has ‘liked’ four TV shows and has an average of about 140 Facebook friends. That’s over 400 TV ‘likes’ for most people. We can use this data to highlight the shows in an EPG that your friends are watching. From an operator perspective, it gives them loads of additional data about that user and what they might like,” says Weiss. He says the company can also scan Twitter to see what TV shows people are talking about. “The nice thing about Twitter is that it doesn’t just bring up the most obvious shows and therefore tends to pull people further down the EPG, which is great for operators.”

Most in the industry agree that the real future of personalisation will come in the form of recommendation. Providing viewers with the content they want with very little effort on their part is seen as the key to retaining customers and tempting them to pay for additional content. While the aim is to keep the process as simple as possible for the end user, the technology behind it can be complicated and fraught with potential pitfalls.

One company specialising in recommendation is Aprico. The Philips-backed company has developed technology that can find, recommend and target content from various sources, including linear broadcast, video-on-demand, web video and targeted advertising. According to Thomas Dvorak, Aprico’s chief marketing officer, operators are initially looking at deploying recommendations based on their own VOD library in an attempt to increase on-demand sales: “The main activity in the MSO field is to use content discovery for finding content in their own VOD libraries. The first step is to make people more aware of the latest movies in their libraries.”

He says that some service providers, particularly IPTV operators with green-field deployments, are starting to take a real interest in redeveloping the EPG towards a user interface that embraces recommendation results. However, other operators, particularly more established players with large customer bases and legacy infrastructure are taking a more cautious approach.

One of the main issues in the area of recommendation is that operators can’t be sure who is watching the TV at any time and so targeting individuals becomes problematic. It is unlikely that children will have the same tastes as their parents, and even amongst children of a similar age there can be great variations in viewing preferences. Some technology vendors have developed systems that involve TV viewers logging into the platform so that operators know who is watching the TV and can therefore target content very specifically. But IP Vision’s Bloom is adamant that customers will not warm to such services. “These options verge from the intrusive to the impractical,” he says.

However, algorithms are becoming so sophisticated that recommendation companies are confident that they can be accurate in the majority of instances. Dvorak claims that Aprico’s trials indicate that its technology achieves an average accuracy rate of 80-90%, meaning that between eight and nine recommendations out of 10 make sense to the user.  So what about that final 10-20%? Dvorak says it is within the operator’s interest to offer a “surprise factor” to help users discover new content. “This is in the interest of user’s changing viewing habits and the fact that we want to help users find new content. You need the right balance of accuracy and surprise; that is what good recommendation engines do.”