may june | Page 13

sort of ‘serendipity’ of discovery of something you wouldn’t normally watch.” “We don’t believe people are going to be looking to do wholesale start-over strategies,” reveals Christensen. “The technology has advanced to a level now where you can solve these issues with point solutions – but it needs to be thought through end-to-end. Otherwise it will never work. To increase their level of personalisation, there are capabilities that need to be built into different parts of the platform. There’s no one point solution that’s ever going to get you there. You need insights about the content, you need insights about the end user, you need the ability to present that to the end user, and you need recommendation engines and lean-back capabilities to keep the end user engaged. But If somebody tries to convince you that you have to replace your entire platform – that’s not the way to do it.” ALGORITHMS. “Providers can help to make their consumer facing UIs and apps better by continually iterating upon and improving the algorithms their discovery and recommendation tools employ,” suggests Adams. “Using more granular data to better describe a show along with more contextual data points around that show, such as how long a viewer spent watching and how many episodes were binge watched in succession, will help drive a more targeted set of programme suggestions. Because content recommendation algorithms run on descriptive metadata, it goes without saying that using the highest quality data and richest visual imagery can also have a significant impact on delivering better, more satisfying user experiences.” “UI-UX specialists have done their homework and noted that people do not consume the same content every time they turn on their preferred viewing device,” adds Smith-Chaigneau. “The parameters that influence them are mood, time/day of the week, location, and device.” “Whether a service upgrades its current infrastructure or creates a new platform altogether, the aim is consistent: to create a frictionless experience for consumers,” states Signorelli. “AT&T is employing a revamped service platform with HBO Max to best consolidate its multifaceted content portfolio. Netflix, on the other hand has edited and upgraded their format for many years now. Each operator will need to employ whichever method works best given the structure of their services and content in conjunction with improved use of data and consumer profiling.” BUBBLES. “Successful content presentation and recommendation creates the impression of always discovering something new,” observes ruwido’s Maier. “The recommendation engines of some providers generate content-filterbubbles that continuously provide viewers with the same content and nothing new. To reach user experience excellence, it is essential to provide the perfect “Successful content recommendation creates the impression of always discovering something new.” – Ferdinand Maier, ruwido alignment of personalised content combined with seamless interaction supporting several modalities. ruwido has been investigating for years now, how to manage this by intelligently mixing AI-approaches with traditional modelling approaches of human activities or tasks.” “Providers should focus and create shortcuts on two core cases,” recommends Fröhlich. “First, they should prioritise the continue-to-watch function. This might be an obvious one, but many platforms have not yet perfected this point. Especially cross-device watching often fails to satisfy. Second, discovery should be a priority. Users want to get relevant recommendations. If platforms don’t manage this, they should focus on optimising the search function instead.” “Like in other technologies, there are “TiVo has focused on Natural Language Processing.” - Charles Dawes, Xperi many failed or poor recommendation systems, which simply haven’t delivered any meaningful uplift or change in consumer behaviour,” advises Docherty. “ThinkAnalytics are seeing video service providers coming round for ‘a second go’ at implementing a more comprehensive system which includes not just search and recommendations but also all the analysis/ insight behind that. Today, using a cloud-native learning platform, it is possible to launch in weeks, not months and to see a fast uplift in engagement, delivering a great ROI.” GOAL. “This really depends on what the provider already has and what the ultimate goal is for their content discovery experience,” suggests Xperi’s Dawes. “In our recent experience, many companies are looking for a solution that allows them to make steps into a next generation of service without completely throwing out the baby with the bathwater. TiVo’s Next-Gen Platform was designed with this in mind and allows a provider to actually mix both approaches. This approach allows you to replace the core platform with the latest advancements in IPTV whilst providing access to multiple clients that can be upgraded or replaced depending on the differing needs of the customer base.” There is no easy journey to the top, but it starts with getting your data under control,” says Bergström. “Regardless of your starting position, you will need high-quality data as the foundation for your recommendation and personalisation strategy. Start by enhancing the quality of your data, getting a clear ID-structure in place and begin measuring how users are interacting with your platform. Once you have done that, evaluate the possibilities of your current platform – does a strong foundation in data give you the benefits you need or do you need to upgrade your platform to something that is more dynamic.” Christensen says that if you really want to drive true personalisation – which everybody wants to do – you need to have a broad definition for that. “It goes far beyond recommending content based on what the viewer looked at the last time. It goes into having a deep understanding of what content is engaging the user and, more importantly, why. In order to do that you have to go far beyond the traditional level of content metadata: you have to understand that this is about the psychographics that drive engagement with the content.” NIRVANA. “Here’s the nirvana point with personalisation: The video and offer presented to me were the right things to get me to subscribe; once I subscribed, the service presented me the right things I should be watching to keep me engaged and watching more; any ads were the right ones to keep me engaged in the monetisation opportunities; and finally, when I was thinking about cancelling, the service put the right asset or offer in front of me to keep me from churning. There are many services trying to solve various points of that puzzle but not many are managing it as part of an ecosystem,” declares Christensen. EUROMEDIA 13