Personalisation helps you to keep viewers engaged and loyal to your company at each viewing session. Among the seven ways to personalise TV listed below, some have been tested and proven effective by Adobe engineers, while others have been proven by the marketplace. You’re the best judge of which ones will work for your audience.
1: Personalisation for new visitors (a.k.a. “The Cold Start”).
When a new visitor finds content to watch on your service, he or she is more likely to become a loyal viewer. Yet it can be hard to recommend the right content to new visitors because of “the Cold Start” problem, where you don’t yet have historical viewing data to work with. You can solve this problem by using data from existing customers who share similar demographics to the new visitor or who watch content in a similar context.
Let’s say that you have anonymised data about the age, gender, location and website visitation history of each new visitor. You can use this data to show millennial women who visited cooking websites one type of TV experience and baby-boomer men who visited sporting websites another type of TV experience. Similarly, you can deliver a customised TV experience to all your most important audience segments.
When a TV service is personalised like this, first-time visitors will be more likely to watch a show in their first session and come back again to watch similar content.
2: Continuous personalisation for return viewers.
The task of personalising TV viewing is never finished because each interaction and piece of data about a user can be used to adjust the experience. Make sure you’re tracking what’s been watched and what hasn’t. This can be used to recommend content related to what’s already been viewed or to promote a whole category of content that may be liked by viewers within a similar audience segment.
3: Personalised browsing experience.
As viewers browse within a personalised TV interface, they should be exposed to optimal content, presented in a way that will compel them to continue watching. This could include personalised navigation, personalised recommendations and even a personalised look and feel of the overall experience.
Netflix provides a good example of a personalised browsing experience. A presentation on the future of recommender systems by Justin Basilico of Netflix and Xavier Amatriain of Quora shows how it works. For each user, Netflix personalises the top-ranked recommendation and gives that one recommendation a lot of screen real estate. This top-ranked recommendation includes a predicted rating of the featured item for the user and evidence to support the rating from the user’s viewing history. Netflix also determines the categories of recommendations that a user may like, organising them by row. And it determines the rankings of recommendations within each category, organising them from left to right within a categorical row.
4: Personalised video recommendations in search results.
Similar to the browsing experience, when viewers search within a personalised TV interface, the results page needs to bring them deeper into the service. The best way to do this is by personalising the search results based on the data you have about each user.
Location is one piece of data that every TV provider has about its users and it can be used to make search results more relevant. For example, Google has been personalising search results for signed-out users since 2009. And it’s using personalised search on YouTube. This can be illustrated in the search results for “football” on YouTube in the US versus in the UK.
Of course, location is just one dimension that could influence video recommendations in search results. A user’s search results could also be influenced by viewing history, device type, time of day, favourite actors, favourite directors and more.
5: Personalised recommendations at the end of a show.
The moment a viewer finishes watching a show is a critical moment in their viewing session. They will either stay engaged or end their viewing session. Many TV services, such as HBO Go, keep viewers watching by automatically transitioning to a new show.
Some transitions between one show to the next are easy to predict, while others require better data and technology. For example, it’s easy to predict that viewers will want to move from one episode to the next in a series. However, it’s harder to predict what viewers will want to watch at the end of a series or after watching a film. Personalisation technology can assist in this area and keep viewers watching.
6: The pick-up-where-you-left-off capability.
Another critical feature for return viewers is the ability to continue watching a show they left. This capability is especially helpful if it works across devices so that viewers can jump around from big screen to tablet to desktop to smartphone as needed.
TV streaming service Crackle provides the pick-up-where-you-left-off capability even for viewers who aren’t signed in.