Adeia Blog

All Blogs

June 13, 2023

Personalization and the Future of Streaming

Personalization and the Future of Streaming

While “channel surfing” was a common theme for many years among cable TV subscribers, nobody who’s switched to streaming services talks about it anymore. It still conjures up an image of a bored teenager, slumped onto the sofa, remote in hand, endlessly clicking through the cable channels looking for something — anything — to pique their interest.

This was about as personalized as linear cable (and satellite) TV could get: the provider would offer dozens (and in some cases, literally hundreds) of channels to choose from, with the implicit message: “there MUST be something here you’re interested in watching.”

The trouble was that all too often, either the range of choices was paralyzingly large, or what we wanted to watch wasn’t on when we were in the mood to watch it. So, we waited, or we channel surfed. We planned our lives around the TV schedule. Eventually, the DVR came along and allowed us to record to a hard drive to watch (and rewatch) later, but even then, we were still somewhat bound by the linear TV schedule.

This evolved into the promise of streaming, where not only was most everything available on-demand, but we could pick and pay for only the “channels” we wanted (in this case, Netflix, Hulu, Amazon Prime, Disney Plus, etc.) instead of buying a cable package and having to take all the excess channels we had no interest in. People began to drop cable subscriptions or “cut the cord” en masse, as they switched to their own custom-tailored set of streaming packages.

The early forays into streaming didn’t help much at first. Most Netflix subscribers have, at one time or another, been in the same “channel surfing” mode on that platform too — endlessly scrolling through the thousands of movies and TV shows available, hoping to find something that fits our current mood.

Linear cable and satellite TV still exist, and many people access their home entertainment that way today. For live sports and other live broadcasts, linear is still the dominant method of delivery. But for those of us who have entirely cut the cord, so to speak, we’ve traded one problem for another.

With linear TV there was essentially no personalization, just dozens of channels on a rigid schedule; while with streaming, everything is viewable on-demand, but there’s more content than we could possibly watch in a lifetime.

Having a personalized experience with streaming makes users happier and more satisfied with their service. The proliferation of streaming video-based content has led to major innovations in terms of personalization, which helps match viewers with content that’s interesting to them.

Personalization has big benefits for streaming providers too: when people feel like the provider offers content they want to watch, they are more likely to keep their subscription active. When viewers start to feel like the bored, channel-surfing teenager, they tend to cancel and look for a more relevant content provider.


Nowadays, content personalization happens at several levels, the first of which is driven by users themselves, when they decide which streaming service(s) they will subscribe to. There’s a dizzying array of choices out there now.

There are broad, general-interest providers like Netflix and Disney Plus, with seemingly endless libraries of movies and TV shows. Live sports is a huge market too, with providers like Hulu, ESPN+ and FuboTV carrying every sport imaginable (in addition to other content), some live and some pre-recorded.

Anime fans can get their fix at CrunchyRoll and RetroCrush, art film nerds can subscribe to the Criterion Channel, horror aficionados have Shudder and ScreamBox, and the list goes on.

Just like there were channels for many of these affinities in the cable era, there are streaming services for people who are into history, home & garden, outdoor adventure, travel and more. It’s a smorgasbord of content, and each viewer can load up their plate as they see fit (and their monthly budget allows).

Building a Behavioral Profile

It’s common knowledge to anyone who’s been using the internet for a while: our service providers know a lot about us, and maybe a lot better than us. This same principle applies to streaming providers.

Over time, each provider assembles an increasingly detailed profile of our viewing preferences, based on our behavior. They know what shows we watch (and which we stopped watching) and what time of day we tend to watch (and on which days, and for how long each time). They know when we pause and rewind.

Each data point helps them build a profile that enables them to suggest content they think we’ll enjoy. And each time we select something they suggested, that becomes another data point which further helps refine the profile.

If I watch a lot of action movies and historical documentaries on Netflix, their algorithms will probably suggest similar or related content. If I’ve never watched a teen comedy or a horror film on their platform, over time, they will probably not suggest these genres to me. And it goes beyond genre. If they know that I only ever watch TV shows and not many movies, I probably will get more TV shows suggested to me.

This personalization isn’t instantaneous or even interactive in any real sense. The profile evolves over time, and the longer you subscribe and the more content you consume, the more data is collected and analyzed and used to help personalize your experience with that provider.

Putting Ads in the Mix

Platforms like Netflix used to be (for the most part) reluctant to put ads into their content offerings. But now they, along with other providers, have decided it makes good business sense to keep their subscription costs lower by moving toward an ad-supported model. Hulu and YouTube are two prominent examples.

In these cases, ads are served up during the viewing session, or at the beginning of it. Some are skippable, others are not. But even skippable ads give a provider the ability to gather some data. Did you skip the ad? If so, how quickly? If not, did you click the ad and make a purchase? Based on the content of the ad and your behavior in relation to it, the provider keeps gathering data that can be used to personalize your experience.

Over time, different ads can be inserted for different viewers. If you and I were watching the same movie at the same time, we might get served very different ads, based on our prior behavior and the unique profile that provider has built on each of us.

Ads are fairly straightforward to include in pre-existing content like films and TV shows. Ads can be inserted — just as commercial breaks were back in the linear TV days — at appropriate breaks in the action.

This is more challenging, but not impossible, for content like live sports. Ads can be inserted in appropriate pauses in the games, like after each inning of baseball or after every two games in tennis. But some sports, like soccer, don’t have many pauses in the action, so ads that interrupt the stream don’t work as well.

From a delivery standpoint, there are different ad models too: client-side and server-side. In the client-side model (YouTube is an example), the client can decide which ad to insert during the consumption of the content, sending a request to the server for that ad. In the server-side model, the server decides which ad to push to the client. Both models have pros and cons, and neither appears to be dominant yet.

Streaming is obviously the way of the future, and it will one day replace linear TV, at least for most people. Streaming subscribers appreciate watching (and paying for) only what they want, without the clutter that came with linear TV, and the flexibility to watch on any device. Providers like it too, for its ability to personalize the experience and capture a greater share of revenue per subscriber. Personalization has come a long way, and while it has room for improvement, it’s trending in the right direction.

Exploring the Impact of Artificial Intelligence Applications on the Semiconductor Sector

Technology-enabled Strategies Form Basis for Differentiation for Video Entertainment

Clickable Objects: A Convergence of Video and e-Commerce

Leveling Up: The Evolution of Video Gaming Monetization

Tao Chen

VP of Advanced R&D, Media IP

Tao is VP of Advanced R&D, Media IP at Adeia. He is responsible for initiating and leading R&D projects, supporting the Adeia CTO defining the future technology roadmap and research strategy. Prior to joining Adeia, Tao was Senior Director, Applied Research at Dolby. He managed R&D teams to develop core technologies for various applications including Dolby Vision, which won a Best of CES Award in 2020. Earlier, he was with Panasonic Hollywood Lab and led the development of specs and encoding systems for Blu-ray and 3D production. Tao was an honoree by the Primetime Engineering Emmy for his contributions to Dolby’s Philo T. Farnworth Award in 2021, and a recipient of Emmy Engineering Award in 2008. He won Silver Awards at the Panasonic Technology Symposium in 2004 and 2009. Tao received the award of Most Outstanding Ph.D. Thesis from the Computer Science Association, Australia, in 2001.