How Netflix Stops On-Demand Channel Surfing

Channel surfing used to be a fundamental part of the television-watching experience. Even with the advent of digital TV guides, consumers are still locked into scanning hundreds of channels to find something to watch, giving each show just a second or two to catch their attention before moving on.

In the age of digital streaming services, Netflix doesn’t have the luxury of television’s captive audience. Netflix users who have to “channel surf” for a good movie or show for too long have plenty of other platforms with differing libraries to search through. In fact, Netflix researchers have been hard at work quantifying just how the video on-demand platform’s international growth has affected the time viewers spend searching instead of watching content — after 60 to 90 seconds of scrolling, clicking or swiping through Netflix’s menus, the company knows that their users are more likely than not to abandon them for another provider, via NBC News.

Netflix’s Neil Hunt and Carlos Gomez-Uribe, both vice presidents of product innovation, spearheaded the research, which comes at a critical time for the company. Six days into the new year, Netflix announced the expansion of its service coverage from 60 countries to more than 130 nations under its umbrella.

“From today onwards, we will listen and we will learn, gradually adding more languages, more content and more ways for people to engage with Netflix,” co-founder and CEO Reed Hastings said in a January statement. “We’re looking forward to bringing great stories from all over the world to people all over the world.”

While the doubling of its geographic service base is cause for celebration, the expansion also comes with a good deal of technical headaches for a company twice the size — and with twice the data. The task getting consumers in front of media they want to watch in no more than 90 seconds is daunting for its previously small user base, but for almost the entire world? For that, Netflix researchers explained in a separate blog post, the company had to create a global recommendation algorithm to predict the preferences of the worldwide media consumer.

Instead of pulling apart localized data, Netflix’s transcontinental reach allowed it to analyze the watching preferences of billions of consumers. This new perspective helped identify what Netflix is calling “worldwide communities of interest,” users who share similar tastes in shows despite differing nationalities or socioeconomic backgrounds. The company’s old recommendation used an algorithm to group users by physical proximity, only further segregating them based on “individual models” to differentiate based on taste profiles. By incorporating this globalized data into their already streamlined algorithms, Netflix was able to quickly tailor their recommendations to an audience the size of which it had never interacted with before.

If the size of its customer base was the only thing Netflix had to react to, their executives would be sleeping much easier than they are. Rather, Fierce Online Video reported that Scott Meyer, vice president at Netflix, told a crowd at the Mobile World Congress that half of its users regularly watch videos on their mobile devices, despite just 10 percent of Netflix’s total streaming bandwidth being sent to smartphones and tablets.

More confounding for how Netflix chooses to react to these consumer behaviors is the way the average viewer supplements normal Netflix watching with mobile devices. The majority of mobile content viewing takes place a full hour after the company’s desktop- and TV-based streaming comes off its peak. Netflix explained this as customers embracing a new way of on-demand content; they might begin watching a movie in the living room, but when it comes time to head to bed, they’ll finish it under the covers on their smartphones.