Three years ago, YouTube had a big problem. People were watching YouTube videos on blogs, news sites and social media — everywhere but YouTube. As a result, the site's 1 billion monthly viewers were only bringing the company $US4 billion in revenue, less than what Netflix was skimming off just 50 million subscribers. Here's how they solved the problem.
At the core of YouTube's trouble was rusty, outdated infrastructure that made keeping up with the deluge of other video streaming services difficult. And as non-desktop platforms — mobile devices, Xboxes, etc — proliferated, YouTube's app development process became ever more splintered. Inconsistent software across dozens of platforms made it hard for the company roll out new features and study user responses efficiently.
To bring development back under a single roof and out of the dark ages, in 2012 YouTube launched a massive, cross department initiative, code named "InnerTube." The project would overhaul everything about YouTube's backend, from its development platform to machine learning algorithms, enabling engineers and designers to keep pace in an world increasingly dominated by mobile devices, where constant iteration is necessary.
InnerTube has amounted to nothing less than a complete overhaul of YouTube's development process. A new, streamlined API now allows engineers to crank out everything from minor tweaks to major updates in a week's time or less across the site's various mobile platforms. Changes can now be easily undone, offering developers the freedom to experiment with new features in a way that wasn't possible before.
Project InnerTube has also rebuilt YouTube's recommendation algorithm by tapping into Google Brain, the AI system that now powers Android's voice recognition. YouTube now feeds Google Brain (as creepy as it sounds) the activity of millions of users and asks the computerised learning system to draw connections between the videos we watch, to study how long we remain engaged, and to see what sorts of content we're favoriting the most. Smart recommendation algorithms are a critical part of maintaining an engaged user base — an army of YouTubers that will plunge deeper and deeper down the cat video or let's play rabbit hole like there's no tomorrow.
According to FastCompany, the project is paying off in a big way:
"All of the things that InnerTube has enabled — faster iteration, improved user testing, mobile user analytics, smarter recommendations, and more robust search — have paid off in a big way. As of early 2015, YouTube was finally becoming a destination: On mobile, 80% of YouTube sessions currently originate from within YouTube itself. Even on the desktop, where users have been sprinkling YouTube embeds for more than a decade, 55% of views are happening on YouTube.com. By early 2015, YouTube says it's seeing 50% year-over-year growth in total viewership. And a majority of those views now originate from within YouTube's own apps or website."
So, it looks YouTube managed to pull itself out of its brief spiral toward obsolescence. At least for now, the world's greatest one-stop shop for cat antics, Star Trek recuts, and dangerous DIY projects will live on.
You can more about how project InnerTube is transforming YouTube over at FastCompany.