The NBA’s digital transformation is a game-changer

The NBA’s full-court press on digital technologies has revolutionized the fan, player, and team experience, thanks to accelerated deployment of cloud, analytics, AI, and computer vision technologies since the association launched its digital transformation in 2020.

NBA EVP and CTO Krishna Bhagavathula, a former NBC News CTO who has been with the National Basketball Association for almost six years, says the launch last October of the NBA League Pass streaming application along with a new partnership with Hawk-Eye Innovations for 3D optical tracking demonstrates the advanced pace of digital change happening in the NBA.

“We’re trying to create a truly differentiated basketball experience for our fans,” Bhagavathula says, noting the return on investment to date has been significant. “We got about one billion video views this season, which is more than triple last year’s total.”

The NBA has also enjoyed a 50% growth in subscribers and 52% increase in viewership this season, the CTO adds, attributing the wins to the NBA’s tight partnership with Microsoft and its three-point action plan that began in earnest three years ago.

“We had a few major initiatives that we were planning, including the flagship launch of League Pass, deploying the streaming product and [associated] new products on a cloud infrastructure and realizing it was time to modernize our entire application stack,” Bhagavathula said during a telephone interview from Denver on June 1 when the NBA Finals kicked off.

Aside from the consumer experience, the NBA’s embrace of Microsoft Azure also initiated a transformation of the association’s internal applications to a hybrid cloud/on-premises mix, with its modernization efforts resulting in between 300 and 400 apps migrating to the cloud while others requiring less flexibility and scalability have remained in the data center.

The third pillar of Bhagavathula’s cloud game plan includes the development of advanced applications not just for fans but for referees, players, teams, and coaches, all of whom rely more on data to improve their performance.

The drive for digital, AI

When Bhagavathula first arrived, the NBA was just beginning its cloud migration and, like many enterprises, continued to manage an on-prem infrastructure filled with servers running thousands of workloads on virtual machines. But that’s all changed.  

“We had a data center that was heavily virtualized,” the CTO says of the state of play five or six years ago. “We were very focused on creating a truly delightful direct to consumer experience powered by the cloud.”

Like the NFL, the NBA CTO opted to partner with Microsoft to leverage its Azure cloud platform, which Bhagavathula says contained all the digital components necessary to build the association’s streaming platform, while providing a cloud data lake and machine learning models the NBA could capitalize on for next-generation applications.

The NBA App, which launched in 2022, is a free application that consolidates a multitude of personalization services for fans, including enhanced integration with social media and live games streamed over the subscription-based League Pass, which offers basic and premium versions.

Azure AI-based “For You” app offers personalized content based on fan preferences as well as real-time video highlights of teams targeted at each consumer. Another new application, called NBA ID, offers members-only experiences, game ticket giveaways, personalized content, and enhanced voting for All-Star picks and other NBA-related issues.

Bhagavathula highlights the availability of League Pass within the last year as a major step forward for the NBA, but he is also quick to point out the increasing use of Azure analytics and AI to provide content and new digital opportunities for NBA staff, teams, and players.

The development of an app Bhagavathula refers to as REPS, a referee engagement and performance system, is “specially designed to aid referees and management to evaluate, collaborate, and regularly focus on athletic performance,” the CTO says.

Bhagavathula also cites the development of several machine learning models that evaluate “customer churn metrics” for how engaged fans are with the NBA’s products and a “game recommendation model” that relies on fan behavior and usage of the application to improve the personalization and customization for each fan.

“The features we are using gives us a sense of both implicit and explicit preferences so we can come up with a recommendation model that’s personalized on a per-fan basis,” he says, adding that the NBA then “targets fans who have a high propensity score to win with those recommendations so they are increasingly engaged with the app.”

The data continuously amassing in the NBA’s cloud lake is also collected from and made available to the NBA store and to merchandising partners, such as Fanatics and Ticketmaster, as is fan survey data. It’s all part of a plan to build up a repository to enable the creation of next-gen AI- and analytics-based features for all parties, the CTO says.

Taking performance to the next level

As for the NBA’s teams, Bhagavathula points to CourtOptix, an application that tracks and analyzes action on the court. The tool, also developed on Azure, packages up data for each team free of charge and distributes it to them to improve their game. “After each game, the teams get a cache of the data and can then use this to gain additional insights and stats for their own analysis,” he says.

Such sports-related analytics initiatives are significantly altering how professional sports are played across a range of disciplines, including tennis, where the International Tennis Federation is leveraging computer vision to help enhance player performance.

Craig Powers, an analyst at IDC, sees no end in sight to that trend.

“Data in the NFL and NBA has changed the way the sports have been played,” Powers says. “In the NFL, strategies that were once thought to be risky are now the norm — such as going for it on 4th down versus kicking field goals or punting. This is the product of analytics — taking decades of NFL play-by-play data and assigning probabilities and point values to each decision.”

Like the International Tennis Federation, the NBA is betting that the increased use of sensors, cameras, analytics, and AI will have a major impact on enhancing player, team, and officiating performance. While players today do not wear sensors during live action games, the introduction of wearable technologies and new types of IoT sensors will no doubt impact data collection and player performance in the years ahead, Bhagavathula says.

The NBA’s partnership with Hawk-Eye, for instance, will yield far more detailed data that will be used by players and teams to improve their play and plan strategies for games.

“It essentially uses skeletal tracking technology to evaluate how high an athlete is jumping, their body posture when making a shot or how they land on their ankle,” the CTO says. “We think this empirical data will actually unlock insight into basketball.”

It is crystal clear that the unleashing of advanced digital technologies such as AI and cloud computing will continue to change how teams, players, and fans participate in the evolving game of basketball — once simply played with a single ball, two hoops, and high hustle players. 

The introduction of digital technologies at arenas, such as contactless concession stands, and the creation of NBA Top Shot — nonfungible tokens sold as digital basketball cards — are having an enormous impact on merchandising and the basketball economy.

IDC’s Powers believes increasing use of analytics, AI, and sensors to build next-generation game plans, assemble rosters, monitor players’ health, schedule player availability, and manage injuries has changed the game forever.

“In the NBA, we’ve seen the three-point shot rise in prominence as the midrange jumper is frowned upon. Why? Because teams have a better understanding of the value of each shot on the floor,” Powers says. “The NBA has taken it a step further by using video to track the exact spot every pass and shot is taken. Coaches and players can more specifically identify their strengths and weaknesses, and this dictates how offenses run and defenses defend. AI uses this data to recommend specific player matchups and tactics.”

While Bhagavathula did not provide comment on the NBA’s plans for generative AI, AI guru Tiago Cardoso, group product manager at Hyland Software, which counts Formula 1 as a partner, says the introduction of large language models (LLMs), such as GPT, have the potential to transform the game and fan experience further.

“Large models can perform generic tasks, so they could identify players, understand actions, tactics. They would be able to recognize how engaged the crowd would be with a play, and it can express all that in the form of text or voice commentary” — a reality Cardoso suggests is just a few years out, adding that LLMs could one day be used to catalog all past televised games to generate tailored highlight compilations based on prompts such as “Show me videos of LeBron James and [Michael] Jordan doing the same dunks,” Cardoso says.

And it’s all made possible by a foundation of well-orchestrated information technology.