Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


How data ‘heartbeats’ are at the centre of DAZN’s customer experience

Stefan Kalcher, SVP data and analytics at DAZN, talks exclusively to TVBEurope about the streamer's use of data to enhance its customer experience

In the battle for audiences in the fragmented sports TV market, the viewer experience is key. This is particularly important when you’re attempting to wrestle viewers away from traditional pay-TV channels. DAZN, often described as the Netflix of sports, has disrupted the sector and built a growing audience by securing key rights (Formula 1, UEFA Champions League, major boxing bouts) across much of Europe.

Anticipating customer needs, creating personalised content, and delivering consistent, engaging experiences help create valuable pieces of data for the streamer, which can be analysed to ensure DAZN’s viewers (and customers) keep coming back for more.

For DAZN, data around the device a viewer is watching on, their internet service provider, and their location is key. “It means we can monitor if there is a problem with a certain kind of device or location,” explains Stefan Kalcher, SVP data and analytics at DAZN.

Stefan Kalcher

For every device that connects to DAZN, the streamer collects what Kalcher describes as a ‘heartbeat’, which includes all of those data points and is tracked every couple of seconds. “It enables us to see if lots of sessions have been closed for good or if the device is throwing an error, and all of this helps us monitor what’s happening with our customers,” he explains. “This helps us understand our quality of experience. And if it’s not good, then we know the customer has a bad experience and that’s when we then need to act upon it.”

DAZN uses Conviva’s SDK, which sits inside its platform, to track the data, while technology from AWS processes the ‘heartbeats’ to help DAZN understand what’s going on. “We also have an internal output that enables our engineers to zoom in and deep dive into devices that have problems” adds Kalcher.

“We’re able to do that in both real-time and after an event. We collect the ‘heartbeats’ and then we summarise them on the fly to understand if there is a problem in a specific area of Austria, for example, and lots of devices are disconnecting at the same time. We can also use it to look back up to two years and see if we are performing better now than we were then.”

Probably the biggest bugbear for any viewer watching their favourite team on a streaming platform is buffering. Connection Induced Rebuffering Rate (CIRR), where the issue isn’t caused by the customer, is among the metrics DAZN keeps a very close eye on. “We look at how long it takes to load the stream when the customer starts watching, how many different CDNs are involved in order to transport the content through the supply chain to the viewer and then optimising it,” says Kalcher. “I think this is the most interesting thing, if there are so many millions of devices out there, how can you optimise the best route of the video for each device? 

“Understanding how many customers are going to watch a specific event in a specific region is the really interesting stuff,” he continues. “We use this information in order to predict viewership, understanding we might have 1,000 viewers at one moment in time and then 10,000 and five minutes later 100,000. We scale our AWS infrastructure according to the predicted viewership. and all of this is data-driven. That’s what we’re doing in order to save costs and at the same time guarantee a good experience to our viewers.”

Extrapolating data is a perfect example of where artificial intelligence and machine learning can help streaming performances, and Kalcher says it’s something that DAZN is making full use of. One example is when customers drop off during half-time in a football match. DAZN needs to know whether they’ll return for the second half. “In order to understand that we need to understand the data first and then scale our machine learning to trigger specific values inside the platform,” he explains. 

“That is our main use case for machine learning and streaming services right now, but it’s only one part of the customer funnel. You also then need to understand from an experience point of view, what is the right content that you want to share with the customer in order to keep them engaged. What might be the next best action for the customer to watch directly after they have been consuming content on your site to keep the stickiness, keep everyone on the platform as long as possible and create a connection with the customer.”

As DAZN continues to grow its content, and its audience, Kalcher believes data will become more important as it aims to understand its viewers even more. “As the company acquires more sports rights, we need to be prepared to immediately create a good experience when the customer is coming to the platform because that’s how you translate the customer’s interest into an active subscription,” he says.