Analysis Paralysis: How to Manage the Streaming Data Deluge
When it comes to live stream monitoring and analytics, the problem is usually having too much data rather than not enough, along with identifying the key data points to pinpoint problems and optimize streams. NBCUniversal’s Seb Emin and Warner Bros. Discovery’s Subhrendu Sarkar explain how two major content providers keep analysis paralysis at bay in this discussion with Dillon Media Ventures’ Rob Dillon at May’s Streaming Media Connect.
Tailor Based on Context
Dillon Media Ventures Streaming Media Strategist Rob Dillon opens the discussion by acknowledging the plethora of data available. He asks, “How do you, your teams, and the people you’re distributing this data to avoid what I’ve always called analysis paralysis, where you start chasing numbers?” He also wonders how to keep your concentration on the right data.
NBCUniversal VP, Playback Experience, Seb Emin says it’s about staying on message based on who is asking for the data and what, specifically, they’re asking for. “There’ll never be a kind of one-size-fits-all solution,” he asserts. “If we are providing an update to an executive, for example, we know exactly what that person is looking for. … Our team acts as the filter for that data.”
Emin doesn’t worry about analysis paralysis because working with data has become routine: “We know every single metric and what we would use a specific metric for in a given situation. And so we sort of take that responsibility from [the recipients of the data] so that we can literally just give them the information that they’re looking for.” NBCUniversal has found that to be a highly effective approach, he says. “We use it not only on live streaming and live monitoring of events, but also when we’re building features; if we need to make decisions about how we build things, we use this data as well. And again, we fine-tune what we provide based on the context.”
‘Slice and Dice’
Warner Bros. Discovery Engineering Lead Subhrendu Sarkar joins the discussion to add that it’s necessary to “slice and dice the data” to understand device-specific performance metrics. His team can drill down into a device’s capabilities and its playback experience “to basically figure out both how important it is and what the impact of it is in terms of that device either getting a 4K stream or not getting a 4K stream, or what its VSF and VPF or VST numbers look like.”
Approaching data in such a granular way is more effective than looking at overall metrics, Sarkar says, noting that “really drilling down to very specific devices and their capabilities and figuring out what the optimal numbers should be for those specific platforms becomes very important in really telling the full story of the platform and providing [an] experience [that] is optimal for each platform and every device.”
Join conference chair Andy Beach and other streaming media experts in person Oct. 6–8 in Santa Monica, CA, for more thought leadership, actionable insights, and lively debate at Streaming Media 2025. Registration is open!
Related Articles
Bitmovin VP, Solutions Igor Oreper explains how to mean clients where they are when it comes to various levels of streaming metrics expertise.
04 May 2022
Atmosphere VP Business Insights Ethan Moore discusses the importance of measuring what matters when trying to serve customers effectively with streaming metrics.
02 May 2022
Disney Streaming's Eric Klein explains how Disney works to improve streaming performance and reliability and guarantee a top-notch end-user experience through the use through a mix of raw and inferred metrics across their range of CDN partners in this clip from Streaming Media West Connect 2021.
29 Dec 2021