💡 How Server-Guided Ad Insertion (SGAI) Could Redefine the Streaming Ad Experience

Paul Boustead's journey started "as a researcher doing research for telcos in Australia" back in 2000. Hel worked on real-time CDNs during what he calls "a super interesting time" when it was "very hard to actually get good scalable streams out there and working."His path led through gaming technology—"voice communication, massively multiplayer computer games"—before founding a company that Dolby acquired in 2007. This gaming background proved invaluable, as many low-latency challenges in sports streaming mirror early multiplayer gaming requirements.Defining Purpose-Driven LatencyPaul offers a clear framework for 2025: "Ultra low latency to us is sub-second," followed by "very low latency which is around about the sub 3 seconds" and "low latency to us is sub 7 seconds." Each category serves distinct purposes—sub-second for sports betting and auctions where "they sell very expensive objects, even houses," while 3-7 seconds addresses the broader sports streaming market.The business case crystallized during our conversation when I shared watching Wimbledon with my 10-year-old son, who received Sinner's victory notification 30 seconds before we saw it. "I didn't want to ruin it for the family," he later admitted, "but I knew it before."This illustrates Paul's key insight: streaming's primary challenge is "enabling people to watch together without getting spoilers from social media." As Paul notes, "A lot of people have a dilemma. Do I put my phone on silent, put it over there or to watch the game and trying to get younger generations to put their phone away."The Adaptive RevolutionThe breakthrough innovation that Paul describes addresses varying viewer needs within single events. "We have one streaming service that switches between the required technologies to meet the customer use case," he explains. "If you've got someone watching a sports event, the majority of people may want to be below 3 seconds because they're watching it socially. But if someone's betting on it, they might want below a second."This adaptive approach eliminates complexity: "Our streaming solution enables our customers to do one integration and then pick the latency."Technical RealityFor sub-second delivery, Paul relies on "WebRTC... Plus there's Media over QUIC," both using UDP networking for controlled retransmission. However, scale differs dramatically—ultra-low latency supports "250,000 plus" users but "you rarely see something above 100,000 because they're particular events." Broader sports streaming scales to "millions" over existing CDNs.Platform fragmentation remains challenging. As I noted from "doing 14 different platforms for clients," device diversity impacts optimization. Dolby’s response: acquiring THEOplayer to ensure "a reliable player that large sports organizations would be comfortable deploying across all platforms."Quality BalancePaul acknowledges the eternal trade-off: "There's a big limitation with low delay streaming at the moment when you're really trying to get the delay going down. Sub 3 seconds we’re really doing at a high quality." His Dolby heritage shows in prioritizing perceptual improvements: "You want to get things like the colors right and you want it to be in high dynamic range. You want to get all of that right first before you start increasing the pixel count."Future EngagementLooking ahead, Paul sees AI's biggest impact in fan engagement rather than pure streaming optimization. "Your younger generations aren't that used to or not that inclined to watch long form content," he observes, pointing to Thursday Night Football's predictive analytics as early examples of AI-enhanced viewing.ConclusionThe challenge isn't just moving data faster—it's intelligently matching technology to purpose. Adaptive latency solutions represent the next evolution, promising the right experience at the right time for every viewer, preserving sports' communal joy regardless of underlying technology.

Ep. 55: Paul McGrath - From Cannibalization Fears to YouTube Success: How CBC Cracked the Creator Economy CodePaul McGrath, a 20-year CBC veteran now leading strategy in the entertainment department, shared how Canada's national broadcaster evolved from fearing digital cannibalization to embracing the creator economy through scientific methodology.Three Phases of Digital EvolutionPhase One: Cannibalization Concerns"The first phase was concerns about cannibalization," McGrath explained. "There was concerns about publishing on digital services, cannibalizing a linear audience." This decade-old fear dominated industry discussions about digital distribution.Phase Two: DTC LearningCBC invested in their streaming platform, CBC Gem, building new competencies. "We had to learn things like how do you run a DTC model? How do you do all of the customer support and customer service?"Phase Three: Platform StrategyCurrent focus centers on creator partnerships after realizing platform consumption scale, particularly among younger audiences.The Retention RevolutionCBC Gem achieved its best year ever by focusing on audience retention from major events like Olympics and breaking news. "We really looked at what are the retention rates that we're getting off big events," McGrath said. "What percentage of that audience do we keep after one month, after three months, after six months?"This leverages CBC's "superpower" as a premier news brand: "We don't have to do a lot of marketing for audience acquisition because the news events will drive a lot of audience in."Debunking the Cannibalization MythMost compelling was CBC's scientific test of cannibalization fears using 50 titles across control and test groups. Results shocked the industry: "Overall engagement on the streaming service went up, not down. In some cases, some of those titles almost doubled in their engagement on the streaming service after we published on YouTube."The new hypothesis: YouTube's algorithm creates word-of-mouth marketing driving search behavior back to CBC Gem. "We think that word of mouth converted into search, which led more audience into the streaming service."Creator Economy StrategyCBC's three-pronged approach includes:Production partnerships with creators for development and fundingLicensing catalog content from creators for FAST channelsOpening content libraries to let creators access CBC's archiveIndustry ConvergenceMcGrath observed the merger of traditional media and creator economies: "I used to say YouTube was like Hollywood on a different planet... But those two planets are getting closer together."He attributes this to economics: "When traditional television producers realize some creators can garner a million people for an hour at a fraction of the budget of a TV show, that becomes inevitable."Call for CollaborationMcGrath concluded with an industry invitation: "If you're experimenting around this stuff, please reach out. Let's share our results together."His vision: collaborative research moving beyond anecdotal evidence to establish data-driven best practices.CBC's journey proves that embracing scientific methodology and testing assumptions can transform digital fears into growth opportunities.