The Ad Tech Middle with First Party Capital General Partners Kevin Flood, Rich Ashton, and Ciarán O'Kane

How Telecom-Grade AI Personalization Can Solve Streaming Retention with Hemant Soni, AI Architect

Have a question? Send us a text! [https://www.buzzsprout.com/2512452/fan_mail/new] The conversation is a companion to Hemant's recent piece on AI personalization at scale [https://www.stateofstreaming.com/articles/hemant-soni-ai-personalization-at-scale]. Hemant Soni [https://www.linkedin.com/in/hemantsoniatlanta/], AI Architect and SOS. contributor, joins Tim Rowe [https://www.linkedin.com/in/troweactual/] to unpack what streaming platforms can learn from the infrastructure telecom companies built under survival pressure — and how to start applying it now.  On Combatting Churn Streaming platforms collect mountains of behavioral data — what you watch, what you skip, when you disengage — and respond days or weeks later with a generic retention email. By then, the subscriber has already left. The model that actually works isn't coming from Netflix or Spotify. It's coming from telecom. * 1:32 – About how telecom is mastering personalization — and why it matters for streaming * 3:05 – T-Mobile's customer decision hub: processing 140M+ subscriber signals in under 200 milliseconds * 5:33 – Comcast's convergence advantage: telecom-grade AI infrastructure applied directly to Peacock's 32M broadband customers Why "Feeling Understood" Matters Most The shift Hemant describes — from responding after a customer cancels to predicting intent before they act — is the key unlock. Customers don't feel retained. They feel understood. That's the difference between a churn intervention and a relationship. * 6:08 – What Comcast is actually deploying: not experiments, proven telecom intelligence * 7:32 – The four pillars of AI personalization and what each one means for a streaming operator * 11:17 – Where to start: a practical framework for media companies beginning the AI journey Get Hemant's 90-day Fast Start Framework  Start with personalization. It's the highest ROI use case, and once you show impact there, scaling AI gets easier everywhere else. Hemant closes with the most actionable thing in the episode — month one: identify use cases and clean your data. Month two: build and test AI models at small scale. Month three: optimize and scale what worked. * 12:42 – Building a unified data foundation: what to connect, clean, and make reusable * 13:56 – Why personalization is the highest ROI AI use case * 14:30 – Language barriers, content hypergrowth, and what AI-enabled localization actually unlocks Connect with Hemant Soni on LinkedIn [https://www.linkedin.com/in/hemantsoniatlanta/] and read his full article here [https://www.stateofstreaming.com/articles/hemant-soni-ai-personalization-at-scale]. Support the show [https://www.buzzsprout.com/2512452/support]

The Ad Tech Middle with First Party Capital General Partners Kevin Flood, Rich Ashton, and Ciarán O'Kane

Have a question? Send us a text! [https://www.buzzsprout.com/2512452/fan_mail/new] Bonus Episode: Tim Rowe on the FPC Podcast — Open Infrastructure, Closed Loops, and the Ad Tech Middle Tim joins Kevin Flood [https://www.linkedin.com/in/knflood/], Rich Ashton [https://www.linkedin.com/in/rich-ashton-fpc/], and Ciarán O'Kane [https://www.linkedin.com/in/cpokane/] — General Partners at First Party Capital [https://firstpartycapital.com/] — for a conversation on their investment thesis and what it reveals about where streaming advertising infrastructure is actually heading. FPC crossed State of Streaming's radar when their corporate innovation model announcement landed the same week as a notable product update from The Trade Desk. What followed was a conversation worth sharing with the SOS. universe. Did You Know? AWS is projected to generate 3x the margin on AI services versus traditional compute in 2026, with 40% of year-over-year growth coming from AI and machine learning — not the media vertical.  The cloud is not neutral infrastructure.  It has a point of view, and it's shaping ad tech. The FPC partners are direct about what they are: picks-and-shovels investors.  They don't back the next sexy format or the latest AI wrapper — they back the companies building the infrastructure nobody sees but everybody needs. * 1:10 – What is State of Streaming? * 2:37 – First-party data as the new center of gravity. Should you: build, own, outsource, or layer on top? * 4:00 – Why the value in ad tech has shifted away from software — and toward data, finance, and infrastructure * 8:30 – Live sports as the unsolved problem in streaming advertising: the gap that Bedrock and Index Exchange are positioned to close The Trade Desk vs. Publicis spat wasn't really about advertiser outcomes.  It was two economic models colliding — a Demand-Side Platform (DSP) maximizing margin on one side, principal media buying maximizing agency take rate on the other. The advertiser is mostly a bystander. * 10:17 – Margin compression as the real story: where the 40–50p in the pound is actually going * 13:52 – Tim puts the question directly: what gave FPC the conviction to back Bedrock early? * 14:28 – The ad tech middle: why global scale DSPs don't serve everyone, and what Bedrock was built to solve * 17:26 – The cloud tax nobody talks about: how AWS, Google, and Amazon costs get embedded in every startup's pricing * 22:01 – What SOS. found when it looked at how much of the streaming universe runs on AWS * 23:21 – Lumen: how a first-party attention signal company became one of FPC's largest portfolio positions — and why the Netflix UK deal matters [https://www.stateofstreaming.com/articles/lumen-netflix-ad-measurement] Most ad tech money flows toward the obvious infrastructure. FPC is betting on something different: composable, containerized, margin-efficient layers that the big incumbents can't easily replicate — and that the ad tech middle actually needs. Learn more about First Party Capital [https://firstpartycapital.com/] and their portfolio. This episode originally aired on the FPC Podcast.  It appears here as a State of Streaming bonus episode. Earlier this year, SOS. covered how the Trade Desk's platform strategy is reshaping the open web buying debate — read it here. [https://www.stateofstreaming.com/articles/blood-red-moon-trade-desk-built-portal-fpc-way-around-amazon-winning] Support the show [https://www.buzzsprout.com/2512452/support]

How Comcast & Cox & Spectrum Reach Unify Linear And Streaming with Ampersand VP of Integrated Sales, Seth Mittman

Have a question? Send us a text! [https://www.buzzsprout.com/2512452/fan_mail/new] Tim sits down with Seth Mittman [https://www.linkedin.com/in/sethmittmanmediarevenueman1/], VP of Integrated Sales at Ampersand [https://www.ampersand.tv/], unpacking how the ad infrastructure jointly owned by Comcast, Charter, and Cox unifies linear and streaming inventory — and what that means for brand marketers, agency planners, and anyone trying to understand where Multichannel Video Programming Distributor (MVPD) inventory fits in a modern media mix. Did you know streaming-only buys are leaving reach on the table? Brands that added linear inventory saw an average 131% lift. The ones that haven't are paying for it. Mittman breaks down what Ampersand actually is, what it isn't, and why the distinction between "one-stop shop" VS. "end-to-end solution" matters more than it sounds. * 1:00 – What redefining cable actually meant — and why it took eight years, not four * 2:25 – The co-opetition structure: how Comcast, Charter, and Cox became one ad sales entity * 3:16 – The sports footprint: 90% of live sports events on television, cross-platform How do you actually buy live sports through Ampersand? Mittman is direct: programmatic live sports isn't the product yet. Direct Insertion Order (IO) is. But the real move is running the same campaign simultaneously across linear and streaming — hitting the audience wherever they went, without upfront guarantees. * 4:49 – Activation: what placing a buy actually looks like across 210 Designated Market Areas (DMAs) * 6:10 – End-to-end vs. one-stop shop: why the framing change matters for buyers * 8:07 – Political: local scale, programmatic expansion, and why it's a separate team Ampersand is the ad infrastructure behind Comcast, Charter, and Cox — yet most buyers have never heard of them. Everyone knows Comcast.  Everyone knows Spectrum.  Everyone knows Cox. But almost nobody knows the backbone sitting behind their combined ad sales effort. Mittman walks through how he educates buyers, what questions he needs answered in every meeting, and why "who are you trying to reach and what are you trying to accomplish" is the only brief that matters. * 9:01 – Core business: holding companies, independents, and the education gap * 10:25 – Multichannel Video Programming Distributor (MVPD) explained: the barbecue version * 11:16 – Trends: why live sports content is pulling buyers back to context alignment * 14:42 – What excites Mittman most after three decades: convergence, not disruption Connect with Seth Mittman: seth.mittman@ampersand.tv [seth.mittman@ampersand.tv] Learn more about Ampersand [https://www.ampersand.tv/] at ampersand.tv Earlier this week we covered Ampersand's closed-loop attribution partnership with Fandango and Kochava — read it here. [https://www.stateofstreaming.com/articles/ampersand-fandango-kochava-movie-measurement] Support the show [https://www.buzzsprout.com/2512452/support]

How Attention Follows Content (Webinar Replay) with David Sanderson, Founder of Reelgood

Have a question? Send us a text! [https://www.buzzsprout.com/2512452/fan_mail/new] Tim sits down with David Sanderson [https://www.linkedin.com/in/davidaesanderson/], Founder of Reelgood [https://data.reelgood.com], unpacking how content availability data reveals where attention actually goes on streaming — and what that means for media buyers, platform teams, and content strategists. Content availability isn't a metadata problem. It's an attention problem.  Most of the industry is making catalog, licensing, and buying decisions based on incomplete information. Reelgood built the infrastructure to fix that — and the insights that fall out of it reframe how you think about viewership, platform strategy, and where the real gaps are. * 1:00 – What Reelgood built and why Google was one of their first customers * 2:36 – The 88% problem: what's actually available across the major streaming services * 4:00 – How a single canonical content ID changes the data picture entirely How did Loudermilk go from 800th to 8th the moment Netflix picked it up and merchandised it?  Find out why that's not an anomaly. When you overlay content availability data against viewership data, platform moves become explainable and predictable. The data also reveals what competitors are quietly doing by genre, what's sitting unlicensed, and where your catalog has white space. * 7:10 – Loudermilk: the case study that shows platform placement is a viewership variable * 9:16 – How to use competitor catalog data to inform your own content strategy * 10:20 – The advertising angle: content availability as a media buying arbitrage lever Why does Franchise content compound? The Squid Game data makes the case cleanly — each new release created a measurable halo effect on every other title in the universe. And catalog gaps are a buyer problem too: Reelgood found 81 titles missing from the Prime/Max channel, including Moonlight and Dune. * 13:00 – The Squid Game ping-pong: how franchise releases drive attention across an entire universe * 15:30 – Catalog blind spots: what media buyers may not be getting when they buy bundled channels * 17:00 – How Reelgood's popularity score is being integrated into SOS.'s Unified Streaming Power Index Download the deck 🔽 [https://podcast.stateofstreaming.com/downloads/sos-reelgood-deck/] Connect with David Sanderson here: https://www.linkedin.com/in/davidaesanderson/ [https://www.linkedin.com/in/davidaesanderson/] And Reelgood [https://reelgood.com] here: https://data.reelgood.com/ [https://data.reelgood.com/] Support the show [https://www.buzzsprout.com/2512452/support]

How Data Turns Live Games Into Stories with Mark Holland, SVP Media Products at Sportradar

Have a question? Send us a text! [https://www.buzzsprout.com/2512452/fan_mail/new] Tim sits down with Mark Holland [https://www.linkedin.com/in/hollandmj], SVP Media Products at Sportradar [https://sportradar.com/contact/], unpacking their new sports media report and the five pillars powering the next generation of sports viewing — from real-time probability overlays to AI-driven personalization at scale. Data isn't a stat — it's a story machine. The box score is table stakes. What Sportradar is building goes further: real-time context that turns an ordinary hit into a career milestone, a routine shot into a geometry lesson in probability. The fan doesn't consume data. They feel it. * 1:00 – How the report frames modern fan behavior and content personalization * 2:36 – From box scores to broadcast context — what data actually does on-screen * 7:10 – What "interrogating the data" means and why it changes storytelling Foresight and GameFrame: two products redefining the live broadcast layer. Foresight surfaces real-time probabilities inside the viewing experience. GameFrame virtualizes player movement from tracking data — not X's and O's, but the exact path a player took and why it worked. Both reflect the same thesis: interactivity is inseparable from insight. * 3:59 – Tim's son's math project — and why it mirrors what Peacock and the NBA are already doing * 5:48 – GameFrame: visualizing the "how" behind the play, not just the outcome * 5:25 – How AI lets partners personalize different experiences to different fans at scale Innovation is coming from every direction at once. Leagues are collecting more data than ever. Media companies are closest to the end user. Sportradar sits in the middle — translating all of it into experiences that scale across broadcast, streaming, digital, and international markets simultaneously. * 8:09 – Why the push is league-driven, media-driven, and platform-driven all at once * 9:16 – Scalability across platforms — why one-size-fits-all is no longer the operating model * 10:20 – The real bottleneck: time and resources, and how Sportradar helps partners do more with less The World Cup is next — and it's a scale stress test unlike anything else. Three Super Bowl-sized audiences a day for a month straight. Sportradar is expanding its soccer data sets heading into June. The full roadmap isn't out yet — but the numbers make it worth watching closely. * 11:04 – What Mark is most excited about on the 2025 sports calendar * 11:42 – The World Cup stat: three Super Bowls a day, every day, for a month * 12:07 – What to watch for from Sportradar around the tournament Connect with Mark Holland on LinkedIn [https://www.linkedin.com/in/hollandmj] and Sportradar [https://sportradar.com/contact/] here. Support the show [https://www.buzzsprout.com/2512452/support]