Streaming Learning Center Where Streaming Professionals Learn to Excel
Related Articles
I recently sat down with Marc Todd, CEO of Skreens, to discuss how multiview has evolved from a niche feature into a must-have capability for sports and live-event streaming. Skreens has been at the center of this transformation, powering multiview deployments for millions of subscribers. What follows is a slide-by-slide look at Todd’s presentation, told mostly in his own words.
If you prefer to watch the video, it’s available here on YouTube.
Contents
The Industry is Evolving with Multiviews
Jan: Start at the beginning. What did the client-side era look like?
Marc: “For a long time now, we’ve had client-side multiviews, and client-side multiviews have gotten spectacular… if you watch Formula 1 and Tiled Media, these companies are doing fantastic work.”
Jan: Define client-side.
Marc: “Client-side means the code that decodes and creates the multiview on the screen happens locally for the viewer. You need the right hardware power level. Apple TV and certain Roku devices can bring all the video down, do the decoding and mixing, and output a multi-channel multiview… it’s fantastic because it lets you personalize.”
“The issue is you can’t reach every big-screen viewer. Even with the hardware, big screens sometimes can’t support 4K60, and when the screen is 80 inches, quality and bitrate become crucial. It’s hard to do with client-side.”
Jan: What changed in 2023–2024?
Marc: “YouTube TV put server-side multiviews out. They acquired the NFL rights and said, ‘What if we put four games up and just create new channels on the server side?’ Compute is done at the server, I put a higher-quality image down to the subscriber, and they just tune the right channel… their success was well known, gaining subscribers left and right. They’re now 40 percent of the market. With that success, they said, multiviews are great, but we’re going to need build-your-own multiview (BYOMV).”
“Anybody who watches sports knows the feeling. You’re watching a four-up and think, I need South Carolina–Georgia, and it’s not in the four I have. You back out and change the channel. You want to choose your own. They announced build-your-own, rolled it out, and I think they’re expanding it for the NFL season… that success is happening on the server side.”
Jan: From the viewer’s perspective, server-side feels like what?
Marc: “It works on every screen… big screens, PCs, mobile devices, even in the car, because compute is in the cloud. The only thing sent down is a single transport stream, a single decode… just like any other channel.”
Jan: ESPN’s recent rollout?
Marc: “They integrated betting, fantasy, and dynamic interactive elements. On Apple TV, you can create a customized four-up: the four games I want, bet on them, watch fantasy, and buy stuff. That’s the future of what sports fans want.”
Jan: And if you’re not on Apple TV?
Marc: “That’s the client-side conundrum. Without the hardware, you get static multiviews you don’t choose. They’re fine, but if the four-up isn’t the four I bet on, I have to change channels.”
“The voting public has voted, especially with sports. They want multiviews. There isn’t a future where a sports viewer says, give me the non-build-your-own multiview service. If you’re a sports fan, you feel how addictive it is… providers need a path to build-your-own on big-screen TVs.”
Takeaway: Client-side gives flexibility with limited reach. Static multiview gives reach with limited choice. BYOMV gives both.
Introducing Tessera
Jan: Can you give us a quick overview of Tessera before the demo?
Marc: “Here’s the marketing slide. Skreens Tessera is a build-your-own multiview (BYOMV) platform. We’re not going direct to consumers. We’re empowering providers, including content and service providers, rights holders, leagues, or anyone who wants to enhance their direct customer solution.”
“It’s easy to deploy and it’s universally compatible. It works on anything, and you don’t need an SDK. You can do everything from your user interface with a couple of API calls. It’s dynamic server-side cloud orchestration and scheduling. More importantly, it’s real-time, up to 4K60, and beyond. You get a lot of power here, and it’s a quick turn.”
Tessera Demo
Before we explore Tessera’s architecture, here’s a brief demo of Tessera creating a build-your-own multiview in real-time.
This clip shows the provider UI triggering BYOMV, PixelFusion composing the layout in real-time, and the result being returned through the provider CDN as a standard channel. Watch for the sub-200-millisecond pipeline and the ‘Launch’ step that immediately tunes the returned channel.
Now to the plumbing that makes that possible: first the static server-side model, then the Tessera BYOMV back end, and the STB/App UI pattern.
Server-side multiview: high-level architecture
Jan: Walk us through the architecture at a high level.
Marc: “We have a product deployed today, which we call server-side multiview. It’s a cloud-based engine that sits between the provider’s input channels, like CBS, ESPN, and so forth, and their CDN. We’re sitting in the cloud.”
“The provider says, at two in the afternoon on Saturday I want a three-up multiview with these three games, a four-up with these four games, and send them back. Those are engineered and targeted for a region. They give us that on a Tuesday for a Saturday. It’s scheduled, orchestrated, it spins up and spins down.”
“Skreens runs this as a managed service. We watch things; we have the monitoring baked in. Video comes to us, our multiview engines build the multiview, we output to the CDN, and the user simply tunes to the channel the provider wants. That channel is now the multiview. This is essentially the static concept.”
“If the provider wants to create 10 multiviews for a couple of million subscribers, you can do that. You want 50, 100, 10,000, you can do that.”
Jan: Where does this hit limits for build-your-own?
Marc: “At a certain point, they say, this control is fantastic, but I need build-your-own. You could try to create all the combinations a user might choose. If you take 25 channels, the user can select and do all the permutations, you’re in 56,000 static multiviews. If you spun up 56,000, the user will select one. For build-your-own, that’s not the right mechanism, which is why we suggest moving to Tessera. Tessera build-your-own multiview has a similar architecture, but it’s built for BYOMV.”
Slide 4 — Skreens Tessera BYOMV: high-level architecture
Figure 4. Tessera BYOMV architecture — the user asks, PixelFusion composes in real-time, and the result is returned as a channel through the provider’s CDN.
Jan: How does Tessera BYOMV work at a high level?
Marc: “We take the streams from the provider and target the provider’s CDN. It works the same way. The infrastructure is exactly the same. Conceptually, the user asks for a multiview.”
Marc: “One huge property about building your own is you don’t have to spin up 57,000 channels. When the user asks for something, there’s immediate value. Cost and value are tied together, so we can step up the cost as people ask for more specific or unique multiviews.”
Jan: Walk me through the app’s flow.
Marc: “This is API driven. The service provider controls the subscriber’s journey. You press the button, and it pops up ‘build-your-own multiview.’ You select and say launch or tune in, and it will tune. Skreens end-to-end is sub-200 milliseconds. We’re not the latency factor. You’re literally creating a brand-new TV show and editing it in real time.”
“From the UI, the set-top-box code sends a message back: I’ve got ABC Miami, this user is building a multiview. We can surface channels that already have ABC Miami, plus a list of channels that are hot. This enables what we call Skreens Smart Throttle. The provider can throttle how many new multiviews get created in the backend.”
“The user picks the next channel, say Fox Miami. Same process. You’re in complete control. Eventually, they say, ‘launch this.’ Launch either makes the multiview or gives me the multiview, and I tune the station. The provider keeps the numbers and can free up multiviews no one is watching for reuse.”
Jan: Any concrete example of how this behaves in the wild?
Marc: “If somebody is in Boston and says I want the Bruins and the Celtics, perfect. I ask for that. Tessera makes it for me, bam, it’s up. Right off the bat, the provider gets value because I’m watching longer. When others in my neighborhood ask for the same multiview, we don’t create another. They get the same channel. With one multiview that might cost twenty bucks, I’m entertaining ten thousand people, and at a twenty-dollar CPM, that’s two hundred dollars you just made.”
Jan: Where does PixelFusion fit?
Marc: “Tessera lets you buy the server-side build-your-own solution. It works with your existing technology, and it’s our PixelFusion technology running from the cloud that makes this possible.”
Tessera service: STB/App user interface model
Jan: Now walk it from the UI and product side.
Marc: “From the engineer and product perspective, I need to let the user build a multiview. This representation is a build-your-own multiview UI. Under the hood, it’s just API calls. You’ll have a pop-up or a guide list that says, ‘Pick the first channel for your multiview.’ I select ABC Miami.”
Marc: “Right away, the set-top box code sends a message back: I’ve got ABC, this user is building a multiview. We can surface channels that already have ABC Miami, whether they were pre-created or created by neighbors, plus a list of channels that are hot. That lets the UI and provider re-order the list, put popular stuff on top, and push other things down. If the user picks something like Cooking Network or History Channel that is not already in a multiview, we generate a new one.”
Jan: And the throttle you mentioned?
Marc: “That is Skreens Smart Throttle. The provider can control how many new multiviews get created in the backend and push people toward existing ones. A nice side effect is discovery. Maybe CNN is hot. I might click it to keep an eye on the news while I watch the game. You’re helping the user entertain themselves with a couple of simple APIs.”
“Then the user picks the next channel, say Fox Miami. Same process. You are in complete control. Smart Throttle keeps it from exploding into a million multiviews in two hours. Eventually, the user says Launch. Launch either makes the multiview or gives me the multiview, and I tune the station. The provider has the numbers and can free up multiviews no one is watching for reuse.”
“We keep this as straightforward as we can so any provider, any app, any client can use it. If you already run a client-side multiview, we are synergistic. If you don’t have enough compute on the device, have us build it and stream it back down. The rich ecosystem benefits from one facet we provide, the dynamic real-time PixelFusion technology.”
Jan: What are providers actually doing in the field? Stingy or generous with multiview channels?
Marc: “The providers we work with do a great job. They know their KPIs. They are opening up multiviews and getting smarter about why and when. At a massive scale, you also have lots of ad zones and local affiliates, so Boston ABC is different than Miami ABC, and that creates more multiviews. There’s also a time element. With NFL you tear down the one o’clock set and spin up the four o’clock set.”
Jan: Order of magnitude per area?
Marc: “Inside twenty, depending on what’s happening. Monday or Tuesday nights might have one or none. But once you introduce a build-your-own option, users will want it on Tuesdays and Wednesdays too. They’ll watch a show and keep an eye on a game because they have five bucks on the over. Give people personalization, and people personalize.”
Tessera Implementation Made Easy
Jan: You pitch Tessera as rapid to deploy. Walk us through the phases.
Marc: “We’re trying to make Tessera rapid to deploy and quick and easy. We’ve partnered with serious equipment, service, and cloud providers to fill out the ecosystem and offer the best to the end customer. Phase one is integration and pilot: spin up a Skreens system, get your live video flows, make sure they work into our system, make sure we can output back to your CDN, IP addresses, manifest files, and make sure it’s working. Your UI engineer builds whatever they’re going to do for phase one, and makes sure the APIs are coming back. We run tests, burn tests, employee trials.”
Jan: So phase one ends when it’s up and running?
Marc: “Phase one takes about two to six months. We charge $50K a month. It’s a managed service. We’re involved in the installation, provide the technology and engineering, and support live events. This isn’t a huge revenue driver for us; we’re getting you set up. At the end of this phase, if you have 10 million subscribers, they’ll all be able to build their own multiview. Phase two is initial deployment.”
Jan: At the end of phase one, everyone has build their own or static?
Marc: “Everybody will have build-your-own, Tessera BYOMV.”
“In phase two, the question is: how many multiviews do you want to build? If you open it to 10 million people, do you set a limit of, say, 50 multiviews? Those 50 are in the cost structure of about $7 to $12 per multiview hour. If I run 50 for 10 hours, that’s 500 multiview hours. You can cap it at 50 or let users do more. Saturdays bring more requests and uniqueness than Tuesday nights. Remember, you have Smart Throttle, totally under provider control. We’ve seen customers launch in a couple of ad zones or on a Thursday night, get their feet wet, then grow. The multiview hours increase, and that’s phase three.”
“You get into it fast, shared risk, shared success. At the end of phase three, two things happen. Ops engineers say, I need to make this cheaper. Product says, I need to make more money. Phase four is optimization and scale. Multiview-hour cost can come down. Cloud cost will come down, AI clouds will drive it down, and if you’re really interested, we can drive it further with on-prem data centers; make it CapEx.”
Jan: Just to be clear, multiview hour is irrespective of how many people are watching; not one per user, one per channel?
Marc: “That’s correct. If you want to stand this up and pay twelve bucks for one multiview and give it to 10 million people, you’re good to go. That’s the benefit. But cost is one thing; don’t miss the forest for the trees. The cost is here, and the revenue is here. If I decrease cost by one notch and increase revenue by ten notches, the gross margin in between is what we’re all looking for. PixelFusion provides something amazing at this phase.”
Dynamic Advertising Opportunities
Jan: You said advertising is just one facet. What does Tessera enable?
Marc: “Dynamic advertising. This is just one facet. There are many others because Tessera, using PixelFusion, allows you to edit the video in real time. You can insert ads into the multiview, and the multiview shape gives you all kinds of new advertising opportunities.”
Jan: Client-side guys already dabble here, right?
Marc: “Client-side multiviews already do interactive stuff and targeted ads. You’ve got real innovators like Play Anywhere and others doing tremendous technology for sports rights and interactive rights and creating those ads.”
Jan: Walk me through the difference between in-stream and out-of-stream.
Marc: “These web pages and others have a model for in-stream vs. out-of-stream ads. It hasn’t really been present with MSOs, but here’s how it works. Say you have two videos up. The first is Fox, and a Geico ad comes up. I’m listening to Fox while I’m watching ABC. The Geico ad runs on Fox, and I hear it. That’s in-stream. I get my full $20 CPM as a provider. On ABC, the Geico ad runs, but I only see it, and there’s no sound. That’s out-of-stream. The industry has been billing and paying 50% of the CPM — ten bucks — for that same ad.”
“Using that concept, just by having multiviews up, doing nothing, the provider can make more money as they work deals with ad companies. And because PixelFusion can place an ad, we can actually insert an ad in-stream or out-of-stream. We can do squeeze backs and serve ad impressions.”
Here’s the video that Marc referred to in the presentation.
Marc: “The customer might be watching three channels, but the ad is running in other areas of the screen. Impressions, sponsorships, custom backgrounds, polls, quizzes, all kinds of things. You could run four ads if you really wanted to get psychedelic, just like on web pages. We’re not an ad company; our job is to provide a platform with the speed, flexibility, and structure.”
“So I’m in Boston watching the Celtics and Bruins, and my provider says, I have a personalized Bud Light ad I can insert here, and I’m going to charge more CPM because there are 10,000 people watching that two-up with the Celtics and Bruins. Insert the ad, you’re making more money. The amount of revenue you create will start to matter more than the lower cost. That’s the power of personalization: the more personal it gets, the higher the CPMs, the more information and analytics the provider gets, and the more money everybody in the food chain makes.”
The Evolution of TV & Multiviews
Jan: Let’s finish with the evolution slide. How do you see this playing out?
Marc: “For my entire life, we’ve lived with linear broadcast TV, and linear has been fantastic. One-to-many has solved size, quality, and reach. The reason things are happening now is that the cloud and the semiconductors have finally made this possible. That’s where Skreens’ patents lie; we’ve been a deep-tech startup waiting for this moment.”
“From straight linear, we go to preset multiviews. People love these; many providers are doing them. The next phase is letting the subscriber build their own. The subscriber is in charge, and they’ll get addicted to it. PixelFusion provides all kinds of upside with revenue. The ad package is an add-on; the four-to-seven or twelve bucks per multiview hour isn’t part of the ad package.”
“And this isn’t the last stop. The endpoint is hyper-personalization: third-party integrations, unmatched interactivity, and personal agentic AI agents. I should be able to talk to my TV, bring this up, that up, place this bet; all of it.”
Jan: You’re pushing BYOMV hard. If I’m a sports provider and the season’s about to start, is static faster to stand up than build-your-own?
Marc: “A little shorter, because you can pin up channels and even just add them to the end of your EPG without changing the app. Lots of customers start with statics. Product teams need to get their heads around dynamically creating brand-new channels and tearing them down. But if I were a service provider, I’d go after build-your-own immediately. This isn’t going away. It’s like DVR; would you launch without DVR?”
Jan: If I can get static a month quicker, I don’t want to lose subs while my competitor launches. Can I accelerate by going static only?
Marc: “The way to accelerate is commitment. We’re not the long pole in the tent — the long pole is internal: priorities, UI decisions, engineering time. Subscriber churn will focus the mind. YouTube TV was small; now they’re large. Personalized multiview is a step into full personalization.”
Jan: What about second-screen value, betting, fantasy, socials? Can Tessera help reclaim it?
Marc: “Everything I’ve shown is basic and still powerful. We can do seven-ups, eight-ups; we have the APIs and patented structures for second-screen. If a provider leans in, it’s massive: place bets, pull info back, drive something dynamic that comes down to the big screen. I actually think the second screen and the first screen are the same screen — just different pixels. With hyper-personalization it’s the same screen. We have other technologies, such as pixel leasing. Skreens is focused on hyper-personalization with AI agent assistance.”
Jan: There’s always a cost vs. revenue trade-off: extra transcodes on the server-side, but new income streams and anti-churn.
Marc: “Even without ad insert, plain chocolate and vanilla, say a multiview costs twenty bucks at low volume. If a thousand people watch, the very first $20-CPM ad covers the entire cost. And with build-your-own you know it’s watched because someone asked for it. If he asked for it, there are another thousand, ten thousand who look like that. Pretty powerful.”
Jan: And the broader market trajectory?
Marc: “This shift is like boiling a frog; the temperature goes up a little at a time, and you don’t notice until it’s too late. We’re seeing the on-ramp to personalization force providers to address what customers want. It’s unleashing a race with four hundred horses; YouTube TV is seven furlongs ahead, and everyone’s chasing. It’s going to keep going like that, and it’s all around personalization.”
Tessera In Action at IBC
Jan: For anyone who wants to see it live?
Marc: “Come see Tessera at IBC in Amsterdam. We’ve got a room and we’re setting up meetings. Go to the Skreens website to request a meeting on our calendar and get more information about build-your-own multiview in Tessera.”
Jan: This has been very useful. I won’t be at IBC, but good luck at the show.
Marc: “Cheers.”
