Roku: One Stop Streaming Shop
What this covers
This Business Breakdowns episode has Joe walking Jesse through Roku, beginning with why television was slow to move online. The cable bundle paid operators around $100 a month per home, broadcasters and cable networks had overlapping ownership, and Hulu was a joint venture of ABC, Fox and CBS testing the internet without committing. Netflix went to pure subscription video because the internet lacked cable's structure, and Joe expects advertisers to follow audiences. Roku's scale is the baseline: over 70 million active accounts, more than Comcast, Charter, DIRECTV and DISH combined, and by Conviva's estimate 40% to 45% of U.S. streaming hours.
The middle of the conversation is operating-system economics. Joe cites IBM failing to unseat Windows and Android displacing Nokia and BlackBerry to argue for Operating System Standardization, a contested forecast for TV. He credits Roku's win over Apple TV to price (about $100 against $300) rather than timing. On licensing, he explains why rivals struggle: Google bars major TV makers from licensing forked Android, retailers such as Walmart are wary of Amazon-powered sets, and VIZIO competes with its own licensees. Samsung's roughly 30% of shipments against about 10% of streaming hours is read as buyers adding a Roku or Amazon stick.
The closing stretch turns to strategy and risk. Joe applies Subsidize Then Monetize with eBay's loss of China as a cautionary case, and shows how a $10 subscription yields either $2 of revenue or $10 less $8 of cost, mostly an accounting difference. He also covers Netflix dropping its player after Reed Hastings spoke with Steve Jobs. He raises The Trade Desk aggregating ad dollars across operating systems as a threat to Roku's reach advantage, and why moving into its own TVs makes sense now but would not have five years ago.
Joe Frankenfield argues Roku's U.S. dominance comes from a cheaper, simpler purpose-built TV operating system that TV makers license, and that its platform economics require subsidizing reach before monetizing, with the main risks being international expansion and content consolidation.
- A purpose-built, lower-memory OS gave Roku a cost and simplicity advantage that drove early adoption and now supports licensing to TV makers
- Roku subsidizes content and device economics to build active accounts, then monetizes mainly through advertising
- Durability in the U.S. is strong, so the open question is whether Roku can replicate its position abroad and survive consolidation of content and ad dollars
Roku's platform scale and licensed OS dominance give it a durable advantage beyond price.
- Advertisers will not move budgets to streaming until services first attract large audiences, so Netflix's subscription model built the eyeballs that now bring advertisers.
“to convince advertisers to move over, you need eyeballs. And so first, you have to get the eyeballs, then the advertisers will move over”
- Roku is the largest TV operating system in the United States with over 70 million active accounts, a number larger than Comcast, Charter, DIRECTV and DISH combined.
“They have over 70 million active accounts. And if you compare that to the traditional pay TV distributors, whether that's cable or satellite, that is larger than Comcast, Charter, DIRECTV, DISH combined.”
- Legacy TV suppliers were slow to move online because the cable bundle was highly profitable, broadcasters and cable networks had mutual ownership, and cable operators earned around $100 a month per home, while internet video had no comparable monetization.
“the cable business was a very profitable business. It was a bundle of bundles of bundles.”
- Conviva estimates Roku has about 40% to 45% of U.S. streaming hours, more than the next three or four TV operating systems combined, with Fire TV at 15-20%, Samsung Tizen near 10%, and LG and Apple TV in low single digits.
“they estimate Roku has about 40% to 45% of market share of streaming hours in the United States and that's larger than the next three or four TV operating systems combined.”
- Over the past decade the U.S. streaming device market has split roughly 40% each between Roku and Amazon Fire, with the remaining 20% spread among many smaller players.
“the streaming device space is split between Roku and Amazon Fire with each having about 40% market share in the U.S.”
- Roku's largest user base creates a two-sided network effect in which content goes where consumers are and consumers go where content is, reinforcing its lead.
“content wants to be where the consumers are. Consumers want to be where the content is. So there's a two-sided network effect.”
- Price and simplicity alone will not remain Roku's strong differentiator, so its growing scale and network effects now carry more of its advantage.
“Now that we look at the playing field today, that alone is not really going to stay a strong differentiator.”
- Roku won because it executed well on cost advantage and ease of use, not because it was first: Apple TV launched in early 2007 at about $300 with limited apps, while Roku launched about a year later at about $100 with broad content.
“Roku did benefit from being early, but they actually weren't the first mover. The first mover was Apple TV, which launched in early 2007.”
- Operating systems economically tend to standardize around a few winners, as shown by IBM's failure to beat Microsoft in PCs despite greater scale and R&D, and by Android displacing Nokia and BlackBerry in mobile.
“the natural economics of operating systems is to move to a standard, to move to a few key winners.”
- Samsung holds about 30% of U.S. TV shipments but only about 10% of streaming hours, implying that roughly two of three Samsung TV buyers add a Roku or Amazon stick because Samsung's own operating system is not good enough.
“Samsung has maybe 30% or just under 30% market share of TVs in the United States, but they only have 10% of streaming hour market share.”
- Google TV holds only about 4% of U.S. streaming hours, which is comparable to PlayStation and Xbox, partly because Google has faced bugs and shifted focus between Android TV and Google TV.
“Google TV streaming hour market share is only around 4% in the U.S., which is pretty comparable to PlayStation and Xbox”
- Google's Chromecast struggled because it was phone-controlled without a remote, while Amazon's low-priced Fire TV Stick succeeded through Amazon's ecosystem integration.
“they didn't come with the remote control and you really had to control how you accessed content through your phone, which customers didn't really like as much.”
- TCL and Hisense went from essentially no U.S. TV market share in 2017 to over 20% by 2021 after licensing Roku's OS, while VIZIO fell from about 30% to the low teens over the same period after not licensing it.
“TCL and Hisense went from essentially no market share in the U.S. in 2017 to over 20% by 2021. That's after they decided to license their OS from Roku.”
- Roku's U.S. market dominance is probably durable given its licensing model and scale, though founder dependence is declining as the company matures.
“The question is, how durable is it? Probably it's pretty durable.”
- Anthony Wood has said Roku's market opportunity is huge, with about one billion broadband households worldwide and about 120 million in the U.S., so selling now would be far too early and the price today matters less than future opportunity.
“There's one billion broadband households across the world and there's maybe 120 million in the U.S.”
A TV-built, neutral operating system wins licensing by simplifying TV makers' work.
- Roku's app-first, iPhone-like interface, which lets consumers click directly into Netflix, Hulu or YouTube instead of navigating content, drove early consumer adoption.
“They tried to make it very similar to iPhone, where it's very app-first-based as opposed to content-first because content can get very muddled and confusing for the consumer.”
- Roku was cheaper early on because its operating system was built from the ground up for TV, while Google TV and Amazon Fire use forked Android that needs more RAM, memory and powerful chips designed for mobile and web browsing.
“They built their programming language from the ground up for television as opposed to Google TV or Amazon Fire, which is actually a forked version of Android that's built for mobile phones.”
- The competition for the TV operating system has shifted from streaming sticks plugged into TVs to operating systems built directly into the TV, as smart TVs become computers that run third-party software.
“What's happening now is that the TV operating system is being integrated into the actual hardware of the TV.”
- Google restricts major TV makers from licensing any forked version of Android, and because Amazon Fire OS is a fork, TV makers have been hesitant to license it.
“Google restricts major TV OEMs from licensing any forked version of Android, which includes Amazon. So TV OEMs have been hesitant to license from Amazon.”
- Google began licensing its TV operating system in 2010, Roku in 2014, and Amazon in 2017, yet Roku's licensing has won more TV-maker adoption despite starting later.
“Google was first to market. They started licensing their Google TV in 2010. Roku started licensing in 2014 and Amazon more recently in 2017.”
- The TV market is very price-sensitive, so small hardware cost differences matter a lot to retailers and buyers negotiating deals like Black Friday TV pricing.
“TV space is very price-sensitive. A couple of dollars goes a long way”
- Large U.S. retailers such as Walmart and Target view Amazon as a direct competitor and are less inclined to sell Amazon-powered TVs than Roku TVs, which makes TV makers reluctant to license Amazon's OS.
“major retailers who sell a large share of TVs in the U.S., largely like Walmart and Target, they view Amazon as their direct competitor.”
- VIZIO started as a hardware company and moved into the operating system from the hardware perspective, and because it competes directly with its licensees, other TV makers are reluctant to license a direct competitor's OS.
“VIZIO started as a hardware company. And then they got more into the operating system because they saw the opportunity, but they approached it from the hardware perspective.”
- Google reportedly subsidizes TV makers, largely TCL, with revenue-share agreements to license Google TV, which shows it has the resources to buy OS adoption.
“supposedly, Google already subsidizes TV OEMs to have a revenue share type of agreement to license at Google TV. I think that's largely with TCL.”
- Roku wins OS licensing by doing much of the work for TV makers: it maintains frequent software updates, designs the main board, certifies factories, sources components, and secures retail placement, while Google TV leaves maintenance to the OEM.
“not only does Roku provide the software, but they also maintain the software updates, which happen frequently. Google TV largely leaves it up to the OEM to maintain it.”
- Roku's move into its own TVs would have been wrong five years ago because it needed to license its OS first, and it can now make hardware because smaller OEMs already want Roku as a partner.
“the question is if they would have done that five years ago, it would have been wrong because they needed to license their OS. And now all the smaller OEMs want Roku as a partner.”
Advertising is a growing share of Roku's revenue and makes its business work.
- Third-party advertising is an increasing share of Roku's business because The Roku Channel grew in viewing hours and Roku manages the inventory under revenue-share agreements, which lowers the margin mix over time.
“that's been an increasing part of their business over time, especially as The Roku Channel has grown in viewing hours”
- Advertising is Roku's largest revenue source; a couple of years ago about two-thirds of platform revenue was advertising and one-third subscription or on-demand, and advertising has grown faster since.
“they said that two-thirds of their platform revenues were advertising, one-third was subscription or on-demand.”
- Roku earns around 80% to 85% gross margins on first-party advertising, where it sells its own share of inventory, but only around 50% on third-party advertising, where it sells on behalf of content partners.
“that's been an increasing part of their business over time, especially as The Roku Channel has grown in viewing hours”
- Netflix and other subscription streamers have been losing share of streaming hours to ad-supported streamers, which are growing faster than subscription services.
“Netflix and the share of streaming hours among all the different streamers... have been losing share over time.”
- All TV content, including library catalog, live sports and news, will eventually be viewed over the internet, and connected TV will overtake the combined share of cable and broadband TV, with ad dollars following viewers.
“The big picture is that all TV will eventually be connected TV. All content... will be viewed over the Internet.”
- The Roku Channel, launched in 2017 as a free ad-supported aggregator, moved from variable content licensing to larger upfront rights purchases as it grew past roughly 100 million viewers.
“as The Roku Channel has gained more scale -- I think it reaches over 100 million people now -- Roku is able to invest more in the rights of the content”
- Roku's acquisitions have been small and opportunistic, such as dataxu and content deals, because its core strategy is to grow active accounts and monetize through advertising rather than buy large companies.
“It's been more like small ones with content. Nothing overly material because the core of their business is just land grab”
- The main difference between EBITDA and operating cash flow is the timing of payments for content assets for The Roku Channel, which shifted from variable license fees to upfront rights that amortize over time, with content cash outflows around $80 million in 2022 and $100 million in 2021.
“So that's probably the biggest discrepancy between EBITDA and operating cash flow.”
- Cable monetized TV through bundled subscriptions plus network advertising, while the internet lacked that structure, so Netflix chose pure subscription video on demand, and advertisers will follow once audiences arrive.
“to convince advertisers to move over, you need eyeballs. And so first, you have to get the eyeballs, then the advertisers will move over”
- Roku monetizes beyond ad breaks through user-interface advertising such as home-screen promotions, remote buttons and search placement, which function like retail endcaps where suppliers pay for prioritization.
“It's similar to having those endcaps in retail.”
- Roku's platform segment generated roughly $2.8 billion, just under $3 billion, making it the majority of revenue, a shift from earlier years when devices were larger.
“The platform segment, which is where they're monetizing their active accounts, those were around $2.8 billion, just under $3 billion.”
Roku subsidizes device accounts early, then monetizes its platform later.
- Roku's total 2022 revenue was $3.1 billion with $1.4 billion in gross profit, and the company had EBITDA losses in 2022.
“Their total 2022 revenue was $3.1 billion. They had $1.4 billion in gross profits, and they had EBITDA losses in 2022.”
- Roku splits into a device segment that sells TV sticks at breakeven to small losses on a gross profit basis, and a platform segment that monetizes active accounts and has driven revenue growth over the last five or six years.
“The device segment is trying to sell the little TV sticks that people buy and plug into the back of their TV, and they really try to target breakeven to even losses on a gross profit margin basis.”
- Roku does not run consumer TV advertising for its devices; instead it spends on retail endcaps, display build-outs and OEM integrations, plus its advertiser sales team.
“They're infamous for not advertising Roku. So you won't see a Roku commercial saying, Buy this device.”
- The platform segment generated roughly $2.8 to $3 billion in revenue and most of the company's revenue, with gross margin around 60%, which has trended down over the last five to ten years.
“The platform segment, which is where they're monetizing their active accounts, those were around $2.8 billion, just under $3 billion.”
- Roku's device business is not a profit center and never will be; its profit comes from monetizing the platform, which can evolve through several advertising and content-placement formats.
“The device segment is trying to sell the little TV sticks that people buy and plug into the back of their TV, and they really try to target breakeven to even losses on a gross profit margin basis.”
- Roku's customer acquisition cost is reportedly negative now, having started near zero, and was about $10 per acquired user last year once the subsidy is counted.
“Anthony Wood would say that they used to have a $0 customer CAC, but now it's negative. So let's say, last year... it was like $10 per acquired user”
- The device segment had roughly $400 million in revenue and losses last year, with negative gross margin because Roku subsidized supply chain inflation caused by COVID.
“COVID caused a lot of supply chain disruptions where they subsidized those inflationary costs that they experienced, so at a negative gross profit margin last year.”
- Roku's management targets EBITDA breakeven and reinvests essentially all gross profit into growth because the market is in land-grab mode for active accounts.
“Historically, the management has targeted EBITDA breakeven. They've always wanted to invest all the gross profits they have because this is a huge space”
- When building a platform, a company should subsidize one side to gain share and only later monetize, as eBay lost China to Alibaba by monetizing too early, and Roku has let Netflix and YouTube onto its platform with low take rates during land grab.
“the lesson is know when to subsidize and when to monetize when you're making a platform.”
- A company should know what it does well and let third parties do the rest, and Roku has done this by licensing content and delegating hardware while moving selectively into its own TVs where it sees gaps.
“The other thing is just focus. You can go to any company but knowing what you do well and knowing what others do well.”
The TV operating system market naturally consolidates toward a few standard winners.
- Whoever controls the living room TV operating system loses earning power if viewing shifts to virtual reality or a metaverse-like platform, though that is probably far off.
“if consumers watch more and more content through virtual reality of some sort, then whoever controls the TV operating system would have less earning power.”
- Anthony Wood said as early as 2009 that there would be a few TV operating system winners, that all TV would be streamed, and that Roku would win, and he has given the same answer since.
“There's going to be a few winners. All TV is going to be streamed.”
- The natural economics of operating systems push toward a standard with a few winners, so the TV operating system market is likely to consolidate even though its fragmentation is debated.
“the natural economics of operating systems is to move to a standard, to move to a few key winners.”
- Nokia and BlackBerry had the largest mobile market share in the early 2000s on proprietary operating systems, but Google's Android was licensed by smaller phone makers and took the standard position.
“Nokia and BlackBerry had the largest market share in the early 2000s, and they had their own operating systems. But then eventually, Google came out with Android”
- IBM licensed its PC operating system from Microsoft in the 1980s, and when it tried to enter the OS market in the early 1990s it was too late because Windows already had the most adoption and applications, despite IBM's greater scale and R&D budget.
“It didn't matter that IBM had greater scale and multiples of R&D dollars compared to Microsoft. It's just that Microsoft was now the standard.”
Roku's licensing model travels internationally, and its hardware and performance outperform rivals.
- Roku announced it would make its own Roku TVs, and it sees pockets like higher-end TVs, where few Roku-licensed sets compete with Samsung, as places it may need to invest in hardware.
“They just announced this past year that they're going to make their own Roku TVs.”
- Roku licensed its first Canadian TV in 2016 and became the largest smart TV OS in Canada by the past year, then pushed into Mexico from late 2018 and became the largest TV OS there, showing the licensing and account-growth model can travel.
“now they're the largest smart TV market share by the TVs sold in the past year in Canada. Now they're doing that with Mexico as well.”
- Roku is now targeting larger TV markets such as Brazil, the U.K. and Germany, and international expansion is the big open question for whether its model consolidates globally.
“they're starting to be very targeted about where they want to move abroad, moving into the larger TV markets.”
- Roku consistently has the lowest video start failure rate, the shortest video start times and the least buffering, according to Conviva measurement.
“they consistently have the lowest video start failures, the shortest video start times and they have the least buffering times as measured by Conviva.”
- Google's Chromecast underperformed because it was built as a phone-controlled device without a remote, forcing customers to control content through their phones, which they disliked.
“They didn't come with the remote control and you really had to control how you accessed content through your phone, which customers didn't really like as much.”
Reported revenue and cost lines reflect accounting presentation and timing more than economics.
- Investors evaluate Roku like a cable company by asking how many active accounts it has and what average revenue per user it earns, with the goal of growing revenue per account over time.
“It's how many active accounts do they have? What is the ARPU for those accounts?”
- On subscriptions, Roku may keep about $2 of a $10 partner subscription under revenue share, or recognize the full $10 as revenue and pass $8 through as cost of goods when it runs the channel itself, so much of the difference is accounting presentation.
“Roku will recognize the $10 as revenue and then forward on the $8 in COGS, so then they'll have $2. So it's just accounting.”
- Roku's SG&A is mostly people, about three-quarters of it, mainly software engineers in R&D and sales and marketing staff, with administrative costs being the third and smaller line.
“I think they said maybe three-quarters of their SG&A are the people.”
- The dataxu demand-side platform Roku bought in 2017, now called OneView, creates working capital timing differences because Roku pays for ad inventory on behalf of advertisers on a different schedule than it receives advertiser payments.
“the demand-side platform, dataxu, which they acquired in 2017. It's now called OneView.”
Incumbents' cable-bundle incentives slowed streaming, and founder dependence is declining.
- Anthony Wood's control of voting shares and his vision were very important early on and still matter, but the business is reaching scale so its dependence on him is less important than five or ten years ago.
“But there's a point where you reach scale. And they're still early on in their life cycle, but it's not nearly as important versus 10 years ago”
- Hulu was a joint venture among legacy broadcasters ABC, Fox and CBS testing internet distribution without fully committing, because cable bundle profits gave incumbents mutual reasons to protect the old model.
“it was a joint venture between ABC, Fox and CBS.”
- The best ways to learn about Roku are past filings, earnings transcripts, Anthony Wood interviews, The Business of Media Distribution, Media in the Digital Age, and Amanda Lotz's We Now Disrupt This Broadcast.
“There's nothing like reading past filings and earnings transcripts.”
Cross-platform ad buying and content consolidation could erode Roku's reach advantage.
- A cross-platform ad buyer like The Trade Desk can aggregate ad dollars across all TV operating systems, which would neutralize Roku's reach advantage and give smaller operating systems similar scale for ad sales.
“a company like The Trade Desk, which can aggregate advertising dollars across all the different TV operating systems, would then mitigate or, I guess, make that advantage that Roku has and they scale less.”
- If TV content consolidates among a few players such as Netflix, Disney and YouTube, those suppliers would gain power over distributors like Roku, and Netflix alone has about 230 million subscribers worldwide.
“the largest one being Netflix right now has maybe 230 million subscribers across the world.”
Roku grew out of Netflix's abandoned player and Wood's earlier DVR work.
- Netflix abandoned its player reportedly after Reed Hastings spoke with Steve Jobs, who did not want Netflix on Apple TV, and because a device distributing other content conflicts with Netflix being a content supplier.
“it was obvious there's a conflict of interest between Netflix, who is trying to be a supplier of TV content and what this device would be, which would be a distributor of it.”
- Reed Hastings at Netflix hired Anthony Wood in 2007 to build a Netflix player for the living room, but Netflix pulled the plug weeks before the December 2007 launch, and the product was spun out as Roku in 2008 with Netflix as its initial sole content.
“they were weeks away from launching this product, and Reed just pulled the plug on it. And so they spun out the Netflix Player into Roku.”
- Netflix invested roughly $5 million in Roku and held about a quarter of the equity at one point, then sold its stake to avoid appearing to have a conflict of interest.
“Netflix had invested, I think, maybe $5 million into Roku and then also had maybe a quarter of the equity at a time, but Netflix shortly after sold their equity”
- Roku founder Anthony Wood invented the digital video recorder, started ReplayTV in the late 1990s, lost a marketing battle to TiVo, and sold ReplayTV in 2001.
“he's the inventor of the digital video recorder, the DVR. And he started a company called ReplayTV in the late '90s.”
- After the spinout, Roku added an app store so other content suppliers could reach its device, positioning it as an aggregator of all suppliers instead of a Netflix-only player.
“soon after they started an app store, so the other content suppliers could then go on to Roku.”
- Wood founded Roku in 2002 after selling ReplayTV, initially building streaming products such as a music player and an image-display device, because he saw the internet would change how media was distributed.
“So he started Roku in 2002. And that was before they started selling the dongles”