Description

ROKU has traded off ~60% from 2021 highs in part on the selloff in growth stocks, but also on fears of increasing competition and supply chain issues. I believe the market is myopically squinting at near-term squiggles and failing to appreciate the emerging crown jewel asset, The Roku Channel. I believe The Roku Channel alone is worth $250+/share today (far more than the current stock price), even though it will likely make up under 10% of streaming hours in the long-term.

Roku was written up previously with some background and a long discussion thread, so I won’t re-hash the core business, and instead, I’ll focus on The Roku Channel.

The Roku Channel is an app controlled by Roku that allows users to watch free ad-supported content. The content is primarily from third-party content owners, which get a revenue share (roughly 50%) on the ad revenue generated by their content. Given that TRC is an ad-supported model, driving more engagement on TRC is key. So, Roku has been building out the content housed within TRC by expanding content partner relationships, and has been supplementing third party content with a small amount of licensed or produced original content. This is a flywheel business – more eyeballs drives more monetization, which drives more interest from content owners to put content on the platform. Initially, TRC started with longer-tail, dated films and TV series, but is gradually improving quality and attracting more premium content. This is how aggregator businesses are built.

Roku has been dropping breadcrumbs to investors to indicate that TRC is becoming a very large business, but they have never disclosed the revenue so most investors don’t appreciate how big this business is, or how big it can be, so I will walk through the math below.

In Q1 2020, Roku said that TRC reaches households with 36m people, which assuming 2.5 people per active account (consistent with Roku overall), implies 14m active accounts watched TRC. Roku last updated the total TRC reach number in Q1 2021 (70m, which works out to ~28m active accounts) and given Roku’s continued growth, I believe TRC will reach 32m active accounts by the end of 2021. So, how much revenue is TRC generating? In order to estimate that, we need to know the # of hours watched per account per day, the ad load, and the CPM. While Roku doesn’t disclose a ton of information about TRC, TRC’s closest comps are PlutoTV (ViacomCBS) and Tubi (Fox), where disclosure is a bit better. Fox has given MAUs and total hours watched for Tubi, so we can see the average MAU watches ~15 minutes per day. Applying Tubi’s 15 minutes per day to an average of 28m active accounts watching TRC in 2021 implies 2.6bn hours on TRC in 2021 (note that this implies TRC is 3.5% of total hours streamed on Roku devices). Roku in the past has said their target ad load is roughly half of TV (16 mins/hour), but we also know fill rates are low as Roku is still demand constrained on ad sales, so let’s assume 4 minutes of ads per hour (8 30-second ads per hour), which is roughly consistent with my experience on the platform. At $32 CPM (consistent with TV CPMs, despite better targeting and measurement on Roku), that implies Roku is generating $650-700m of revenue from TRC in 2021. Roku pays out roughly 50% of that to content providers, and has highly scalable fixed costs below that. Most importantly, since TRC sits within the Roku OS, Roku pre-installs this app and can drive engagement very cheaply, so customer acquisition costs are very low. (This is in contrast to Tubi and PlutoTV, which have similar revenue share arrangements, but then also have to pay significant amounts to acquire viewers, so their models are much less profitable at scale.) I conservatively assume TRC spends about $325m in opex in 2021 (30%+ of total Roku opex of ~$1bn), and so I think TRC is roughly breakeven today.

TRC revenue and profitability are now on the verge exploding. If you think…

1)      Total Roku hours streamed can go from ~72bn in 2021 to ~140bn by 2026

a.       This should be very achievable considering active accounts are still growing 17% Y/Y in 2021, and considering the average account streams 3.5 hours/day vs. the average household watching 7 hours of linear TV/day)

b.       Note: I think hours streamed will be far higher than 140bn and I am being purposefully conservative; consensus is at ~156bn hours in 2026.

2)      TRC can go from 3.5% of total Roku hours streamed today to 7% by 2026

a.       For context, Roku launched TRC in September 2017, so gained ~3.5pts of streaming hours within 4 years vs. expecting another 3.5pts in 5 years.

b.       This should be very achievable as Roku really ramps the quality of content going into TRC. This includes their Quibi acquisition and their plan to develop 50 original shows in the next 2 years.

c.       Additionally, Roku has several levers at its disposal to drive higher engagement. Perhaps most importantly is a TRC button on Roku remotes. I have not seen Roku TVs shipping with a TRC button on the remote. I believe this is because Roku is still demand constrained on CTV ad sales (advertisers always lag eyeballs…) and so there is no reason to drive higher engagement if monetization isn’t ready. I also believe Roku wants to build out amazing content on TRC before really pushing users into that environment. But when the time is right, Roku will make it very easy for users to hop in to the TRC environment by adding that remote button. And Roku also has the ability to adapt the home screen and menus to further push users into TRC.

d.       Note: again, I am being conservative. If the flywheel really kicks in, more and more high quality content will move into TRC and there is no reason TRC shouldn’t grow to >10% of hours streamed. If you go back through interviews with CEO Anthony Wood, he has discussed turning TRC into the home screen of the operating system, such that when you turn on the OS, you see content first, not apps first. From Anthony Wood in 2018: “We do think the future is a different UI, which is more content-focused. More recommendation-focused. And we have that UI! It’s called the Roku Channel…The Roku Channel is our sort of sandbox for building a next-generation, content-first user interface. And someday, when we think it’s ready and good enough and has enough content in it, it’ll probably become the home screen. But that’s not going to happen right away.” Suffice to say, if that plays out, TRC will have significantly more than 10% of hours streamed. For now, let’s just assume 7%.

3)      Ad load can go from 4 mins/hours today to 8 mins/hour by 2026 (still half of linear TV)

a.       This is totally in Roku’s control, and they are comfortable with 8 minutes/hour

4)      CPM grows at a modest 2.5% annual rate

5)      Content split stays flat at 50%

a.       I believe this may prove conservative as Roku generates higher effective net take rates on the content it licenses or produces (given Roku is taking on more risk, but also because Roku can subtly push viewers to view that content by pushing to the top of displays and therefore prioritize monetizing their own content)

6)      Opex doubles from $325m to $650m

a.       Costs below the content revenue share are largely fixed

b.       Again, there is very limited CAC in this business

…then TRC revenue will grow 9x from ~$650-700m in 2021 to $6.0bn by 2026, and TRC EBIT will explode from breakeven today to $2.4bn by 2026. At a 23% tax rate, that is $1.8bn of net income, growing 30%+ in 2026. At 35x P/E (dominant aggregator growing 30%), that is $64bn of equity value at the end of 2025, discounted back 4 years at 15% is $37bn of current equity value, or ~$250+/share just for TRC vs. the whole company trading at $180/share today.

And again, the above assumes TRC will only be 7% of hours streamed on TRC and gets you to a stock price 40% higher than the current level. Roku is actually very profitable on the remaining 93% of hours. I think ROKU will be worth $800/share (4.5x / 45% IRR) in 4 years.

Catalyst

Improved disclosure of TRC (no view on if/when this could happen)

Stable active account growth disproving the competition bear case