Roku, Inc. ($ROKU)

I'm excited to have Dan Jedda with us. I have said before it's first time, long time, first time Roku's been here. Thank you for being here. We appreciate that.

Thanks for having us. We're very excited to be here. It's my first time. So it's great.

Well, I don't usually offer praise, but I have to -- this is my first question, right? So you joined Roku, in May of '23 and 3 years in and your impact on the organization has been tremendous. It really has been -- we've been skeptical of the company you've proved us wrong in terms of the way you operated it. So thanks for being here. And given all that's changed for the better at Roku over this time frame, can you share now your biggest opportunities and priorities? Like now that you've done what you've done, where do you focus on to improve and what are your priorities?

Yes. Well, first of all, thank you for the nice introduction. Very thoughtful. It's been an amazing 3 years. I've just had my 3-year anniversary. And I'll answer that question, but let me just quickly give me a minute to talk about the migration of call it, over the last 3 years because it really has been transformational like, Roku is a fundamentally different company today than it was 3 years ago, not because of me, because of the whole company, but I've only been there for 3 years. And I want to just -- give me a minute to talk about that transition and then I'll answer that question directly. So if you go back to the second half of 2023, which is when I started, as I came in, Roku was a very device-centric company, still is with a growing scale of -- at that time, I believe it was around 80 million streaming households. Now we're 100 million streaming households. And so -- the focus on scale was there. There was less of a focus on the monetization of that scale, and that is the pivot that the company has undertaken over the 3 years. And so I'll give you a couple of examples, like in the advertising business, like Roku was very focused on its top, call it, X 200 big advertisers. It was direct sales force led. It was focused on getting more from the top brands, as many companies were. And then the M&E space, the media and entertainment space is space you're very familiar with, also as a vertical, like that was the focus. And that is a relatively small number of advertisers. You fast forward today, we're very diversified. We added more ad product. We diversified by integrating with demand-side platforms. In 2023, we had our own demand side platform called OneView at the time. So if an advertiser wanted to come via a DSP, we force them to come through our DSP, which again, was very limited. So we basically said, "Hey, let's open this up. Let's bring in more advertisers, let's skate to where the puck is going, meaning programmatic, the shift to programmatic, like let's be the best at -- on both the programmatic pipes as well as the direct sales led". And so we added more ad products focused on this. We integrated with all the DSPs. We tapped an entirely new market, entirely new TAM through the ads manager focused on SMBs, I'm sure we'll talk about that. We basically really focused on where this market was going. And from that standpoint, we're fundamentally different now. And then we did the same thing on the subscription side. It was not heavily focused on subscriptions in 2023 as we exited '23 now that we rightsized our cost structure, we said very clearly externally, we are going to focus on subscriptions. We're going to make this a real revenue stream that's going to grow tremendously. We added new subscription product. We focused on premium subscriptions. We brought in a leader of subscriptions to do just this. And we included in our home screen changes, we put personalization that drove subscriptions. So fundamentally, we're just a very different organization. And I'll just end by saying we reallocated capital to invest in these areas to grow them, not went out and hired a lot more people. We said, listen, we have plenty of capital. We really just reallocated to what we thought was our highest ROI initiatives. Now to answer your question, we're still focused on all that, increasing diversification of demand, going deeper within demand-side platforms, adding new ad products like on our home screen, looking at our new home screen and how it can drive both engagement and monetization, new subscription product, new countries, we just launched premium subscriptions in Mexico, we'll eventually launch more countries. We're building that now. So basically, we are focused on all these initiatives across the company, and that is where a lot of our allocation of capital goes, which again is just fundamentally different than 3 years ago.

Okay. Another fundamental difference, which I think was really important is improved transparency on revenues, it give us now advertising and subscription revenues. You eliminated it, you made it more easier to understand 606 revisions, why do you make those decisions? And there are other changes you see coming that you're focused on to make it more transparent and easier to understand company from the outside.

Right. First of all, the ad business is complicated enough, like you want to actually make it simple and easy to understand for our investors. And so one of the -- some of the changes -- we talk about this internally a lot, specifically within my org but even across the companies, how do we make it easier for advertisers -- for investors to understand our business. And so one of the reasons we did adjust 606 because it was extraordinarily complicated, and we're able to simplify that. On the segment reporting side, we've been working on that for about a year, very excited that we rolled that out. And even that, we took to the next level, like we even debated should we just give revenue out. But no, we said no, like we want to give revenue and gross margin out because it's going to help our investors understand not only how we've diversified our revenue streams, but also there's fundamental differences between these 2 now segments. And so we are constantly asking ourselves how can we give our investors and potential investors the information to understand our company, but also not give out too much competitive information, which I'm very sensitive to as well. So we dropped a lot of data points along the way. We're always talking about what more data points that we can drop so people can understand our business. And the segment reporting is something I was very excited to do. And it was not obvious, but I did feel that people didn't understand how big the subscription business was. And so we thought it was important to tell everybody. And so rather than just drop a data point on it, we said, "Hey, let's just break this out and let's start reporting on it". And then similarly, on the margin side, I constantly would say, "Hey, ad margins are not going down because we're blending out to 51% to 52%. That's down from 56%." And the common thought was lower CPMs are driving ad margins down. Integrating with DSPs are driving ad margins down. Despite my telling everyone that's not what's going on, you're having mix within the platform business, ad margins are doing very well. Integration with DSPs does not drive margins down. Lower CPMs for us do not drive margins down. Now we're showing that in the advertising business, of course, we just reported just north of 60%, which is up 450 basis points from a year ago.

Right. Is there anything that you want to tease that potentially is coming down the road in terms of changes in...

The only thing I'd say on this is we optimize very effectively for gross profit dollars and gross margin. Within advertising, we're getting very good. Not all impressions are created equal, not all DSPs are the same. And so we optimize our floor pricing across our inventory. We optimize our floor pricing across demand-side platforms basically maximizing the demand and the gross margins. We fill our highest margin inventory first. We have algorithms built to do this. So we're very good at optimizing for both revenue and gross profit.

Okay. I focus on this because there are so many companies we cover that have taken a disclosure away from us in the past year or 2. So it's -- we love companies actually tell us more and not less. So thank you for doing that. We also love companies that focus on free cash flow. And that's something that you've really honed in on since you've gotten there, right? So you talk a bit about the low capital intensity of your business, you also have an NOL that's benefiting you. How long will that NOL last? And kind of your -- you talked about $1 billion by 2027, '28 if not sooner, but give us your bridges on cash flow and why it's such a strong number for you versus maybe other companies?

Yes. That also was a change that we really focused on is making our absolute North Star free cash flow and free cash flow per share, which we've stated many times over the last 3 years. So we are very focused on free cash flow. So what does that mean? Well, it means that first of all, we are CapEx light, which we're very fortunate to have an accounting that's CapEx. It doesn't mean we don't invest, by the way, our investment is in R&D, which is not -- we don't capitalize R&D, we fully expense R&D. So we are very CapEx light. We are single-digit millions in CapEx. I don't anticipate that changing anytime soon unless we put in an ERP or something significant, but even then, it would be onetime, it wouldn't be ongoing. So we're very CapEx-light. We are one of the few companies, and this is why I'm okay, focusing on EBITDA versus operating income. We do focus on both, but we are one of the few companies that actually generate more free cash flow than EBITDA because we're so CapEx light. So our conversion of free cash flow -- of EBITDA to free cash flow is well over 100%. We have -- we're very effective at managing working capital, and we're also very effective now at managing dilution. So it is a North Star for us. And you're right, thank you for bringing up the NOLs. We have about $1.2 billion that's been impaired because of accounting purposes on our balance sheet. We'll write that up now that we're profitable, probably sometime in H2. But the point is we get to utilize a $1.2 billion net operating loss as an asset for us to offset cash taxes. How quickly are we going to use it? As quickly as we can because that just means we're generating a lot of profit. And I think it will be a couple of years, but for a couple of years, we will have EBITDA -- free cash flow above EBITDA. Even once we utilize it, and that will be a great day for us because that just means we are having a lot of profit flow through our business. The only difference between EBITDA and free cash flow would be cash taxes in that case.

And then priorities for cash flow, right? So you've done a little tuck-in acquisitions, don't have much CapEx. So you sit there and you have the benefit of this cash flow coming in, how are you prioritizing where you're going to use it?

Yes. Capital allocation in the form of OpEx and capital allocation in the form of cash is something I spend a lot of time on. What we've been focused on -- and you're right, we've done some tuck-in acquisitions. We're focused on, of course, anything that can help us from a monetization standpoint, from a scale standpoint, we look at -- there's nothing to talk about imminently in the form of acquisitions. We're very disciplined in how we focus on acquisitions. We're only going to do acquisitions that really make sense for us both strategically and from a positive ROI NPV perspective. But we are also -- we also did -- 2 years ago, we basically changed our vesting of RSUs to be net share settlement. That's where instead of paying cash taxes in stock, we pay them in cash, quasi-stock purchase -- repurchase program. So that offset 1/3 of dilution. And then last year, we had a share buyback, we authorized a share buyback and approved a share buyback. And we've been utilizing that to fully offset dilution. I think Q1 was the first negative dilution in the company's history.

Yes, sequentially.

Yes, sequentially. And so -- sequentially, correct. And so my goal -- I stated this, my near-term goal is to fully offset dilution with cash, as you said -- I said publicly, like I believe we can do $1 billion of free cash flow by 2028, if not sooner. And so when you're generating $1 billion of free cash flow, we can use some of that to offset dilution. We'll continue to do that. I might even go beyond that at some point. We'll wait and see on just what's the opportunity from the cash...

What I also appreciate is that you're old school focusing on GAAP EPS, GAAP EBITDA, right? And so historically, Roku had been a big issuer of SBC, right? So how do you think about the right balance of stock-based comp and why not pivot to more GAAP-based targets, right, and not adjusted EBITDA but there's no GAAP EBITDA definition but moving to include SBC and kind of all the key metrics that I communicated?

Yes. We do that now internally. We treat stock-based comp as a very real expense, is embedded throughout everything we look at from a P&L ROI perspective, we look at stock-based comp. So it's something that we've -- again, coming in, I just thought it was high on a relative basis. It was driving a lot of dilution. We've rightsized -- when we rightsized our cost structure, we rightsized our stock-based comp. We made some other changes to manage our stock-based comp. And so I believe stock-based comp in 2024 was just north of $375 million. We improved it by $25 million in 2025. Our guidance implies another $25 million improvement in 2026. That's despite adding headcount. So we feel like stock-based comp is now well managed, but it's always considered a real expense. So as we do focus on operating income in addition to EBITDA. We do focus on net income in addition to operating income in addition to EBITDA. So -- and then of course, all this equates to how does free cash flow and free cash flow per share which stock-based dilution that's why we say per share dilution is a part of that. And so every metric that we look at is across all those financial metrics, which again is relatively -- which is a change from, call it, 3 years ago.

Which is -- we're in favor of. So you just touched on it before. I think one of the areas of the company that seems less discussed is the international opportunity. I know it's relatively new. What have you learned as Roku has garnered more scale outside the U.S., right?

Yes. So let me talk a little bit about international. International is a huge opportunity for us, and we're at different stages in different countries. So -- let me just give you a couple of examples of countries like, in Mexico, we have nearly as much scale in Mexico as we do in the U.S. I mean it's very impressive what Roku has built in terms of scale in Mexico. The ad market is nowhere close to where the U.S. ad market is in the shift to digital. So CTV ad market still has a way to go. But it will go more to digital because all the hours are shifting just like in the U.S., all the hours are shifting to CTV and to digital. So the ad dollars will eventually follow. It's just taking time, but we're in a great position. But where we're starting to monetize very effectively as we wait for -- I mean we are doing well in the ad market. It's just not where I think it should be on a relative basis. The market -- the ad market needs to catch up to where the hours have gone. But where we are really focusing now is on subscriptions, again, which is relatively new for us to really focus on subscriptions. We've always had the subscription business. Roku Pay has been around for a long time, but we've not had a huge focus on driving the monetization via subscriptions. And we're doing that now in Mexico. We launched premium subscriptions this year, in Mexico, it's doing well. We launched with Apple, for example. And we have other premium subscriptions, and we're adding more all the time in Mexico. We launched Howdy in Mexico doing very well with a great content offering at a relatively inexpensive price, I believe that the equivalent is around $2.60 in the U.S. So we are driving the subscriptions part while we continue to wait for the ad market to shift, and we're getting better and better end market. M&E has always been there, but the in-stream video has taken some time to move. Transition over to Canada, we have less scale, about half the scale in Canada that we do in Mexico, but the ad market is very -- doing very well. And we've really been focusing on driving advertising and subscriptions in Canada. We don't have premium subscriptions yet in Canada. That will happen. But we do have a robust ad market. So we put feet on the street, people physically located in Canada, we're driving that. The ARPU looks great. It's 4x the ARPU of Mexico with less than half the broadband penetration just because of the ad market and how it's driving. And I think there's incredible opportunity. The rest of the countries, U.K. and Brazil and the rest of Latin America, we're just still building scale. We are -- there's the U.K. we are starting -- we are monetizing that. We just don't have a lot of scale yet. The rest of Latin America and Brazil, we're building scale. The good news is because we're first -- we're a first mover in many of these Latin American countries. The scale that we're building is just not an expensive CAC, and I love that. We're getting ahead of the game, just like we did in the U.S.

Right. Well, can you -- I get the North America strategy, LatAm, what about like Asia Pac? Is that just too hard because you've -- a lot of incumbents there.

Yes. We constantly look at countries outside of our focus countries. These are the countries that we're launched into. So I don't have much to add. We are looking at it. We look at Europe a lot, we look at Asia. We look for the opportunities. I don't have much more to add on that yet except to say it is on our radar. And that really is about distribution. It's not so much the operating system. That's fairly easy and not a big investment to convert the operating system to be compliant, what we'd want to do in Europe or Asia. It's about distribution and what is the cost going to be? You can't -- you don't want to be a #5 player in this space, and you know this, right? Being a #5 or #4, you need to be #1 or 2 or 3 player to get massive scale as an operating system to really drive the monetization. So we're constantly asking ourselves, like, how do we get significant scale in these countries.

Got it. Okay. Let's turn to advertising. You mentioned this before, too. Can you help us frame what's under the hood of the ad business. You have so many different types of products and now so many ways to go to market. So what are the KPIs that you think are most important that you can share with us? And any kind of metrics against those KPIs?

Yes. The KPIs -- it's a great question. So one, we focus on controllable inputs, first and foremost, not the output. The output is what it is. We have a relentless focus on what we can control. And I'll give you a couple of examples on how we think of that. Like for example, like if you look at just broad brand-based advertising, reach matters. So a KPI is reach within the Roku channel and reach across the platform, focus on both. And we've really started to focus on that. Again, as we entered into 2024, we said, what are the KPIs that are going to drive the monetization of the ad business and the subscription business. And reach was one of the important metrics that we could control that we wanted to drive. And we've seen tremendous improvement in reach. TRC is the #2 app on the platform. As you might notice it, we used to call it a top 5 app then we called it a top 3 app in the U.S., now we're saying it's the #2 app. So that just tells you how much reach and we look at daily reach of course...

And reach has declined in the linear world tremendously except for broadcasting which -- come in. The reach is no longer available.

Reach has grown significantly for us. Again, not just in the Roku channel, although that's critical for us. But across the platform because we get a share of inventory from others that we can serve our ads on. So we've really focused on reach within the, call it, the brand marketing budgets. You flip all the way down to the SMB market, very different market. You do not really focus on reach, what you're focusing on are traditional KPIs of a number of advertisers that start an account with us all the way down to the conversion to a paid advertiser, the repeat rates of a paid advertiser because these are SMBs. These are things we control. We control the UI, we control the funnel, we can get improvement in conversion rates. We have other ways of, of course, how we bring advertisers in, there's a CAC involved with marketing et cetera. We'll sign agreements with larger companies, medium-sized agencies to drive volume. But ultimately, we want to control the KPIs within the funnel of conversion. And by the way, performance because it is a performance-driven market is one of the most critical inputs. That's something we control, like the performance of our platform. So we are constantly working on integrations, maybe we'll talk about Ads Manager. But the KPIs are what we can control to bring more SMBs and get them to stay with us. So those are the KPIs we focus and that is the exact opposite end of the ad market. Now we've talked about the largest 200 brands all the way back down to the SMBs and then everything in between.

Right. So I think a major pivot was embracing third-party DSPs, you again mentioned that upfront. Was the Amazon tie-up or tie-in the most meaningful given the leveraging of Amazon's first-party data. So was that a major unlock from the outside looking in, what do you think?

Yes, it's a great question. Yes, it was a major unlock. I was so excited when we could announce the Amazon deal because for the prior year, we had heard a lot about the walled gardens are going to rule. Everybody else, game over. I cannot tell you how often that was being repeated. And we were certain the industry was wrong on this, that the walled gardens are not going to stay walled that they're going to open themselves up. And if they do that, what is the best way to open yourself up is come to the platform that has the most scale and not by a little, by a lot in the U.S. with our over 50% broadband penetration. So Amazon was a walled garden. There was others, people said YouTube was a walled garden. You fast forward now DV360 is getting more into CTV. We just signed an agreement with DV360 to adopt their unique identifier of the hashed e-mail. So our first-party data can be matched up with their first-party data. The Amazon deal, which we announced about 9 months ago was utilizing our first-party data with their first-party data. It's a win for Amazon. It's a win for Roku because now they can recognize that it's me, match me and my purchase profile on Amazon. This is all done in anonymized clean room, very safe for Roku. Our data is very safe. And we ultimately can help Amazon expand their demand side platforms in the meantime, we benefit from the media side and the Roku channel, and we benefit from a platform level as they pay us at the platform level. So it's a huge win for both organizations. That relationship is going incredibly well.

And clients love it too because you can use first-party data from Amazon...

100%.

Not paying a huge toll to a third party?

Correct. They would not come through it if they did not see the performance on their side. So we give Amazon instant scale. They give us a way to monetize our platform, 1P data plus 1P data, a huge win.

So one of the questions that people have the time was how do you make sure that the DSP doesn't undermine your -- your first -- your go-to-market directly relationship we have with the brands or the advertisers. So how do you manage the potential conflicts in channel?

So we have -- first of all, we have ways that we protect our data, first and foremost. So our 1P data is fully protected. I won't get into the details, they're complicated, but this is something we take very seriously. So that's one piece. The second, to your point though, it's not so much of cannibalizing. It's meeting the advertiser on where they want to transact, and Charlie is really good at discussing this. So essentially, our thought is like we don't want to force an advertiser to come through a specific DSP or directly to us if they don't want to. What we want is we want the most performant media out there. And because we have 100-plus million logged-in users with all the data behind it, we are a very performant platform. So we've solved and will continue to get better even. But we solved this idea. We know who our users are. We have the data behind it. Then we say the advertiser like, again, some of the top 200, they want to transact through a single DSP. So we're not going to come and say, no, we're going to say, okay, we'll partner with that DSP because we know our media is very performant. And we know the advertiser wants to run on our media. We know we're integrated with the DSP. We've integrated as deeply as the DSP is willing to go with us and we're always looking to go deeper. So it's not a cannibalistic function at all. It's a meeting the advertiser where they wish to transact. We have seen no indication of cannibalization on Roku by advertiser, by vertical, every which way we've looked at it. There might be cannibalization within the competition of demand-side platforms, and that's out of our control. That is a function of the demand side. That's why we're integrating with all of them, and we want to go as deep as we can with all of them.

Another question came up on the post call is, look, as you add DSPs, is that a onetime bump or that we worry about a comp? Or is this going to be an unlock of a higher revenue stream into the future. So how do you think about maybe the comping of some of these big DSP additions to your system...

I think the -- it's not a onetime step-up. These are all gradual because the way these DSPs work is, they're all focusing on whatever their unique character is. Some of them are focused on top 200 brands. Some of them are focused on the next 200 and some of them are focused even on like app installs, which I would say is the lower CPMs. So they're not onetime bumps, what happens is that you plumb into the integrations and you're constantly optimizing across the DSPs. Also, we're always becoming more performant because we're using our 1P data. We're having new ad product. We're trying to ingest more data. We're getting more signals to become more performant, whether it's the auction, whether it's a one-to-one deal, we're always trying to be the most performant platform. This all means that as the shift to programmatic happens, and it is happening, whether it's top brands, SMBs and the torso in between as the shift goes to programmatic, we're going to play at every piece of the CPM demand curve. By the way, huge difference than 3 years ago. We were only at the top of the CPM demand curve in in-stream video. Now we play along the whole aspect of the CPM demand.

As someone who's covered linear for a long time and now digital for more than a decade and you guys and Trade Desk, you realize the value of SMBs, right? So talk to us, you've hinted at it. What do SMBs want from you or what they need from you to bring that money because that's the differentiator on Meta and YouTube and Google is just the vast size of that advertiser?

Yes, yes, it's huge. So we've looked at some of the -- some of the TAM on this. And we've stated like there's a market out there that says SMB spends $600 billion on total advertising. The digital piece of that is a subset, call it, $180 billion to $200 billion. I'm not sure exactly how much. But imagine if 10%, 15%, 20% of that can get carved out and moved over into CTV. So you already have -- and you mentioned this, you already have linear shifting to digital. So the total linear plus digital, $90 billion. About $33 billion has shifted to CTV, but that CTV piece is growing double digits even though the total TAM is growing low, low single digits at most. The CTV piece growing double digits, that's $33 billion. Now you have this entire new TAM that's shifting over into CTV. Why are they shifting -- why didn't they shift over sooner? Two reasons. One, the performance capabilities weren't there because this is a very performance-driven. They're not trying for reach. They have specific KPIs, site visits, conversion, calls to my call center, et cetera. That has been solved, not fully solved, but greatly improved with CTV. The second reason was it's very difficult to create a 15- or 30-second video. And GenAI has removed that impediment as well. By the way, if you didn't do it before, you had to go to someone and say, "Hey, I need a video that's going to cost $30,000. I don't have $30,000. I want to spend $30,000 on media not on creating a video. GenAI has not only made the creation doable. It's -- it can be done very inexpensively. So you've got the 2 largest impediments being fully removed. The shift is happening on CTV. We have a product called Ads Manager, which is a self-service, we have some inside sales focused on it. But by and large, it's a self-service click-through where it takes 5 clicks, we're about 5 clicks, where you can create an account, say what you -- how you want your KPIs that you want to advertise, say, where you want to advertise, do you want to do it just on the Roku Channel, do you want to do it across the platform. We let them choose. And then you can upload a video or we will help you create a video, you hit publish, you're off and running. And so I'll give you an example of where this has been very successful. And it's just going to get better and better and better like we had some recent success with insurance companies of putting a 1-800 number on where we would run a video, the 1-800 number would show up, and they could track who saw the ad and then who actually called because of the way we can API and pixel sites. So that -- their performance was we want to track how many calls we get based on our ads. We can now do that in a very effective way. We can do conversion data through integrations with Shopify. We can pixel sites and do site visits. If you want to actually measure site visits, we can do that. And it's getting better and better. And I'll end by saying this way. This is -- this product is only going to get better on performance and the ability to create a video. I was just looking at the funnels that we have. We start with a huge funnel of people who start an account to people who actually say what their KPIs are, they say where they want to advertise across from a publisher perspective, the biggest falloff we still have in that funnel is upload or create your video. So as GenAI improves this, and it will improve it, it's just going to get better and better. And our goal is, as a KPIs, how do we get that conversion higher and higher and higher through having more templates, more integrations with other GenAI companies to help the SMB create the GenAI video.

Got it. I have to pivot to subscriptions. I think and you tease it, we're surprised by how large the business was and also how fast has been growing, right? I think if you sat in the office of MoffettNathanson, we would debate whether or not subscriptions are mature as an industry, but you're growing wildly fast. So what are you doing to basically -- I know it's relatively new sector -- segment for you, but what are you doing to tap into that demand? Because in aggregate, it looks like it's slowing, but for you, it's accelerating.

Yes. So a couple of things going on. First of all, we've always had a subscription business. We have said in the past and this is before premium subscriptions became a real important strategic initiatives as we monetize tens of millions of subscriptions. That's true. It's growing. It's growing very well. The focus -- the pivot that we've done is really focused on premium subscriptions because premium subscriptions are very strategic to us because you can embed the content throughout the user interface. So if you are a -- if you have what we call direct-to-consumer or D2C subscriptions, you have basically 2 ways to get to the content. You either go through the app tile, maybe you have a button, you can go through the button. And you might show up in personalization, but if you are a premium subscription, you are going to show up throughout the UI. So we ingest the content, we put it throughout the UI. So we're driving ingress into that content partners app throughout the UI, whether it's the personalization row, whether that's the left nav, could be through a sport zone, it could be through other zones that we have. It could be embedded in content tiles that we have to drive traffic to the sites we monetize. So the ingress is all over the place. It's not just the app or if you have a button, the app plus a button. It's throughout the UI. That's very strategic for our partners. It's very strategic for us. So we're getting more and more partners who come on in premium subscriptions. I'll also say that there's a thought that there are -- there's more consolidation in the subscription space where I'll call it the mega apps are where most subscriptions are consumed. And think of that mostly for yes, Tier 1s, and there will always be Tier 1s who will not do that. But there'll be a lot of Tier 1s and then the whole torso and tail of which there are many. They do not have the ability to show you what their apps are and what their content is. So it's going to be done in these what we call like premium subscriptions or Amazon channels. And that's what we're focused on. And that business, we're actually gaining share in that business. We're growing in direct-to-consumer as well, but with specifically in premium subscriptions, we are gaining share of the overall channels business because we're focused with new ad product, personalization, driving subscribers into that content throughout the UI.

You also -- you basically bought a company and you created a company. So why did you -- when you think about Frndly and Howdy, which built from scratch, what drove the decisions and how those new assets faring under your management?

Both are doing very well. We're really happy with the Frndly acquisition. We're very happy with Howdy, it surpassed our expectations in terms of growth. And the reason why is, again, because we have the UI, because we have the platform, we feel that we can drive subscriptions. And if there's a subscription that we can own and operate, we will take a look at that. Now Frndly is both on and off Roku. And now Howdy is both on and off Roku. So with Howdy, we felt there was a niche of a low-priced ad-free content offering that many streamers would love to watch. And so we basically built the product and launched it, and it's doing very well. As soon as we learned a lot on Roku, we launched it off Roku, it's available on Amazon channels doing very well there. And it's also in Mexico. And that's a product where we can launch in many countries, including countries we don't have the platform in. And we believe that there is this offering. Also, we believe that AI content long form, could change the trajectory, could change the business dynamics of SVOD services. And if we're right on that, what a great opportunity to have a low-priced offering that can take on long-form AI-generated content when it's ready, and we can monetize that.

Yes, really disruptive. Gross margins lower than we thought, but there are some reasons why the trend has been against us. So what's going on with gross margins? That's just going to be a structurally lower business in average?

Are you referring to platform gross margins?

Yes, yes.

So platform gross margins. One of the reasons we bifurcated subscriptions and advertising down at the margin level. So we could say very clearly what the margins are on our subscriptions, which is a bit of an annuity stream and then what it is on advertising. And so -- on the subscription side, it's right around 40%, just north of that. Premium subscriptions, which is slightly lower gross margins being the biggest driver is what's driving that down. But it's a lot of incremental revenue. It's a lot of incremental gross profit. Those margins, I think, will be around 40% for the rest of this year. We also have some high-margin business in our subscription business like D2C. On the ad side, again, this notion that all these activities that we're doing are driving margins down, it was -- despite us saying it all the time, now we show it ad margins at 60%, just north of 60%, up 450 basis points year-on-year. I think it's going to stay there. I think there's an opportunity to potentially grow it. We've got a lot of new ad product. We've got the home screen, which is very high margins that were coming out with the redesign. There's opportunity to drive more higher margin ad product in the home screen. We're getting -- we're very efficient at optimizing all our inventory, the highest inventory that we have. We optimize first. And then, of course, we focus as much on managing the economics of it. So we always optimize for campaign, completion campaign goals. But then after that, we optimize for gross profit dollars.

Okay. I have a minute. I have to ask you about devices, I have 2 questions for you. One is, what do you see as your moat versus others in the space? And then do you worry that some -- especially those with a growing ad businesses, will loss-lead TVs even more in the future.

So yes, the moat is -- first of all, we have a brand that streamers actually love. That is the moat is streamers love the Roku OS. They love the simplicity of it. They love the remote, they love the functionality. It's an amazing product that the team has built. That's one moat. The second moat, which is a less known but equally important is our BOM cost that we have because we have a very low memory footprint relative to everyone else in the U.S. space, our memory footprint is lower, and that was intentionally done by Anthony to basically facilitate an advantage to our players and our OEM partners and our first-party TVs when we're the hardware manufacturers. So the BOM cost has always been there in a rising memory cost environment of which we are in, the BOM cost gets more and more. So we are seeing more OEMs coming to us saying, "Hey, in addition to having this awesome OS, you provide a BOM cost advantage that no one else can provide to us. Can we do more units with you." So that also is a very effective moat for us. And something that I think the team had the foresight to be very smart. And remember, it's a purpose-built OS not taking other types of OSs and making it available to CTV. We purposely -- Anthony purposely built this operating system for CTV.

Okay. We ran out of time. Dan, thank you so much for being here.

Appreciate the time. Thanks, everyone, for coming.

Thank you.

Thank you.