Magnite, Inc. (MGNI) Earnings Call Transcript & Summary

Hi. Looks like we're live. I'll make a quick intro. I'm happy to have Michael Barrett, David Day and Nick Kormeluk here with us from Magnite.

And we'll dive right in. Michael, if you don't mind, could you take a minute to describe the business for anybody that's kind of new to Magnite and doesn't really understand all the big working pieces of the digital ad ecosystem?

Sure. Happy to do so, Kyle, and thanks to the audience. So Magnite, it's an industry [ riff ] with acronym, right? And so we're referred to as SSP or a sell-side platform. That acronym is also interchangeable with exchange. So when you hear people refer to themselves as the exchanges, generally speaking, SSP exchanges. They work for the publisher. They help publishers like Hulu, Disney, Business Insider, Wall Street Journal take their digital ad inventory and put it into a platform in exchange where buyers like The Trade Desk called the DSP, demand side platform -- Trade Desk works with that agencies and marketers, and they put their programmatic ad dollars into the Trade Desk console. Trade Desk plugs into our exchange, the Magnite exchange, and they're -- that enables them to be able to sift through the inventory with the data that they have to find the right user in the right environment and bid on it in real time. And so we conduct that auction. We aggregate all the inventory from thousands of publishers. So it becomes a meaningful experience with the buyer. We can segment that audience. We can create audience segments and profiles to make it easier for a buyer to find the person they are looking for. We make sure the inventory is fraud-free, free of bots, high inventory quality. And we also are the payment collector. When that auction is consummated, we take the money. We take our slice out of it, which we call as our take rate, and that's how we get paid. And the publisher pays that. And then we pass on the rest of the check to the publisher. So it's kind of at the 1,000-foot level, what we -- how we play in the ecosystem.

Very helpful. And you're in a fairly crowded segment of the ecosystem. You guys talk a lot about supply path optimization. That process that's going on there. But while we're thinking about this at a high level in simple terms, could you just frame up simply as you can how you win business from your competitors? And then I'm going to probably ask you programmatic, so be thinking in the back of your head in advance.

Yes. Sure. So there's been several phases of programmatic advertising over the last 15 years of its inception. And in the early days, it was mostly about being the exclusive SSP and having exclusive access to inventory. So maybe BBC would only work with 1 SSP. And if you wanted BBC's inventory, you had to go to that SSP. And so that was the first phase of SSPs. The battle was to get exclusive inventory, and that would differentiate you. The middle phase was all about a unified auction. Publishers were like, well, that's kind of inefficient because it's still -- there's still my direct sold, and then I'm creating this waterfall where Rubicon Project gets to see my inventory first. But then if they can't sell it, they pass it on to programmatic. And programmatic can't sell it, they pass it on. So literally a waterfall in no intelligence as to who gets to see what inventory is just whether it got sold or not. So the publisher said, let's -- that's crazy. Why don't we put them all in demand? They're all in a level playing field, and it's a unified auction. It's a jump ball. So I'll send my impression to everyone simultaneously and let anyone bid on it and see if they can win. Obviously, that kind of democratized access to inventory. So it took away that value proposition of, hey, I'm the unique -- I have unique inventory from BBC. You need to buy from me. So then what started to differentiate how SSPs can separate themselves. A lot of it was tools and tech. Some of those tools and techs were built for the buy side to make it easier to buy. And ultimately, the number of players expanded for a moment in time, but then the buyers start to scratch their head saying, why are we trying to buy the same inventory from 15 players? I'm only going to do it from 5. And that began this whole path of supply path optimization. And so who were the guys that survived? They were the scaled guys. They were global. In our case, independents mattered greatly because, hey, I didn't want to give out my money to Google who has a DSP who's competing with me. And same with publishers. Publishers didn't want to give all their inventory to Google who runs in ad exchange because of the idea that they have -- they own and operate inventory. So they're competitive with publishers. So little by little, it started to wind its way back down. And at that point, I think we had done a decent job of becoming the largest general exchange as Rubicon Project. And then there are the specialized exchanges, and CTV was one of the fastest growing, most attractive. But knowing and put them together, knowing and put a general exchange with connected TV capabilities, and that's what the marketplace was crying for. They were like, hey, we need something that looks like the Trade Desk on the sell side, someone that can do it all, that's independent, that's global, that's scaled. And we could answer 3 quarters of that, but we didn't have CTV capabilities at Rubicon. Telaria could answer the CTV piece of it, but couldn't answer -- do everything. So it made sense to merge. We did close on $4.1 million. We're 2 quarters into that journey. And we are clearly differentiated as that player that can do everything that's independent, that's scaled and feel really good about our path going forward.

That's helpful. I'm getting a lot of inbound on programmatic S1. Any kind of high-level thoughts on your business versus theirs, puts and takes, similarities, differences?

Yes. No. It's kind of -- I've known Rajiv and his team since the early days -- it was one of the early members of the company called AdMeld. That was one of the first SSPs that we sold to Google and competed head-to-head with PubMatic. Actually, Rubicon, PubMatic and AdMeld all came out in the same year back in '07, '06. And PubMatic always -- it was born from kind of an ad network based in India. So they always had a pretty low-cost basis. So kudos to them. Never really known for kind of software innovation or build, so you kind of see that in their line of R&D. So I think Rajiv has always run a very efficient operation, has played, generally speaking, in the longer tail mid-torso of the publisher universe. And down there, you're allowed to charge probably higher take rates, so that helps the margin business, but it comes with it some challenges in inventory quality. So it's something that we've kind of issued as Rubicon and now Magnite. We always want to be top-tier enterprise level, good brands, great quality inventory and have never really run into any problems in that area. So yes, I mean, I think that their -- they play a role. I think there'll be -- like, we found at Rubicon, not having connected TV capabilities really matters today and will really, really matter going forward. And when we looked at the build buy partner scenario, we kind of ruled out partnership just because it's too important, and the build was onerous. It was going to be 12 months lag. We didn't even have the talent. We'd have to find the talent. A year later, everyone else would be on version 4.0, and we'll be on 1.0. And so we thought that buy was the way to go. So -- and then it'll be interesting how they address that challenge.

Great. That's helpful. Maybe we can switch over, talk a little bit about your recent results. Could you take a minute, either you or David, kind of talk about the same-store pro forma growth rates for the desktop, mobile and CTV businesses and maybe dive down in any meaningful unit -- pricing unit, volume trends that stood out in 3Q?

Yes. I'll take that. Yes, we were really pleased obviously with our Q3 results. We grew on a sequential basis 44% Q2 to Q3. From a component perspective, CTV grew at 51%. On a pro forma basis year-over-year, our mobile channel grew almost 16% year-over-year. And our desktop channel was almost at the breakeven level. So those results, I think, were definitely at the high end, if not exceeding the high end of our expectations. From a component perspective, it's kind of interesting, from a unit volume and CPM perspective, our business is more dependent on the total ad spend. And so what you often see as you have CPMs perhaps accelerating, you see maybe a relative deceleration on unit volume growth or sometimes vice versa. It's kind of like squeezing a balloon. You're going to somewhere else. But what was interesting, the recovery was so significant in Q3 that we actually saw pretty healthy increases in both unit volume and CPMs. Also on the CPM front, political, it's a -- it was a smaller part of our business. It never really exceeded kind of low to mid-single digits, but it also had an impact in that they tend to pay extremely high CPMs to guarantee placement. And so in some ways, it drags along other CPMs. In other ways, it causes folks to sit on the sidelines to not compete in that environment. So those are a couple of trends that we saw.

On the political side, some of my broadcasters told me that their political was twice their core in October. So I would imagine that sent some local advertisers looking for someplace else as well, probably in September and October. But I can't prove any of that, but it makes sense. Is there some optimal mix in your mind of kind of desktop, mobile, connected TV going forward? Or if we get 3 years out and desktop is still a significantly large piece of the business, are we just as happy and high-fiving just as hard as we would if CTV became 85% or something of the business?

Well, I think we'd be thrilled if desktop grew at 50%. So -- and that scenario will take the same proportions. But I don't know, Michael. I think, obviously, we like growing CTV, and I would hope that, that would reach 30-plus and beyond in the future. But at the same time, the rest of our business is extremely important, and we're not just regarding that part of the business by any means.

Does mix have any margin impact as you look out? Or is it all kind of the same platform leveraging the same infrastructure?

CTV tends to have a lower take rate. It's primarily PMPs or private marketplaces, and that's where a publisher and a buyer have a relationship and are helping to bring part of the demand to the table. And so -- but the great thing about that is, typically, PMP deals are much higher CPM. And so even though you have a lower percentage of the transaction, your actual revenue per transaction is typically much higher.

That's helpful. I think it's -- I think we're at a moment in time where it's a little hard to get a good margin read. Really nice sequential ramp-up of a very challenged June quarter. But with the deal kind of looking backwards in the guide, I think it's kind of 30% 4Q adjusted EBITDA margin. How should we think about margins looking out just roughly? And then remind us what the seasonal impact is in a more normal-looking world, God willing, that's what we get in '21?

I hope so. Yes. So from a seasonal perspective, we typically run about 30 -- low 30 percentage of our revenue in the fourth quarter. So we have kind of a traditional holiday seasonality that a lot of retailers and others have. From a margin perspective, let's take last year pre-Telaria merger. So Rubicon Project -- legacy Rubicon Project. We ran a 16% adjusted EBITDA margin in Q3 and a 32% adjusted EBITDA margin in Q4. Now adding Telaria to the mix, that's only going to help our margins over time. And as you mentioned, we ran a 23% margin in Q3, and our guidance would imply about a 30% margin in Q4. And so we've been -- we've always targeted a 20% top line growth rate and a 25% or greater adjusted EBITDA margin. And so we've been tracking at a very healthy pace to those levels. I think, like you mentioned, of course, COVID has put a little hiccup there, but I think you can see that we're back on pace to get there. I think a few factors to consider for next year. We did get a little bit of a boost on our margins because of some temporary COVID-related cost savings. So you don't have nearly the travel. You don't have some of the marketing events. And as you mentioned, at some point in 2021, hopefully, some of those will come back. So you've got to factor that in. We've got a growing business. In particular, our CTV business runs on the cloud. And so as that grows and scales, you'll see a little bit more of a variable nature in our cost of revenue for those cloud costs. Whereas our legacy platform -- the non-CTV platform is CapEx supported on-prem, and those infrastructure costs tended to run through CapEx and then an amortization kind of profile. So I think, if anything, with the scale and the merger with Telaria, we have probably higher aspirations from an adjusted EBITDA margin on a long-term basis and even the potential we had premerger.

Great. Michael, you mentioned that some of your peers run kind of thinner R&D. David guided to higher EBITDA OpEx in the second half of this year. Is there anything that we should be thinking about that could come in from an investment perspective at any particular line as we think about next year?

I think we've been pretty clear that we've reached within a plus or minus band of investment in the company. And I think we'd be very clear if there were something special on a onetime nature that we felt was a perfect trade off for growth in return for a hit on margin temporarily, but I don't think that, that's something investors should be overly concerned about. Is that something David would concur as the CFO?

Definitely.

Okay. And you mentioned political briefly at kind of a mid-single-digit contribution to total revenue, and I think Trade Desk echo that. It really doesn't feel like the open Internet got its fair share of those dollars. They're obviously hyper cyclical, hyper volatile depending upon which markets and which rates is light up, but is there anything -- what barriers or hurdles need to get removed between now and 2022 so that you guys can get your fair share? Because my broadcast television guys are talking hundreds of millions of dollars, and it just seems like you guys have a really strong play here that hasn't kind of showed yet.

I don't know if political is so much different than any other ad category. Obviously, it's wildly different in terms of its disappears, comes back, disappears. But generally speaking, whenever there's a new media opportunity, the eyeballs arrive before the ad dollars do. And so I think it's just -- we're under -- CTV is under share in terms of a lot of spend. CTV is under share in terms of every spend. So I think that what we saw happen this year was nothing short of phenomenal in terms of the amount of cord cutting and folks adopting streaming platforms, particularly ad supported, given the economics for the consumer. And it's a lag. Advertising, it will catch up. And I think it just -- it takes time, and that's been the story with the Internet all along, hand ringing about brand advertising in the early days of the Internet, like when will the brands believe in the Internet, and that's come to pass. And so I'm not overly concerned that there's something systematic about CTV that isn't conducive to political advertising. As a matter of fact, we showed in a study we did how impactful it can be. And keep in mind by 2022, there's going to be 50% of households that won't be able to reach in linear. So it won't be one of these corner cabinet kind of buy situations or a sample. It will be a must buy.

That makes sense. So let's back out from the small, which is the single-digit political, and talk about the big, which is the targeting, tracking and identity in the open Internet. It's my favorite new topic. I'm obsessed with it, as Nick has probably figured out by now. Sorry, Nick. I'm a -- I'm assuming that you guys heard Scott Howe's commentary on publisher lift using ATS in their September quarter. He gave some pretty eye-popping publisher yield increases even with the Chrome inventory with third-party cookies. It seems like for now, this whole scenario has flipped 180 degrees from Tom Kershaw, our CTO, scaring people the death of the 70% unit pricing decline earlier this year. And now we're getting lift from ATS, even in Chrome inventory. It seems like we're now positioned to come out the other side of this better off, and it happened so rapidly. And we haven't even gotten the UID 2.0 system in place to comp yield versus ramps ATS. So if you could just -- I'll just kind of tick down a few questions here. Have you tested ATS yet?

Yes. We work with LiveRail, correct. Yes.

And do those yield numbers that he was reeling off on his call resonate with you?

We have nothing to say. I mean, I think like -- just like we do with our earnings, we take the highlights and polish them all up. So I think that, generally speaking, it has been a very positive experience across the board. And to your point, in terms of the larger picture, really satisfying that these initiatives that have been talked about for some time are really getting up and in production and working long in advance of deprecation of cookies.

So great segue. What milestones do we need to see in the marketplace so that we're extremely comfortable that UID 2.0 is a viable solution by the second half of next year? What do I -- do I just need to keep looking for people that are willing participants and just add up all of the press releases? Or are there some proof points that I should be waiting to see?

Yes. It's a really good question. I mean, if we -- so the consumer experience is going to be such that in order for that to work, you as a consumer are going to be very aware of the deployment of it, right? You're going to come to your favorite site, and you're going to be prompted to register -- to log in. And you may choose to do it. You may not choose it the first time. Second time, they hit you, third time. But in order for it to be successful, you're going to have a critical mass of users agreed to log in and then that triggers the viability of the program. And we support it, and we're leaning into it. We kind of don't think that it's going to be the only solution, that there'll be plenty of consented users that don't want to log in. And then what are you going to do with them if you're a publisher? How are you going to monetize those users? And that's kind of the problem that we talked about that we're setting out to solve with first-party publisher data, creating audience segments, standardizing the taxonomy through Prebid and federating it over thousands of publishers so that there's this rich data pool of users who aren't logged in but still are consented so you can target them. It all comes down to the fact that browser companies and OS systems like Apple have all agreed that a consented user, first-party cookie from the publisher is perfectly fine going forward in this privacy-first world. And that certainly shows a shift of the first phase of programmatic where the buyer created all the segments, drops third-party cookies, carpet bomb the world with them and track users to now the publisher regulating the control of cookies. And we think that, that's a great value exchange for publishers because they've been undervalued in terms of their role in this. And as an agent of the publisher, we think it bodes well for Magnite.

Another good segue. It's almost like you've seen my questions. You're...

They're found in my den that I got kicked out of. So I could [indiscernible].

I wish everybody else has seen you get chased out. It made feel better to see a CEO getting phased out.