Magnite, Inc. Q4 FY2024 Earnings Call

Good day, and welcome to Magnite Fourth Quarter 2024 Earnings Call. All participants will be in a listen-only mode. Please note this event is being recorded. I would now like to turn the conference over to Nick Kormeluk of Investor Relations.

Thank you, operator, and good afternoon, everyone. Welcome to Magnite's fourth quarter 2024 earnings conference call. As a reminder, this conference is being recorded. Joining me on the call today are Michael Barrett, CEO; and David Day, our CFO. I would like to point out that we have posted financial highlight slides on our Investor Relations website to accompany today's presentation. Before we get started, I'll remind you that our prepared remarks and answers to questions will include information that might be considered to be forward-looking statements, including, but not limited to, statements concerning anticipated financial performance and strategic objectives, including the potential impacts of macroeconomic factors on our business. These statements are not guarantees of future performance that reflect our current views with respect to future events and are based on assumptions and estimates and subject to known and unknown risks, uncertainties and other factors that may cause actual results, performance or achievements to be materially different from expectations or results projected or implied by forward-looking statements. A discussion of these and other risks, uncertainties and assumptions is set forth in the company's periodic reports filed with the SEC, including our 2024 Annual Report on Form 10-K. We undertake no obligation to update forward-looking statements. Our commentary will include non-GAAP financial measures, including contribution ex-TAC or less traffic acquisition costs, adjusted EBITDA and non-GAAP income per share. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our earnings press release and in the financial highlights deck that is posted on our Investor Relations website. At times, in response to your questions, we may offer additional metrics to provide greater insight to the dynamics of our business. Please be advised that this additional detail may be one-time in nature, and we may or may not provide an update on the future of these metrics. I encourage you to visit our Investor Relations website to access our press release, financial highlights deck, periodic SEC reports, and the webcast replay of today's call to learn more about Magnite. I will now turn the call over to Michael.

Thank you, Nick. We once again exceeded top line guidance growth in CTV for Q4 and although our Q4 results in DV+ were disappointing, I'm very proud of the terrific year that our team produced. We generated contribution ex-TAC of $607 million and processed ad spend of over $6 billion. We generated adjusted EBITDA of $197 million and $118 million of free cash flow, all record highs for Magnite. These results demonstrate that we've made the right long term investments, are focused on the right part of the market and are making tremendous progress. We are very confident in our future. For Q4 CTV contribution ex-TAC increased 23% year-over-year, outpacing our guidance of 18% to 21% and for full year 2024 CTV contribution ex-TAC grew 19%. In contrast, our DV+ business in Q4 came in later than we expected at 1% contribution ex-TAC growth as a result of some unusual spend patterns post-election. This caused total Q4 results to come in below our range. That said, we're happy to report that DV+ has rebounded nicely to start 2025 and David will walk through more details in his remarks. In contrast to DV+, our CTV business continued to grow with significant momentum. Performance was driven by overall ad spend growth and a stabilizing year-over-year average take rate, illustrating a better product mix. Our most significant growth in Q4 came from Roku, LG, Vizio, Walmart, Disney, Fox, Warner Discovery and Paramount. Netflix continues to ramp and we remain bullish about our Netflix opportunity as they significantly grow their global ad tier and corresponding ad revenue. They will be a key programmatic partner as they expand the rollout of their ad platform in 2025. Live sports had strong growth in the fourth quarter and we expect that to continue. We are encouraged by Disney's focus on live sports and NCAA Football, specifically in Q4 and their pending acquisition of Fubo to become one of the market's largest live TV subscriber bases. We also continue to strengthen our international sports business with the addition of new partners including FIFA and Sky New Zealand. Lastly, we announced a deal with DIRECTV to expand our partnership in their streaming business on top of our efforts in satellite TV. ClearLine, our self-service direct buying platform, posted very strong growth in Q4 and is showing great promise for 2025. The agencies and brands we work with continue to ramp their buying and when buyers use ClearLine, not only do we receive a fee for use of the product, but many of these buyers also leverage our data leading to additional revenue opportunities. ClearLine along with SpringServe, also power our agency marketplaces. These marketplaces, which are leveraged by GroupM and Horizon among others, provide agencies with their own end-to-end private label platform and establish direct connections with sellers, allowing more spend to go to working media. We believe agency marketplaces are a differentiated product offering for Magnite and will contribute nicely to our growth throughout 2025. Now to DV+. As I mentioned earlier, to start 2025, the business has resumed growing at healthy pace in the mid to high single digits. We've seen this rebound broadly across verticals and we're also seeing some benefits from ramping new deals, leading us to believe that the business has normalized after the unusual Q4 trend we experienced. Audio, which is part of our DV+ supply footprint, continues to be a solid growth driver and we see our partners focusing more and more on programmatic ad revenue. We are excited about the growth opportunities of partners such as iHeart and Spotify this year and beyond as this channel grows. Now I'd like to turn to AI where we have some really exciting new initiatives. We have a long history of using machine learning, neural networks and sophisticated data science to optimize our data centers and efficiently process trillions of ad requests per day. These tools reduce our costs and the cost for buyers, allowing them to spend more across Magnite. In addition, these tools power operational efficiencies for clients behind the scenes. In 2025 we will be releasing a number of new client facing tools powered by generative AI. As an example, we recently launched an in beta generative AI feature for our curator product that allows buyers to quickly and easily identify the optimal audiences to meet their marketing goals while improving match rates. Other AI powered tools include in DV+, a yield optimization engine for our demand manager header bidding solution and in CTV, a tool for automating and standardizing content classification signals. We are optimistic that these tools will help drive significant value for our partners and look forward to additional AI driven product releases throughout 2025. These advances will just further solidify us as the leading independent SSP. Before I turn the call over to David, I'd like to address head on comments by The Trade Desk in their earnings call two weeks ago regarding their OpenPath initiative in CTV and the presumption that it will ultimately displace SSPs. Large parts of this argument are flat wrong. But before I get into our points of contention, first the point of agreement. I can't blame them for thinking most SSPs are headed for irrelevance, but it's not because the SSP model is inherently flawed or inefficient. It's because most SSPs tech is undifferentiated and outdated. Most of our competitors in the streaming space have taken display advertising technology and hacked it into working with CTV. Because they have no CTV ad server, they're limited to acting as a reseller of non-unique inventory in an open market capacity. It's a low value practice. Meanwhile, our CTV offerings are purpose built for the space, including our streaming SSP platform and, critically, our SpringServe ad server, which is the industry's leading programmatic mediation layer. Magnite doesn't need to rely on reselling inventory because our tech has earned us direct relationships with every major streaming platform other than YouTube. We can now reach over 90% or 92 million U.S. households and over 90% or 75 million European households in EMEA's big five countries, which is amazing if you think about it. The Trade Desk also argues that OpenPath is more efficient for sellers and avoids the unnecessary fees of middlemen. Yet, The Trade Desk charges sellers a healthy fee to use OpenPath in addition to their fees on the buy side. In reality, the economics are unchanged for publishers. How's that more efficient? Moreover, they claim that each seller using OpenPath would be better served conducting its own yield management. This misses the very obvious fact that most sellers don't have this technology and would be incredibly expensive and inefficient to build in-house. Building their own tech wouldn't be just inefficient, but disadvantageous, as none have access to the reams of pricing data that feed Magnite's yield management systems. As a result, sellers would be far more likely to leave money on the table. This is why sellers connecting to OpenPath continue to find enormous value in using SpringServe. Further, The Trade Desk's argument ignores the importance of diversity of demand. While they are obviously a huge player, no one DSP can represent all the world's demand for every GEO and use case. In addition to established DSPs, we are seeing an explosion of DSPs focused on SMBs and performance advertisers. These essential demand sources aren't accessible through OpenPath, and it's impractical to think sellers will tap into all of them by integrating and managing dozens of direct connections. That is exactly what Magnite is built for. Simply put, OpenPath doesn't replace the need for yield management or remediation layers. So even as OpenPath grows, Magnite's technology will still be used and valued by publishers in the vast majority of OpenPath CTV transactions, but a future dominated by OpenPath is not in the best interest of sellers. The Trade Desk talks a lot about objectivity, but let's be clear, as a DSP, their allegiance lies at the buy side. Despite their marketing effort, OpenPath is not about providing long term value to sellers, it's about extracting more fees from every transaction that runs through their platform. It's about accelerating the commoditization of sellers' inventory and data. We've seen what happens when one player becomes too dominant on both the buy side and the sell side of this industry, and sellers always end up getting the short end of the stick. As a sell-side company, Magnite's mission is to protect the interests of media owners. Like The Trade Desk, we believe in the open Internet, but we also believe that for it to thrive there must be a healthy and fair value exchange between buyers and sellers. That's the future that technology like SpringServe enables and that's the future we'll always be working towards. In closing, we delivered strong Q4 CTV results with overall quarterly results being negatively impacted by DV+'s post-election spending pause. However, our recent trends give us comfort that DV+ headwinds are not structural as we've seen normalization of trends in Q1. The strategic investments we've made to create the world's leading programmatic CTV platform are clearly paying off and we are incredibly excited for the remainder of 2025. With that, I'll turn the call over to David for more detail on the financials.

Thanks Michael. The fourth quarter was unusual for Magnite. We continued with very significant momentum in our CTV business and came in above the high end of top line expectations with 23% year-over-year growth in contribution ex-TAC. As Michael mentioned, however, we experienced an unusual post-election pause in our DV+ business during the fourth quarter. In the run up to the election we saw normal levels of paid impressions and CPMs in our non-political business with CPMs roughly flat year-over-year. Immediately following the election, CPMs dropped significantly, 15% to 20% down through the end of the quarter. As a result, we did not see the normal holiday ramp in our DV+ business resulting in 1% growth for the quarter and the underachievement compared to our expectations. This was radically different from our experience in both the 2020 and 2022 election cycles and this trend was consistent with the weaker than expected results reported by other open Internet industry players. When all was said and done, we saw soft vertical demand in DV+ in 13 of the 18 categories we track. In particular, we witnessed pullbacks in consumer categories, health and fitness, retail, automotive and food and beverage. Other than the benefit from political, the only other bright spot was technology. As Michael mentioned and as I'll highlight in guidance, DV+ has returned to mid to high single digit growth in Q1, which gives us some comfort that lower than expected DV+ results in Q4 were an anomaly. Turning now to full year results. We generated record contribution ex-TAC of $607 million and processed ad spend of over $6 billion. We also generated record adjusted EBITDA of $197 million and $118 million of free cash flow. We ended the year with $483 million in cash and our net leverage ratio has been reduced to 0.4x. Our business is performing well and our capital structure is robust. Total revenue for Q4 was $194 million, up 4% from Q4 2023. Contribution ex-TAC was $180 million, up 9%. CTV contribution ex-TAC was $78 million, up 23% year-over-year and again above the top end of our guidance range. Ad spend growth and strong momentum with our SpringServe ad serving drove the outperformance in the fourth quarter. DV+ contribution ex-TAC was $102 million, an increase of 1% from the fourth quarter last year as we discussed. Our contribution ex-TAC mix for Q4 was 43% CTV, 40% mobile and 17% desktop. From a vertical perspective, as expected and as we discussed last quarter, political was the strongest performing category at approximately 6.5% of contribution ex-TAC. Total operating expenses which includes cost of revenue for the fourth quarter were $154 million, a slight increase from $152 million for the same period last year. Adjusted EBITDA operating expense for the fourth quarter was $104 million within our guidance range. The increase from $95 million last year was primarily driven by personnel and software development costs. I'd like to touch on our tech stack cost initiatives that particularly impact our CTV profitability and scale. First, our tech team continues to make gains in reducing per unit cloud costs which have allowed us to manage significant ad request volume increases with modest cost increases. To this end, our cost per ad request in 2024 has decreased 26% in DV+ and 45% in CTV. Our focus on scale improvement and efficiency is one of our top priorities, and these efforts will continue in 2025. In addition to our per unit cloud cost efforts, we've also been increasing our focus on optimizing our hybrid structure to move additional functions from cloud to on-prem. We expect that these initiatives will lead to increasing margin expansion rates in 2026 and beyond with some potential for benefit in the second half of this year. We will have more to report in coming quarters on these efforts. Net income was $36 million for the quarter compared to net income of $31 million for the fourth quarter of 2023. Adjusted EBITDA grew 9% year-over-year to $77 million, reflecting a margin of 42%, which compares to $70 million last year. As a reminder, we calculated adjusted EBITDA margin as a percentage of contribution ex-TAC. GAAP earnings per diluted share increased 50% and it was $0.24 for the fourth quarter of 2024 compared to $0.16 for the fourth quarter of 2023. Non-GAAP earnings per share in the fourth quarter of 2024 grew 17% and was $0.34 compared to $0.29 last year. Reconciliations to non-GAAP income and non-GAAP earnings per share are included with our Q4 results press release. Operating cash flow, which we define as adjusted EBITDA less CapEx, was $64 million for the quarter. Cash generation was very strong in the fourth quarter and our cash balance at the end of Q4 was $483 million, an increase of $96 million, or 25% from the end of Q3. Capital expenditures, including both purchases of property and equipment and capitalized internally used software development costs were $12 million for the quarter, bringing the total to $52 million for the full year. Our net interest expense for the quarter was $5 million. As mentioned, our net leverage ratio was 0.4x at the end of Q4, a sequential improvement from 0.9x at the end of Q3. We continue to focus on managing shareholder dilution after having successfully lowered our leverage. For the full year 2024, our repurchase program and withhold to cover activity have effectively reduced dilution by 3.2 million shares for $37 million. As a result, the increase in total shares outstanding at the end of 2024 was limited to 2% compared to the end of 2023. We currently have $110 million remaining in our authorized share repurchase program, which we will look to deploy strategically. I'll now share our expectations for the first quarter of 2025 and our current thoughts about the full year. For the first quarter we expect contribution ex-TAC to be in the range of $140 million to $144 million. Contribution ex-TAC attributable to CTV to be in a range of $61 million to $63 million, contribution ex-TAC attributable to DV+ to be in the range of $79 million to $81 million and we anticipate adjusted EBITDA operating expenses to be between $111 million and $113 million, which implies adjusted EBITDA margin of over 20% for Q1 at the midpoints. For the full year we anticipate total contribution ex-TAC to grow above 10% or mid-teens excluding political. Adjusted EBITDA to grow in the mid-teens with adjusted EBITDA margin to expand at least 100 basis points over 2024 and free cash flow to grow in high teens to 20%. In addition, we expect total CapEx to be approximately $60 million supporting our tech stack efficiency efforts as I noted earlier. Overall the fourth quarter and full year 2024 were strong for Magnite despite the unexpected volatility in the post-election DV+ results. We had some very significant wins during the year including Netflix and continue to add cutting edge capabilities to our industry leading CTV tool set. We've also set a high priority on the cost efficiency of our tech stack and have made significant strides this year. The team is performing well and we are excited about what the future holds in 2025 and beyond. And with that, let's open the line for Q&A.

The first question comes from Jason Kreyer with Craig-Hallum. Please go ahead.

Great. Thank you, guys. So obviously ended the year on a strong note in Connected TV, but the guide for Q1 seems to reflect a little bit of a slowdown there. So just wondering if you can get a little context on what you're seeing here in the early stages of '25?

Yes, I'll take that. In DV+ let's split the businesses there. We're actually seeing a nice rebound, especially from Q4, and so as mentioned, we're seeing current growth in the mid to high single-digits and so that's, we think running nicely. And CTV we've typically had Q1 be kind of a low point in our year-over-year growth.

I think it's within expectations.

Okay. Got it. Thank you. I wanted to touch on some of the stuff you talked about — SMBs you kind of highlighted working with some partners to bring greater budget online and allow that cohort of the industry that hasn't participated in CTV in the past to give them access to that. So just wondering, if you can talk about what your role is there, and kind of what you're seeing, and what the timing looks like for those entities to be able to participate?

Yes, great question, Jason. I think we're at relatively early stages there. It's kind of a two part move. First, these newer entrants have to bring on board these performance advertisers that generally are used to the metrics of the DV+ world. The ability to couple the impact of a CTV ad, coupled with advertising on DV+ at the same time and weighing whether or not that's more efficacious, is the start of a journey. But there is a ton of appetite. If you read every Q4 or earlier earnings from any of the big streamers they've all pinned their hope on expanding the addressable advertiser base — they're not just looking to shift dollar-for-dollar from linear to streaming. They need to expand the pie. So they're really leaning into a world of 10,000 advertisers, not 500 advertisers. Our role specifically is to bring that great supply to them. There's no way they can create direct relationships, plug in directly, and access all the supply that we can globally. We're a quintessential SSP in that respect. But as I mentioned earlier, a uniquely advantaged SSP given our programmatic mediation layer with all the top streamers.

All right, thank you.

The next question comes from Dan Kurnos with The Benchmark Company. Please go ahead.

Great, thanks. Good afternoon. Michael, maybe just on how should we think over the medium term about the growth of your CTV ex-TAC business, now that you have all these wins? I mean, obviously it's going to be lumpy with Netflix and timing is hard. So maybe kind of spread that out for us, so we can get sort of a normalized rate. And then, on some of the data initiatives, are we talking utilizing the data you have to improve conversion or drive more business wins, or are you talking about new business lines that can actually add incremental revenue to the growth story? Thanks.

Dan, great question. On CTV growth rate, we've been pretty consistent in saying we have every expectation to outgrow the market. That market has bounced around given the economic backdrop of the last several quarters. We believe CTV can be mid-teens on a full year basis and we fully expect on a full year basis to outpace the market. There will be quarter-to-quarter differences, but we expect to outpace the market. As far as inventory coming online, that's hard to project because some of that is out of our control and relates to our streaming partners' go-to-market strategies. As it specifically relates to Netflix, I feel comfortable with our previous statements: exiting 2025 they will be one of our largest CTV clients. On the data piece, data is a new bucket of revenue opportunity for us. Historically most data surcharges or data sales have occurred at the DSP level on the buy side, and we're seeing a big uptake of our curation tools and buyers utilizing our data storefront. In that instance our publishers and we can participate in those economics.

I was just going to add also, if you think about our guide for 2025 and the political comp, if you strip out political, our guide—conservative, in my view—not knowing the ramp of Netflix and others, would imply closer to 20% CTV growth next year. So it's important to keep that political comp in mind.

Thank you, both. Appreciate it guys.

The next question comes from Laura Martin with Needham & Company. Please go ahead.

Hi, there. My first one is on GenAI. I wanted to drill down there a little more. It sounds like you're introducing a bunch of products in '25 and I'm interested. Do you expect those to bring in new clients, to increase revenue per existing client, to lower client churn? And then related to the cost side of GenAI, you did $56 million of CapEx in 2024. How much do you—I'm very surprised you're going on-prem, because a lot of this GenAI stuff is with a large language model that's housed at Google or Amazon or over at OpenAI. So I'm curious as to bringing on-prem increasing your CapEx on on-prem versus using the cloud. Could you talk a little more about your GenAI strategy in 2025, please?

You bet, Laura. I'll jump on the first part of the question. Generally speaking, the tools that we have launched and that are launching are definitely focused on existing clients to make their life easier, to make it more efficient, and to have them spend more. So the general focus of the first wave of generative tools is to increase revenue with existing clients.

On the CapEx, we had $52 million in CapEx in 2024 and as you highlighted, we do expect about $60 million in 2025. That's actually not related to the GenAI efforts. It's really more related to primarily our CTV business that primarily runs today on AWS. AWS per unit costs, even as they're being optimized for certain applications, can be up to eight times more expensive than bringing those activities on-prem. And so, there's actually a very significant ROI from that incremental CapEx spend and the ability to bring some of that activity on-prem. So the CapEx is much more about our optimization efforts of our tech stack, and giving us greater opportunity to expand our margins. As we swap out more expensive AWS cloud costs in the future to cheaper on-prem costs, we'll see margin improvement. Michael mentioned that the economic packaging of our new tools can push the cost of additional servers for those specific tools onto the end user, so we don't anticipate tens of millions of incremental costs as it relates to cloud usage because of large language models.

Okay. Super helpful. And then my second one is, I understand you feel like you were provoked by The Trade Desk comments, and I appreciated your pushback on why SSPs are still required. My big picture question is, is the open Internet healthy? We have the largest DSP and the largest SSP sort of in a public war now of words about the relative usefulness of their services, but they are part of an ecosystem. I'm just wondering, do you feel that the rise of Amazon's competitive layer is threatening the open Internet and slowing growth, which is creating more friction among the open Internet ecosystem. Do you feel the open Internet is healthy in today's world?

I think it's very healthy, Laura. This isn't a back alley fight. Jeff has a clear view of how he thinks the world will play out and we have the same worldview about the open Internet; we differ on how it gets there and who enjoys the spoils. Amazon can divert dollars, but more and more what we're seeing is Amazon is a buyer of open Internet inventory. They may have sophisticated buying capabilities, but they are intent on being a large general DSP in the world, which requires buying open Internet inventory. We have a tight partnership with them and they're fast growing on our platform. I think they're more of a validation of the open Internet than someone undermining it.

Thank you very much.

The next question comes from Shweta Khajuria with Wolfe Research. Please go ahead.

Thanks a lot for taking my questions. Let me try two, please, on OpenPath Michael, you gave a lot of good color and I think that was much needed. Thank you for that. My question is, how should we think about your comment that publishers end up getting about the same amount of revenue irrespective of whether they go through an SSP like Magnite versus OpenPath? And second, on CPMs, what drove the CPM pressure in the fourth quarter, and what gives you confidence right now? Is it simply a function of CPMs bouncing back and things being back to normal? Any comment there would be helpful. How did you see January versus February? And then third, could you help us unpack your 2025 guidance a little bit, to the degree you can comment on incrementality from any of these new partnerships? How should we think about that? Thanks a lot.

Thanks Shweta. On OpenPath economics: what wasn't broadly understood was that when a publisher engages with OpenPath, there is a fee charged by The Trade Desk. That fee can range in the mid-single-digits depending on the service. So in terms of fees and the tightening of the supply chain, the economics are largely unchanged to the publisher. We've heard this from multiple publishers. So ultimately there's no huge savings here for advertisers or publishers. Regarding CPMs, I'll defer to David who has all the data.

At the highest level there was a drop in demand in November and December. One can speculate as to why — concerns about messy election results or other factors — but the fact was less demand in the fourth quarter running up to the holidays in the DV+ business. In January and February we have seen a significant rebound. That demand has come back and CPMs have improved, which gives us confidence going forward. We're two months into our quarter and have reasonable visibility that the rebound has sustained for a good part of this quarter, which adds to our confidence. On guidance and incrementality from new deals, we've talked in general terms about what the impact of some of our deals could be. We don't control the timing, so we attempt to be somewhat conservative in our guidance. If you look at our current revenue growth rates we have a certain flow-through profile, but incremental revenue on top of that baseline flows through to EBITDA at a much higher rate. So to the extent we have upside, you'll see better proportional impact on EBITDA.

Okay. That's helpful. Thanks, Michael. Thanks, David.

The next question comes from Matt Swanson with RBC Capital Markets. Please go ahead.

Yes, thank you so much for taking my question. Maybe following up on the drop of demand you saw in November and December, it just didn't seem like it really impacted CTV, or did you see a similar dynamic but it just outperformed anyway due to the strength? Just trying to put some guardrails around that.

No, we didn't really see the same impact in CTV. It's a very different world with lots of demand coming in. If there was some impact, it was masked by other growth in demand. We feel like we're a share gainer in that space. Could CTV have grown higher in that quarter? Maybe, but we didn't see a drop in CPMs or similar activity in CTV.

And then maybe one for both of you. The secular growth drivers that you called out for being investments in 2025 — AI curation, ClearLine, agency marketplaces — we spent a lot of time on CTV. How should we think about those in terms of how monetizable they are this year?

Many of these partnerships are strategic and will be longer term to fully ramp. We are dependent on partners in terms of selling them through. Agency marketplaces are a good example: we've been working with GroupM for years. For the first two years it was a concept, now it's a roaring marketplace. Many of these examples—United, Netflix, Mediaocean—take time. I wouldn't anticipate a hockey stick in any particular quarter for these; they gradually grow and add to general revenue growth in 2025.

Thank you.

The next question comes from Shyam Patil with Susquehanna. Please go ahead.

Hi, this is Aaron on for Shyam. Thanks for taking our questions. Maybe for starters, I wanted to double click on DV+ and the volatility that you saw post-election. How did the CPM and demand drop off you saw in Q4 compare to prior election cycles? If it was more severe, what do you think caused that?

It was very different than 2022 and 2020 cycles. In those cycles CPMs did not really drop. This time it was dramatically different. As far as the cause, we know what the output is but can only speculate on the cause — whether advertisers wanted to stay out of the market or expected political turmoil after the election. We saw the drop and now we've seen the recovery, but I don't have definitive insight beyond that.

With political seasons, political spend crowds out consumer ad spend and it gets expensive. It's common to see some larger advertisers pause and then kick in with seasonal advertising in December after the general election. We simply didn't see that this time. The fact that advertising resumed in January and that the CPM cratering happened exactly the day after the election points to the election as the likely cause, though correlation isn't proof of causation.

Got it. And then if I could ask a second question, appreciated the commentary and response to the OpenPath comments from The Trade Desk. Michael, what would you highlight as the top two or three differentiators of Magnite's SSP versus OpenPath?

OpenPath is a valuable source of demand and the demand is core for publishers, but it's just one DSP. The Trade Desk suggested publishers should build their own yield management and we strongly disagree. Publishers generally lack the capability and the data to build yield management at the scale we operate. Our biggest values above OpenPath are that we bring the world's demand — every DSP, global demand — and we add yield management on top of that. The Trade Desk isn't even operating in Latin America while we are. For the same or similar fees, publishers get The Trade Desk plus the world's demand and yield optimization with Magnite, which is not what OpenPath alone offers.

That's helpful. Thank you both.

The next question comes from Eric Martinuzzi with Lake Street. Please go ahead.

Yes, just one question/clarification. You mentioned political at 6.5% of contribution ex-TAC in Q4, what was it for the year?

For the year, it was 3.2%.

The next question comes from Zach Cummins with B. Riley FBR. Please go ahead.

Hi, good afternoon. Thanks for taking my questions. I wanted to ask about the CTV business. Have you seen any material mix shift within that business? It seems relatively stable throughout 2024. How are you thinking about that mix moving forward in 2025 and beyond?

That's a pretty accurate statement. Our mix of revenue by publisher types and products has been quite stable. In 2025 on the margin we will be bringing more demand as Magnite, rather than publisher-sold programmatic. We expect expansion of media plans to bring on publishers that rely more on us for demand. Over time we anticipate the mix changing, but through 2025 the major premium streamers will still be the center of media plans and those publishers often prefer to sell programmatic deals direct.

Understood. And just one quick follow-up, curious about your recent press article regarding a partnership with ex. Can you talk about your role in that partnership and how to think about potential contribution over time?

That was a press article and we're somewhat limited in what we can describe. You can anticipate further information downstream, but that article was premature.

Understood. Well, thanks for taking my questions, and best of luck with the rest of the quarter.

Thanks, Zach.

Thank you.

The next question comes from Omar Dessouky with Bank of America. Please go ahead.

Hi guys, this is Arthur Ong from Omar. Thanks for taking the question. David, I wanted to touch on your comment about lowering your cost per ad request by optimizing the number of ad requests you send to DSPs. How should we think about seeding that efficiency improvement? If you have a model predicting how DSPs will respond to different ad requests, the model should keep getting better as you ingest more data. Where do you see the seeding of that efficiency gain and where are you on that curve?

There are a couple elements to those efficiency gains. There's a core cost of processing an ad request and a core cost of sending that ad request out to DSPs. There's also the 'brains' behind how efficiently you filter what's valuable and where it should be sent. We've been working on that for a long time and it's been part of our success. There are step function gains in per unit cost by optimizing cloud usage. We're relatively new to putting our CTV businesses together and optimizing in the cloud, so there's upside there. The biggest step function is that certain activities can be run much more efficiently on-prem and we are in the phase now of sorting that out and optimizing it. I think that's where the biggest gains will come over time. We'll have more to discuss in coming quarters as we progress.

We intend to continue improving unit costs, but correspondingly we're doubling or tripling volume. You'll see a steady layer cost but we'll be able to do much more, and our lifeblood is auctions in volume. You shouldn't expect costs to go to zero because we'll keep taking on more volume at an affordable cost.

Got it. Thank you very much guys. Appreciate it.

Great. Thanks, Arthur.

This concludes our question-and-answer session. I would like to turn the conference back over to Michael for any closing remarks.

Thank you. Thank you all for joining us and your support. We have a very exciting 2025 ahead of us that I feel very good about. The Magnite team is performing at a very high level and I love our resilience, even when the market throws us challenges like in Q4, and our ability to bounce back quickly. We look forward to speaking with many of you at our upcoming investor events. We are participating in the Susquehanna Conference tomorrow; Benchmark meetings in Boston on March 6; Craig-Hallum meetings in Chicago and Minneapolis on March 12 and 13; Bank of America meetings in New York on March 19; RBC meetings in Kansas City and Denver on March 25 and 26; and B. Riley meetings in San Francisco on March 31. Thank you all for joining, and have a great evening.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.