AdTheorent Holding Company, Inc. (ADTH) on Q3 2022 Results - Earnings Call Transcript
Operator: Ladies and gentlemen, thank you for standing by, and welcome to AdTheorent's Third Quarter 2022 Earnings Call. At this time, all participants are in a listen-only mode. After the speakers' presentation, there will be a question-and-answer session. Please be advised that this conference is being recorded. At this time, I'd like to turn the conference call over to your first speaker, David DiStefano, Investor Relations. David, please go ahead.
David DiStefano: Good afternoon, and welcome to AdTheorent's third quarter 2022 earnings call. We will be discussing our results announced in our press release issued after the market closed today. With me today are AdTheorent's Chief Executive Officer, Jim Lawson; and Chief Financial Officer, Chuck Jordan. Before we begin, I'd like to remind you that today's conference call will include forward-looking statements based on the company's current expectations. These forward-looking statements are subject to a number of significant risks and uncertainties, and our actual results may differ materially. For a discussion of factors that could affect our future financial results and business, please refer to the disclosure in today's earnings release and our other reports and filings with the Securities and Exchange Commission. All of today's statements are made based upon information available to us today, and we assume no obligation to update any such statements except as required by law. We will refer to both GAAP and non-GAAP financial measures during the call. You can find a reconciliation of our GAAP to non-GAAP measures included in our press release posted to the Investor Relations website at www.edp.com. All of our non-revenue financial measures we discuss today are non-GAAP, unless we state otherwise. With that, let me turn the call over to Jim.
James Lawson: Thank you, David, and good afternoon, everyone. Thank you for joining our third quarter 2022 earnings call. During today's call, I will discuss our high-level results and business highlights and then turn the call over to Chuck, who will provide a more detailed look at our third quarter results and provide guidance for the fourth quarter and full year 2022. I am pleased we were able to generate financial results in line with our guidance and grow active customers 11% year-over-year despite macro-driven pullbacks in ad spend. In the third quarter, we generated $37.6 million in revenue, adjusted gross profit of $24.7 million and adjusted EBITDA of $3.6 million or 14.5% of adjusted gross profit. Performance continues to be led by strong demand across Connected TV or CTV and the healthcare vertical. In particular, we are pleased to see momentum across products and capabilities recently introduced into the market, which were the subject of investment and development in 2022. As we discussed on the Q2 call, the low end of our forward guidance contemplated continuing weakness in the macroeconomic environment and further pressure on ad budgets. Unfortunately, we did see this in Q3, leading to revenue results toward the lower end of our guidance range. We are pleased to note that our performance began to inflect at the end of the quarter and that we are seeing increased demand for our offerings, including record pipeline generation in September. Despite these positive signals, we believe it is prudent to narrow our 2022 guidance towards the lower end of our prior range, factoring in the macro uncertainty as well as more conservative assumptions about the holiday advertising season and the amount of incremental or unforecasted ad budgets that we expect to receive in Q4. Our 2023 planning assumes no material macro rebound, but our strong balance sheet and cash flow characteristics will enable continued focused investment and position us to capture greater wallet share when ad spend normalizes. In prior quarters, we have highlighted investments in data, technology and new products that drive valuable differentiation versus other DSPs and provide meaningful growth opportunities for our business. During the third quarter, we saw positive customer responses to several of these strategic investments, including the mid-October launch of our highly differentiated predictive audience suite of targeting products, the first of which is a highly customized offering for health advertisers. Equally encouraging, our ongoing CTV-focused product development continues to drive consistent quarter-over-quarter growth and pipeline generation, which we see as just the beginning. Finally, we are successfully scaling our direct access business as we continue to develop and deploy enhanced capabilities, including robust verticalized ad solutions. This self-service platform revenue is net new to AdTheorent, opens up immense opportunity in customer types and sets the foundation for developing longer-term strategic relationships with organizations controlling large and sustained media budgets. Let me speak to each of these and the results they are driving. First, I'd like to discuss AdTheorent predictive audiences, AdTheorent Health and our continued advancements towards platform verticalization. We have previously discussed the AdTheorent Health product initiative and our focused efforts to build a dedicated health product and go-to-market infrastructure, heightened scrutiny around the use of personal health information. Combined with our ability to use privacy forward ad targeting methods to drive healthcare advertising outcomes more efficiently than other platforms, positions us to capture a sizable portion of the nearly $16 billion in annual healthcare advertising spending. In Q3, our platform's deep health capabilities helped us win sizable deals with some of the U.S. and Canada's leading digital healthcare agencies and brands, some moving their full programmatic media spend to AdTheorent. We expect even greater momentum in the healthcare vertical going forward, given our mid-October launch of AdTheorent Health predictive audiences. We built this product in consultation with leading healthcare advertisers based on their stated interests and needs and the early feedback is very encouraging. This first-of-its-kind product allows programmatic advertisers to target audiences in a more precise data-driven and less opaque manner than previously possible. AdTheorent Health predictive audiences are fully customized based on advertiser requirements. They use primary sourced healthcare data to identify using machine learning and statistics, the most qualified audiences for targeting consideration without relying on personalized information or user ideas. This de-identified data is aggregated from hospital and healthcare provider claims, electronic health records and pharmacy and contextual data representing more than 300 million patients and healthcare professionals, which equates to more than 90% of the U.S. population. This aggregated health data is mined to create statistical representations of desired target audiences. For example, condition sufferers, caregivers or prescription users based on non-sensitive data such as age, location or gender. Using our platform, an advertiser could create a statistical representation of a target audience comprised of for example, people who have been diagnosed with high cholesterol within the last 12 months, who are prescribed a given drug and who had the prescription filled at a specific pharmacy location. From there, step 2 of the process, activating these audiences as part of a customer campaign on our platform focuses on executing AdTheorent unique KPI specific impression scoring within the created audience. We believe our launch of AdTheorent Health predictive audiences provides a blueprint for delivering unique value through our platform across a variety of growth verticals. As we have discussed in the past, in addition to AdTheorent's core machine learning powered platform, an ID independent impression scoring capabilities, we offer customized vertical solutions to address the needs of advertisers in specialized industries. We continue to see the strongest growth from sectors where we have launched these vertical-specific solutions such as health. Importantly, we are also working actively to verticalize the platform's tools, resources and workflows for self-service or direct access users with health being an immediate focus. The continued verticalization of our solutions and our platform will allow us to compete against generalist DSPs with greater value-added differentiation, and we believe it will drive more rapid adoption of our platform. And AdTheorent predictive audiences are foundational to this as our audience builder tool and the platform enables users to create custom ML-powered predictive audiences, which are focused and unique to specific verticals. Next, I'd like to briefly highlight our progress with CTV, our commitment to innovation helped drive 36% year-over-year CTV revenue growth during the third quarter, a remarkable achievement given the meaningful compression in advertising budgets during the period. CTV was nearly 10% of revenues during the third quarter of 2022, up from less than 3% in 2020 and nearly 7% in 2021. Our use of machine learning models to drive post CTV ad exposure business outcomes for clients is truly cutting-edge and differentiating. We believe we have a sustainable competitive advantage in this channel, and we are already seeing a return on the 2021 and 2022 investments that we made to deliver a more performance-focused CTV offering. Marketers are continuing to shift significant dollars from linear to programmatic CTV due to increased precision of targeting and the ability to report measurable ROI, both of which are core to AdTheorent's CTV offering. Feedback from the market remains enthusiastic, and customers are continuing to choose AdTheorent CTV solution for 5 key reasons. First performance, we build machine learning models for CTV campaigns using the wealth of data we have obtained in order to drive towards business outcomes instead of simply delivering video views or completes. One of the ways we have achieved this is by committing development and other resources to maximizing and normalizing data attributes available for programmatic CTV. We have discussed this in prior calls, but the more data our machine learning models have access to, the better ROI we are able to drive for our customers. Second attribution, we have built our own attribution solution that allows us to tie viewership on CTV devices to outcomes on mobile and desktop devices. This data feeds into custom machine learning models on CTV and also allows us to report back valuable insights to our customers. Third privacy, we use data science and machine learning for ad impression analysis and targeting, not to build user profiles or ID focused targeting segments, relying on sensitive or individualized personal data. Fourth omnichannel activation, we can run CTV as part of an omnichannel campaign and CTV can be used either as an upper funnel or lower funnel driver for our customers. In a prior call, we discussed AdTheorent's full funnel approach and how we drive greater lower funnel campaign performance, for example, sales by leveraging data about consumer engagement with upper funnel campaign tactics. And finally creative, we have a talented in-house design team that allows us to deploy innovative creative executions on CTV. For example, we have the ability to run CTV ads with QR codes, allowing consumers to interact with CTV units in a more meaningful way. We are also able to run dynamic CTV ads that provide a more personalized experience based on a user's location, time or weather signals. Innovation will continue to drive our success in this channel, and we are already seeing strong uplift from innovations deployed in the first half of '22. In particular, we are seeing strong demand for our predictive audiences across the CTV channel. Predictive audiences are a natural fit for CTV because customers are moving from linear to CTV for more precise targeting. And our predictive audiences drive both traditional audience targeting benefits such as audience quality or reach validation as well as machine learning-driven campaign outcomes, such as online sales. During the third quarter, we also expanded our reporting capabilities for CTV and developed custom analysis to demonstrate how CTV can empower brands to reach both known and potentially high-value users. For example, we partnered with a luxury hotel brand and through a custom website analysis, identified more than 160,000 households associated with devices that visited a website but did not complete a booking. In partnership with this luxury hotel brand, AdTheorent reengaged these households with CTV media as an upper funnel tactic, building brand loyalty among a high-qualified audience. For this brand, as part of a true full funnel strategy, our campaign performance was 24% better than the cost per booking benchmark. Our CTV offering gets better every quarter as we operationalize and automate new and innovative machine learning advancements, and this will drive our continued success. Looking forward with sustained investment in the channel, we expect our outsized growth in CTV to continue. Moving to direct access, direct access is our margin-accretive self-service business offering. As discussed previously, we've invested meaningfully here. And while we will not see a significant revenue benefit until 2023, it is already attracting net new customers, and we are excited to begin scaling this meaningful growth initiative. Third quarter progress in market was notable. 9 new brand or agency customers began actively setting up and working towards launch of their first campaign. 21 new brand or agency customers had campaigns actively running through the platform, most of them sophisticated global advertisers, and digital ad impressions served for direct access clients rose 72% quarter-over-quarter. We are also in late-stage discussions with a number of large organizations about potential 2023 engagements. Importantly, direct access is complementary to our managed programmatic offering. It expands our addressable market to include those brand and agency customers who wish to transact in a self-service capacity, allowing us to now transact with all advertisers regardless of size, internal resources or campaign complexity. Direct access addresses the needs of customers with media trading expertise and execution resources, whereas the managed programmatic model is well suited for customers, focusing on complex KPIs who desire additional support and value-added benefits, including strategy, creative and campaign optimization and execution. We are also very pleased with the market validation we are seeing. Customers appreciate our industry-leading price transparency, our platform optimizers and tools, which drive performance and efficiency and the intuitiveness of our enhanced workflows and UI, which operationalize complex ML capabilities. We are having business development discussions with bigger players, and we have a value-adding case for inclusion in their programmatic tech stacks. We are encouraged that sophisticated agencies and brands are telling us that our platform and capabilities fill needed gaps and provide incremental value as compared to DSPs and other solutions they are using now. We are making progress proving our value explicitly through platform evaluations, and we are highly confident that customers can leverage our platform to meaningfully improve return on ad spend. I also want to talk briefly about 1 new capability introduced during the third quarter and provide a quick update on another. First, we introduced search and social capabilities, expanding into YouTube, Google Search and Meta platforms to accommodate our customers' desire to activate with AdTheorent across the entire digital ecosystem. AdTheorent's data-driven approach to advertising is well suited for this expansion because we are able to leverage our data and machine learning to connect the dots between search and social investments and programmatic media, inform our programmatic media buying with search and social data and capture a larger percentage of customers' media budgets. Also, I'd like to give a quick update on our previously discussed work to leverage natural language processing and keyword analysis as part of our impression scoring system. During the third quarter, we scaled our NLP framework to scrap and extract keyword attributes from 500,000 sites daily. Additionally, these contextual signals, which are not dependent on user IDs have been incorporated into all of our CPA predictive models. Since introducing these attributes, 92% of our CPA models as part of the automated feature selection processes have adopted keyword signals as a model feature impacting model performance. And in A/B testing, we have seen significant increases in KPI performance by models, leveraging keyword attributes. Our industry-leading campaign KPI performance does not happen by accident. It is the result of hard work and the expertise of our team in building and using technology to drive positive outcomes. In our next phase of this work, we will incorporate sentiment analysis into our NLP framework. Switching to the macro backdrop and looking ahead. Many of the dynamics we discussed last quarter remain unchanged, including reduced ad budgets, campaign deferments and diminished visibility. But as I noted at the outset of the call, although we are closing out 2022 in a challenging spending environment, we approached the end of the year with optimism and high confidence that we will perform very well in categories that we can control. Advertiser sentiment appeared to bottom out in late August and improved incrementally throughout the balance of the quarter. September showed the strongest demand in the quarter, and these slightly more robust trends have continued in the fourth quarter to-date. In light of the ongoing uncertainty and economic cross winds impacting our industry and business, we took measures to further reduce operating expenses during the quarter. These actions will save us over $3 million annually, starting in the fourth quarter. As the macro headwinds subside, the steps we are taking both in terms of investments in innovation and operating with a leaner cost structure will enable us to fulfill our commitment to generating strong top line growth and returning to historical levels of profitability. We remain convinced that especially in periods of economic volatility, the ability to drive industry-leading performance and deliver best-in-class return on ad spend, all using ID agnostic privacy-forward methods that customers crave will ultimately prove the value of AdTheorent's unique product portfolio. We have no debt. We operate profitably. We have a high-performing platform, and we are bringing to market platform capabilities that customers want and need. Today, I discussed a few of the advantages we bring to market, which we believe will fuel adoption and growth. Our platform is our product, and there are now many new ways for customers to receive the value we deliver. Timing is everything. We used Q3 to complete and deploy critical enhancements to our direct access line of business and expand and refine our vertical offerings, including launching AdTheorent Health, predictive audiences and establishing a foundation to deliver predictive audiences across a range of growth verticals. We are confident that these focus areas will be a big part of our success in 2023, and our excellent team is locked in on getting that done. With that, I'll turn it over to Chuck.
Chuck Jordan: Thank you, Jim. Thanks again to everyone for joining us today. Before discussing detailed financial results, I'd like to point out that in addition to our GAAP results, I'll be discussing certain non-GAAP results. Our GAAP financial results, along with the reconciliation between GAAP and non-GAAP results can be found in our earnings release that is posted on our website at www.adtheorent.com. As Jim covered in his remarks, the impact of softening macroeconomic conditions and the deteriorating visibility in advertising budgets that arose in the second quarter persisted in Q3. While these headwinds did influence buying decisions and weighed on our third quarter performance, we were able to deliver results in line with previously established guidance ranges. Now I'll walk you through our third quarter financial performance and then discuss our guidance for the full year of 2022. Total revenue in the third quarter was $37.6 million, a decrease of $2 million or 4.9% as compared to the third quarter of 2021. We experienced some softening in demand across our BFSI, government education and nonprofit, consumer packaged goods and industry and agriculture verticals. These verticals were down collectively $5.6 million or 31%. Partially offsetting these decreases was continued strength in AdTheorent Health and the real estate and services verticals, which were, in the aggregate, up $3.5 million or nearly 40%. We remain pleased with the momentum in our CTV offering. CTV revenue grew 36% during the quarter to $3.6 million as compared to $2.7 million in the third quarter of 2021, fueled by new customer growth. In discussing the remainder of the income statement, we make references to certain non-GAAP measures. You can find information on the most directly comparable GAAP metrics in our third quarter earnings press release. Our overall campaign profitability remains strong. Adjusted gross profit, which is a non-GAAP metric that removes traffic acquisition-related platform operations costs was $24.7 million in the third quarter or 65.8% of revenue compared to 64.6% in the third quarter of 2021. Moving down the income statement to operating expenses, third quarter operating expenses were $39.4 million, an increase of $5 million or versus the third quarter of 2021. Stock compensation, headcount costs and insurance were up year-over-year. Partially offsetting these year-over-year increases were decreases in TAC and legal and professional costs. Platform operation expenses for Q3 were up 1.9% due to hiring driven increases in our media operations, data science and analytics and technology teams, an increase in equity-based compensation and continued investment in our data infrastructure to support our predictive audience initiatives. The decrease in revenue driven traffic acquisition costs offset these increases. Sales and marketing expenses for Q3 were up approximately $1.9 million or 20.8%, largely driven by a $1 million increase in employee expenses related to sales and customer support hiring and increases in equity-based compensation and travel-related expenses as our customer-facing teams resume more traditional business travel routines. Technology and development expenses for Q3 were up approximately $1 million or 35.8% primarily due to incremental software expense and employee-related cost increases as we continue to invest in our technology and product development capabilities. General and administrative expenses for Q3 were up approximately $1.7 million or 53.9% primarily due to an increase in equity-based compensation, an increase in insurance expense driven by public company directors and officers, insurance premiums and employee expenses related to hiring for the general and administrative teams. Professional services expenses were down year-over-year as the prior year included costs related to public company readiness, including elevated legal and consulting costs. Moving to earnings, we exceeded the midpoint of our Q3 guidance range with our third quarter adjusted EBITDA coming in at $3.6 million versus $8.9 million for Q3 in 2021. Adjusted EBITDA margin for the quarter was 14.5% versus 35% for Q3 of 2021. Now let's turn to our outlook for the full year 2022. For the full year and consistent with previously provided guidance, we expect revenue to be between $160 million and $165 million, with the range representing a decrease of 3.2% to essentially flat compared to 2021. For the full year, and consistent with previous guidance, we expect adjusted gross profit to be between $105.9 million and $109.1 million, representing a decrease ranging from 3.2% to nearly flat compared to 2021. We expect full year 2022 adjusted EBITDA to be between $17.5 million and $20 million or approximately 16.5% to 18.3% of adjusted gross profit. As Jim mentioned, we are seeing more optimistic trends in the demand for our offerings. We continue to invest in our high-performing platform, and we are excited about the market reception to our new products. Operationally, we are debt free, and we have taken measures to ensure we are properly positioned for the current economic environment while continuing to invest in our market opportunity. We're taking a cautious view of 2023 as we work through our 2023 planning process, but we believe we are well positioned to gain market share and grow EBITDA, especially once ad spend returns to historical levels. Now I'd like to turn it over to the operator to moderate our Q&A session. Question-and-Answer Session
Operator: And our first question today comes from Maria Ripps from Canaccord.
Maria Ripps: I appreciate all the color on the quarter. So first, with revenue trends similarly sort of decelerating into Q4, can you maybe just talk about how Q3 progressed? You mentioned things may have improved by the end of the quarter and that sort of continued into Q4. I guess how did October perform relative to Q3? And maybe talk about this trend sort of in the context of expected larger year-over-year declines here in Q4?
James Lawson: Q3 was not where we wanted it to be. Our revenue was at the low end of our range. That was for a couple of reasons. We were impacted certainly by the macro, a temporary pullback from brands on ad spending, some delays with new platform integration due to inflation or for financial results. Large holding company growth in the third quarter was about 0% to 1%, which shows a bit of the slowdown in the aggregate ad spend. And we're a relatively small organization and a handful of pauses or pushbacks in our world are much more visible in our short-term results. And I also think that the current environment tends to favor the status quo tech stack. It takes a bit longer to break through to the bigger flagship accounts. Some pilots and strategic deals take a little bit longer to win. So I mean I think there's -- we saw a bit of that, but at the end of the day, that's fine. I mean this was a quarter where we made a ton of progress. I think the macro conditions and some of the factors that were headwinds for us this quarter, we believe our short-term. We think that in September, we witnessed some of the most robust and strong demand for our offerings that we've seen as an organization in large part fueled by some of our recent deployments, 1 of them being AdTheorent predictive audience specifically AdTheorent Health, predictive audiences. A number of the growth in pipeline is attributable to those recent deployments. So we're very often -- the work that, we're putting in the investments that we're making are going to impact the numbers in future quarters. Obviously, not the quarter we wanted to see this time around. We signaled that it was going to be a range that was going to be less clear than we've seen in the past. There's a lot of uncertainty. But at the end of the day, I think we feel very good headed into the fourth quarter. Still don't have full transparency into holiday spending. I don't think anyone does. And some of the future macro conditions still remain a little bit uncertain. But we are seeing more positive signals. We feel good about where we're going. But again, we want to be fully transparent about the challenges that we're seeing and that we have seen.
Maria Ripps: Got it, that's very helpful. And then secondly, how does an advertiser planning cycle look like now? Sort of anything you can share with us around visibility into next year's budgets at this point? And I know you're not guiding to next year, but how are you thinking about your ability to outperform the broader digital ad market next year, assuming that it grows somewhere in the sort of mid-single-digit range?
James Lawson: I'll answer the -- second question first. I think that we believe that the opportunity for us to grow remains strong. The products that we've rolled into market and frankly, a number of the main focus areas and one that I'll talk about briefly here is our direct access offering. We believe that the opportunity for our direct access offering cannot be overstated. It's something that we've been working on with great dedication and focused investment. In our view, the programmatic market is long overdue for a transformational tech recalibration, how ads are targeted to consumers. Currently, it's done on an ID user basis, focus to user IDs, retarget user IDs. We believe there's an incredible opportunity to introduce a solution that changes that. That is what we bring to the table. We think that it's resonating very well with the customers and the partners that we're talking to. -- to have that reflected in our results, will take a little bit of time in this environment, but we feel very good about where we're going. We believe that many organizations are looking for media activation capabilities that can sit on top of data, many different sources of data, some of the data that they have, some source or C data that we license. Make sense of it and then activate it to purchase digital media intelligently without relying on user IDs. And using our ML-powered media buying platform, we do this very, very well. And we're excited to be having discussions with very big customers about those capabilities. We believe that we can return to a strong growth trajectory next year. Our historic goal has always been 20%. We don't think next year it's going to be 20%, just in the interest of being prudent and careful and cautious given a lot of the unknown factors. But we're going to grow next year. We are going to realize a return on investment from the very, very successful deployments that we've made with direct access, AdTheorent predictive health, audiences, AdTheorent predictive audiences generally and CTV. We think those areas will drive growth. We do not believe that these last 2 quarters are indicative of AdTheorent and what the future holds for us. Our head has been down on driving value creation, differentiation and frankly, products that our customers want and things that will make us stand out. We're having -- we have the luxury of having very good sophisticated customers that we've worked with for 10 years. And they have communicated to us that we're performing at a very high level on their campaigns. This quarter, we grew the number of active customers that we have, the ad spend was down. The economy was bad and that's okay. We're -- we have a balance sheet, we have cash in the bank, and we have a, profitable business that, allow us the luxury of being patient. We have a very big and vast opportunity in front of us. And we're not going to let a quarter or 2 of macro challenges impact our enthusiasm or change us from our course. We believe that a continued heads-down approach, delivering on these major opportunities is the job. And we've rolled out in the third quarter, 2 of our biggest products to-date as a company, and we're very excited to see some of the progress that come from those deployments.
Operator: Our next question comes from John Blackledge from Cowen.
James Kopelman: This is James Kopelman on for John. There's obviously a ton of potential with direct access initiative. You just spoke to a little bit, I'd like to come back. We recognize it's early innings as you begin to scale. I'm curious what is some of the feedback that you're starting to hear from new customers so far? And looking out over, call it, 12 to 18 months. I understand the macro is complicated, but how quickly should we think about the scaling of the initiative? And then, again, looking out over some period of time, call it a year or 2, I'm curious if you could help us on -- in terms of verticals, how much you might expect to come from health versus other key verticals?
James Lawson: On the last question first about health. Health is our largest vertical now, and we expect it to be our largest vertical for the foreseeable future. And that is why we rolled out the AdTheorent Health, predictive audience's product first. We believe that in a data environment and a privacy environment where there are very large budgets and very large organizations seeking to connect with consumers about healthcare products and pharmaceutical products. That there was a gap in the market and a gap in the capability set that we believe we have filled by accessing and licensing, key data sets that we can use health claims data and data about from pharmacies and other just first and source data that we can use to create a very valuable foundational data targeting layers for our machine learning programmatic platform so that we don't have to target users based on inferences about their health condition. That has been a standard fair method of delivering targeted ads in healthcare, a lot of the very sophisticated and thoughtful programmatic healthcare advertisers are looking to find a better way to reach consumers who are interested in various healthcare products. Our offering was designed and developed in close consultation with some of the most sophisticated and large global pharmaceutical advertisers in the world. We talk to them about what was in the market. We talked to them about what we could achieve with our machine learning capabilities. We licensed data. We made that investment, and now we're seeing the beginnings of the return. We're seeing pipeline that is record-breaking for our company, and we're seeing very good test opportunities with household-name pharmaceutical companies even in Q4. So I don't want to get ahead of ourselves here and talk about what financial results we're going to see from healthcare in the fourth quarter or this year, but the response is very, very positive. And I think it's because the customers are smart, and they recognize that making inferences about healthcare condition based on who visits what website is not the future for programmatic digital advertising in healthcare. Your first question, I think, was about direct access. Direct access, we started out our company on the foundation of using machine learning and data science to target ads by predictably scoring ad impressions. And the challenge with doing that is that it's new and it's more disruptive, and it's harder to do. And in the beginning, we offered our service as a programmatic managed programmatic offering or a managed service version. And as we've gone forward in the market, we've realized that we needed to make some changes to our UI and our workflows so that these very complicated executions could be achieved on a self-service basis by brand to have in-house teams or by agencies who are looking to use a really, really powerful performance DSP. That is what we have built. That is what we've been focused on doing. We could have come to market 3 years ago, 4 years ago with another generic DSP that targets IDs. That is not what we are in the business of doing. So the work that we've put in to building our DSP, to building it on a -- to building our capabilities in terms of targeting ads sets us apart. Our direct access now that it's in market, we're talking with very, very large media buying companies. The results have not hit our financials yet, but there are going to be some very good opportunities that we're going to have. We already have trials that we're very happy about and excited about. So I'll just leave it at that. We think that direct access is a very, very big part of our future, and we will see results in 2023 from that effort.
James Kopelman: Great, that's all very helpful. And I appreciate the color and best of luck with the remainder of the fourth quarter.
Operator: Our next question comes from Laura Martin from Needham.
Laura Martin: So I'm going to push you a little Jim, as you can only expect from me. So our competitive advantage here is algorithmic and performance. And now you're talking about CTV, which is cool, except CTV tends to be top of funnel. It tends to be always registered environment, which feels to me like it's not the best use of your predictive scoring. So please refute to me?
James Lawson: The beauty of what we do is that we try to disrupt some of those premises that we think can be challenged. And I think that is an example of 1. CTV need not be a purely awareness upper our funnel tactic. AdTheorent has the ability to generate machine learning insights from post CTV ad exposure, leverage those insights to drive optimizations to make CTV a more performance channel. One of our major investments that we started making in the third and fourth quarter of last year was to normalize and enhance and improve the data received from the programmatic exchanges related to CTV inventory, the content object metadata. A lot of DSPs, a lot of programmatic companies do not use that data because it's messy and it's not normalized. And it's frankly very hard to understand what it is. We invested and we have now completed that investment and we are using, the data that comes through the bid request through SSPs to make our CTV more data-driven. And obviously, the ability to drive focused performance-based CTV is different, and that is exactly what we are in the market to do. We have a full funnel approach, where we can leverage data from the upper funnel to inform our models and our algorithms in the lower funnel, but we also believe that at every level of the funnel, there's also a performance aspect. Even in the upper funnel, there is a measurable winner and a loser in terms of who can drive clicks or engagement. With CTV, we're making that screen and the ability to advertise on CTV, more data-driven and more accountable. And I think that is what's resonating. We saw 36% growth this quarter. It's a little bit down from last quarter of over 100%. But that was just a factor of some of the larger macro trends that we were seeing. The work that we're doing on CTV, I think, is going to set us apart, it already is, and we're very excited about it for the future.
Laura Martin: Okay, okay. And then my second one is sort of an industry design question. So we had Trade Desk report revenue of 31% and a lot of the smaller DSPs like yourself, reported negative revenue growth. So the question is, is sort of the large getting larger. Are these winner-take-all markets in the digital advertising space?
James Lawson: Yes, I mean the Trade Desk got a solid quarter. They have a great company. A number of factors impact companies differently. Their size -- they're almost 10x our size. They've had time to implement strategic partnerships that we're in the process of implementing. This is our first year as a public company. We - became public at a quite an interesting time. There's some consolidation going on. But at the end of the day, the way we defend ourselves against that in the near term is just making sure that when we get trials and frankly, when we go into the market, there is a need for what we do. There is a demand for a DSP that comes to the table with a more data-driven less ID-focused way to drive business outcomes. That is the lower funnel performance DSP opportunity that we are pursuing. The more we verticalize that, the more it resonates with the customers that we're talking to. I think there are big ESPs and other programmatic companies that don't have that same focus. So we have a different opportunity, and we believe we're pursuing it in the right way. We're not going to panic because Q2 and Q3 had a demand environment that wasn't ideal for media buying. But at the same time, yes, the Trade Desk had a great quarter, great. Good for them. We're really happy for them. But we had a great quarter, too. It wasn't shown in the financial results. And at the end of the day, we realize that, that is all that matters for this purpose, but we also have an obligation and a duty to invest in the opportunity and the transformation of programmatic to an ID agnostic world, where you can predictably score impressions and not be not be required to identify IDs and go in that direction. We believe it is an enormous opportunity. And we're the only company doing it, we're the only DSP doing it. It's an efficient -- it's also and efficiency. When you can create audiences using statistics and analysis of data and you can create audiences that are more effective and more valuable than the third-party audiences that are being licensed by every other DSP in the market right now, that's valuable. It's also valuable to be able to go to very large brands that are trying to bring programmatic capabilities in-house. And they want a predictive partner that can help them understand their own data, draw conclusions from that data and then use it to execute programmatic media buys. That is the business that we are in. We think it's very exciting, a little bit different. It's going to take a little bit longer for us to get the scale that some of our peers and market already have with 10x our size and resources. But we're very optimistic about it. We have a better business now than we had a year ago by far. And we're going to just keep doing what we do until the market improves, and we'll have results that are good, and then we'll have results that are great when the market improves.
Operator: And our next question comes from Andrew Boone from JMP Securities.
Unidentified Analyst: Matt on for Andrew, just with an obviously a weakening digital advertising environment, can you give us any color on how we should think about take rates going forward? And then maybe a second one, just on the advertisers that were added in the quarter, are those health advertisers, obviously, predictive audiences being added in October. Just any color there on the strength from the client heads in the quarter?
James Lawson: The third quarter did not reflect any incremental new predictive health customer acquisitions. We believe that those are going to come beginning at a smaller scale in the fourth quarter, but mostly next year. Most of the pipeline creation from this product is focused on 2023. So no, we have not seen any benefit from that work to-date. With regard to take rate, we look at things from a very, very different perspective. We think that when we work with customers, the first question we asked them is, what is your cost per action goal for your campaign. If you're trying to sell something and you have a media budget, what is the cost per action that you want to spend? How do you define success? And then our algorithms are programmed to achieve that outcome. So for us, the take rate is more about can we deliver the campaign performance that the customer came to us and requested. And then from there, if we can do that very efficiently by for example, removing fraud, not wasting our money on impressions that aren't viewable, not wasting our money on impressions that are not accountable by third-party measurement partners and all those other wasteful things. If we can do it really efficiently and we know where to get the right conversions, and we can get conversions without wasting as much money on impressions that are not going to yield conversions, then we can be profitable. And our adjusted gross profit margins in the third quarter increased to a very strong level. They've always been strong because, again, for us, it comes back to the ability of our platform to do hard things. So we don't talk about take rate. We don't publish take rate. But for us, it just starts with how can we drive efficiently customer outcomes and when we do that well, we can have a very profitable business.
Operator: And our next question comes from John Roy from Water Tower Research.
John Roy: So Jim, you've talked obviously a lot about predictive audiences in healthcare. I was wondering, are you going to be able to leverage that into other verticals quickly? Is that going to be something that's going to be a lot of repeatability or is it a lot of a whole new designed to get that type of activity in other verticals?
James Lawson: That's a great question. The reason why we're so excited about the predictive audiences capability is that it is an infrastructure that we've built. There is an audience building capability in our platform, and it is a replicable approach and method to creating audiences using data and facts rather than assumptions. And we rolled out our first predictive audience for health, and it's been very, very well received. And that particular audience consists of health claims data that is the source. And from that, in this case, it's unique. In healthcare, it's different. You need to make sure that you don't -- there's a de-identification layer when you're dealing with healthcare. You can't have a seat audience. You can't have a seat audience and then work directly from that in healthcare, because there are rules under HIPAA about the extent to which actual condition suffer information can be used. And many platforms take, for example, 10% condition sufferers and then they put 90% non-condition sufferers into a segment and they mix it all up, and that becomes a HIPAA compliant targeting segment. Well, in our world, it is a much better and more data-driven and exacting method to understand some of the statistics of the audience and define the statistical parameters of condition sufferers without getting into who they are. And then from their target KPIs based on actual conversion activity within the context of that audience. But the beauty of predictive audience is, we are able to optimize campaigns based on KPIs as we always have. In other words, we score impressions based on the likelihood that those impressions are going to drive a sale or a visitation or what have you. But for those customers who really care about audience quality. In other words, I only want to talk to people. I only want to consider serving an ad to people that are within these audience parameters. We can do that as well. And I think we're doing that and making that something that can be activated on a self-service basis in a turnkey fashion is game-changing for AdTheorent. So we're very excited about it. We're excited to have health resonating so much with our big customers. And we're really excited to talk about this as we go forward across a number of different verticals. It's a big opening for us in the market.
Operator: And ladies and gentlemen, with that, we'll be concluding today's question-and-answer session. I'd like to turn the floor back over to James Lawson, CEO, for any closing remarks.
James Lawson: Thank you for joining us today, we appreciate everybody joining to listen to our Q3 update. In the third quarter, we delivered revenue on the low end of our range, but we believe we're doing the right things to position this company for long-term success and profitable growth. In market, we are providing transformative media buying products and capabilities. We believe the programmatic advertising industry needs to graduate from user ID targeting, and we are the platform and company that can lead this transformation. We are well on our way. We have a materially better business than we did a year ago, and I concisely look forward to sharing further updates. Thank you for joining us today.
Operator: Ladies and gentlemen, with that, we'll conclude today's conference call and presentation. We do thank you for joining. You may now disconnect your lines.
