WPP plc (WPP) Earnings Call Transcript & Summary

Good afternoon, everyone, and welcome to the third in WPP's webinar series, which goes deeper into different parts of our business. I'm delighted to introduce you today to Nicolas Bidon, who is the CEO of Xaxis and also members of his leadership team. As you know, Xaxis continues to be a real growth story for WPP. And I certainly think it's a great example of how we've been able to stay ahead of the curve in terms of the industry changes through innovation and investment and actually demonstrates how digital and programmatic are actually growth drivers for our business rather than a threat. I think you all know the format by now that Nicolas and his team will present to you for about half an hour, I'll say maybe a little bit over the half hour and then we'll open up to Q&A. So I'd like to say formally welcome Nicolas and his team. And Nicolas, over to you. Thank you.

Thank you, John, and good morning or good afternoon, depending on where you're joining us from. So on the call today, you will see 3 members of my global leadership team, and I will let them introduce themselves in a moment, I'll start by introducing myself. So my name is Nicolas Bidon, I'm French, as you can tell, I'm based in London. I've actually been the global CEO of Xaxis for about 4.5 years, at Xaxis for 9, so very much at the beginning of the adventure. Before that, I was in the leadership team of Yahoo for Europe. And prior to that, spent 7 years in the U.S. including 4 in San Francisco, working for various enterprise software vendors. So ad tech, which is the intersection of media and software was very much where I was, I guess, destined to go. And that's it for me. I'll turn over to our Global COO, Sylvia, who will introduce herself now.

Thank you, Nicolas. My name is Sylvia Sparry, I've been with the company since 2013. I've been in various leadership roles in the U.K. and EMEA teams and I'm now global COO. So as part of that, I lead things like platform excellence, automation and partnerships from a global perspective. And prior to Xaxis, around operations and data size for contextual ad tech business for 7 years. So Arshan, over to you.

Thank you, Sylvia. Hi, everyone. I'm Arshan Saha, CEO of Xaxis Asia Pacific. I'm Malaysian. I was born in Malaysia, spent a couple of years in the U.K. and now I'm based in Singapore. I joined Xaxis as the second employee in Asia back in 2012 and back then to start to launch the business in Southeast Asia. Prior to Xaxis, I worked for an ad network spanning across both the U.K. and Southeast Asia as well. So over to you, Sara.

Thanks, Arshan. Hi, everyone. My name is Sara Robertson. I'm the Global VP of Disruption at Xaxis. And I guess I'm kind of a newbie because I've only been with the company for 6 years. But during those 6 years, I've spent -- I've had a ton of fun building AI products and leading product and engineering team based in New York City. Prior to Xaxis, I worked in the music industry. I was at Warner Music Group, and I worked with a lot of rock stars. Back to you, Nicolas.

Thank you, Sara. So in terms of agenda, we'll spend the beginning of the hour speaking about programmatic and Xaxis evolution, and we want to zoom in on a couple of capabilities what we think are very differentiating for our clients and for the group. Give you an overview of where we see the growth opportunities moving forward and then wrap up to move on to the Q&A part. So next week is a very special week for Xaxis. It's great timing because it marks actually the 10-year anniversary from when the company was first founded, which was June 27, 2011, was the initial press release from WPP announcing to the world the birth of Xaxis. And when you look at what the vision was back then was really trying to put together specialists from across the group and technology assets, including from a previous acquisition of 24/7 by WPP to really create a center of excellence around programmatic. And I think credit to Martin Sorrell at the time and Brian Lesser, the original founder of Xaxis, I think they saw how important programmatic was going to be in terms of media and really wanted to lay the foundation to create that center of excellence. So before we move forward, maybe it's worth recapping what programmatic advertising is. And there's no universally accepted definition. So this definition is very much our own. But I think it gives a good sense for what this is about. And at the core, programmatic advertising is a way of trading and buying digital media, making heavy use of data and technology to make use of different data signals to make real-time decisions, often on a per impression basis. To really figure out whether this is the proper setup for a brand, the proper audience to show their ad to and also making decisions on things like what creative or offer should be shown to that audience and what price should be paid for the opportunity to show this advertising. So it's a pretty broad, as you can see scope and definition. I think the simplest way to think about it is programmatic enabled the use of data and tech to make digital media buying more efficient. So it's about digital media. And why does that matter, I think, is because, as you well know, digital media is the biggest part of the total advertising market. In fact, since 2016, it is the dominant category into the overall media landscape. And if you look at the chart on the left, you can see that in some countries like Norway or China. It's close to actually 80% of the share of wallet of advertisers. And worth noting is the fact that this share is increasing in all markets. So within digital media, programmatic, which is what Xaxis really specialized into has also become the predominant ways of trading. If you look back to 2011, which is where we were founded, it was only about 9% of the transactions or the ways of buying digital media. And in the U.S., which is what this chart is about it represented a market of about $1 billion of media investment. You fast forward 10 years to where we are now, the estimation is that it will be close to 90% of the way digital media is traded and representing a marketplace of about $100 billion, if you round up to easy-to-remember numbers. So programmatic, I would argue, has become synonymous with digital media. And that explains why the outlook for programmatic is strong despite already all the growth that have happened before. What you're looking at here is basically an index which is putting together the Wall Street consensus estimates for revenue growth for some of the publicly traded companies that I think are quite representative of ad tech and programmatic. So companies like the Trade Desk, Criteo, PubMatic, et cetera. And you can see that while there was an impact like most businesses last year in terms of COVID, the double-digit growth of those companies and the forecasted growth is still extremely strong as more media goes into digital and more digital media gets traded programmatically. So that was the initial vision for Xaxis creating that skilled specialist with a core competency around programmatic. And in the last 10 years, we've really delivered, I think, on this initial vision, building a specialist within WPP and GroupM, which employs more than 1,500 people across 47 markets and basically stewarding $1.5 billion of spend for clients on all programmatic channels. So 10 years has been very successful, but it's been 10 years of evolution and transformation as well. And if you look at this slide, this is actually a screenshot of our original website back in 2011. And you can see that the way we portrayed Xaxis was talking about what we call a programmatic audience company. So basically, the goal was to help brands reach audiences at scale across all these different digital channels, leveraging the buying methods of programmatic. But what we saw as the years went by is that clients wanted us to go beyond just delivering audiences. They wanted us to focus on delivering actual business results for their brands. And that's why about 4 years ago when I became CEO, we rallied the company around a new vision, which was to basically become the leader in what we are calling outcome media. And I'll explain to you in a little bit what that means to us. So the best way to maybe illustrate this is to talk about the work we do for clients. And I'll give you 3 examples of outcome-driven media in action. The first one is the work we've done for a large FMCG client, where we worked with them on savings. So the outcome they wanted to deliver is basically to make their video investments go further. And what we did is we leverage our Copilot platform, which Sara will talk to you more about in a moment, to achieve amazing results, in that case, reducing their cost per completed view or CPCV on their video investments by more than 50%. So the outcomes they wanted was savings, and that's what we delivered. But it can be also outcome media tool in the toolbox to deliver growth. And a great example of that is the work we've done with a large retailer in the U.S. who wanted to really boost their e-commerce channel and worked with us to incorporate some of their first-party data into our technology stack, including Copilot, where we delivered basically an algorithm that resulted into $17 million of incremental revenue from those campaigns while reducing the cost of investment and therefore, delivering increased return on that spend. And last but not least, for the clients who maybe have less of a digital sort of channel when it comes to tracking conversions, we're working with them to basically create what we call custom outcomes indicators or COI, which are essentially bespoke metrics or composite index, looking at the factors that are the most impactful when delivering the outcome they want for the business. So in the case of a large auto manufacturer, I think I can actually say it's Ford because we've presented that case before publicly. We were able to help them lower the cost of what they call PIE or purchase intent index by 25% and increasing the number of test drive request by 5x. So these are all examples of what we can do when it comes to outcome-driven media, but best to maybe hear from a client in the Nordics, who has worked with us for many years to explain in more details what we've been able to do for them. So let's watch a short video to go over that. [Presentation]

So I think that's a great example of what outcome media can do for clients. And the reason we're able to deliver this kind of performance on campaigns is because of those 5 pillars we believe are the bedrock of our success. So I'll go quickly over them. The first one is scale. We do think that scale matter as programmatic is so now predominant. And that means scale of access to the best quality media at the best possible price, which, of course, is something where we benefit from being part of WPP and GroupM. It's also the scale of our international footprint and the fact that we operate for global clients at a global scale and also the ability to create the center of excellence in this culture that will be attractive to the special type of talent that you need to be successful in this space, which maybe isn't the traditional talent that agencies would typically hire. So the people that work in Sara's team, data scientists, machine learning engineers together with media experts and traders. Data is, of course, key. And I think this is what we call the comprehensive framework to leverage data in the best possible way. That includes our clients' data. And in the case we just saw with Landal that's again maybe having direct access to their sales data as well as to the list of their parks and the geo coordinates of those. But it's also third-party data where, again, our scale gives us the opportunity to create relationship with premium data vendors that we can then leverage on behalf of all of our clients. And last but not least, the last 2 will go deeper into examining them because we feel they will be key to the next tenure of our success AI is at the core of what we do for the last 5 years, and Sara and her team continue to augment our capabilities in that front, which we think are very differentiated compared to other agency holding groups as well as creative, which we feel is becoming more and more important, we think it's the last big leverage to deliver business outcomes for clients, and we've made some specific and significant investment in that space into something we call Xaxis Creative Studio, which Sara will talk to us again about. So this is the 5 things that have been constant throughout our history. But as I said, it's been a story of evolution. And one of the sometimes feedback we got in the very beginning of Xaxis 10 years ago was that it was too much of a black box. It was working for clients, but they wanted to see more details to be reassured about where their media was shown, whether it was the right environment, making sure there was no compromise on brand safety or viewability, all of these important factors of quality. And that's something we've taken into account, and we now deliver full transparency around all of these dimensions as you can see on the slides. One thing we've done as well over the last 10 years is -- sorry, I have trouble moving the slide, is making sure we flex our commercial models to reflect the desires of our clients. So if you look at initially how we launched, the focus was always to include what we call an all-inclusive outcome-based model, which is if you look at this slide, it's a bit complex, so I'll walk you through it. But if you go from left to right, these are some of the components you need to be successful in your programmatic programs and Xaxis initial model, which is still the predominant model for most of our clients, is, again, to bundle all of these costs into an offering, which is centered around a very specific client outcome. But as we grew with some of our clients and talked to them, we realized that some of them wanted to have a slightly different way of working with us. For instance, they wanted to own their DSP contracts with the Googles of the world or Trade Desk or Xandrs. And they also wanted to get more visibility into what components of the total cost was the media cost. And for those, we've offered and are offering basically different ways to work with us, which is more based on a percentage of media model. And last but not least, we continue to experiment to make sure we are client-centric. And for clients who maybe have internalized some of their services around what we provide, but are interested into being at the cutting edge of innovation. We're looking into deploying Copilot into more of a consultancy model or software as a service model. And again, that's something Sara will tell us more in a moment. Another factor revolution has been scaled, which I mentioned a few times already. And if you look at our largest client today, we manage an 8-figure annual programmatic budget for them and deliver basically campaigns in 34 countries and do 125 test-and-learn pilots over the last 12 months to continue to make sure we have a clear learning agenda as the market continues to evolve as well as their goals. And to give you an idea of the scale of operation beyond those metrics, for them, we deliver more than 2 billion ads per month across all programmatic channels. So that includes, of course, traditional display and video, but increasingly new channel as well like out-of-home or connected TV. So it's been very much a story of transformation for those first 10 years. And I've seen it firsthand because, as I mentioned in the introduction, I've been here for 9 years, I joined 1 year after the formation of the company. And we feel we've built on that initial vision of creating a specialist that can deliver value at scale for global clients. And we feel very strongly about the fact that transformation is core to who we are. In our space, things move too fast for us to think that transformation is a moment in time in our history. We believe that we will continue to transform and evolve over the next 10 years. And in fact, I'm going to transition now to Sara -- no sorry, to Silvia, who is going to take us through some of the trends that we think will be important when it comes to transformation of our clients' business and transformation of Xaxis. So over to you, Silvia.

Thank you very much, Nicolas. So as Nicolas has demonstrated transformation is a constant for our business. And I'm going to spend a little bit of time to talk you through 4 key trends that I think will further shape the future of our industry. So the first one is, of course, the end of third-party cookies. So as all of you have probably heard already, Chrome, which has a 60% market share of browsers worldwide, will eventually stop supporting third-party cookies. And while Google announced yesterday that they would move the duplication date to the end of 2023 to allow for some more time for alternative methodologies to cookies to be properly tested and implemented by the industry, the direction of travel towards more privacy and consumer-centric ad ecosystem is clear, and it is irrevocable. And Apple's app tracking transparency initiative is further restricting what advertisers and brands can do with data. And all this will reshape how we plan and how we activate and how we measure results driven by online advertising. Secondly, I want to talk to you about the golden age of video. Again, we see consumer behaviors are changing irreversibly here. With things like cord cutting and companies like Netflix investing billions in content production, we see enormous growth in this channel. Digital video channels are already -- have already got a higher market share than traditional linear TV, and that trend is only going to accelerate. We also see other emerging programmatic channels coming of age. The digital out-of-home market is already estimated to be worth $8 billion globally, and programmatic audio is a $1 billion market in the U.S. alone. And beyond this, we see channels like gaming and podcast emerging, which open up even more growth opportunities. Another channel that has seen enormous growth during the pandemic is, of course, e-commerce. We've all had to adapt the way we shop, and this is reflected in the increasing adoption of e-commerce across the globe. So as a percentage of retail sales, we have seen a jump from 13.6% in 2019 to 18% in 2020, and that represents a massive jump in just 1 year, which translates to transformation and opportunities for the brands that we work with. But what all of these trends also illustrate is that there is increasing fragmentation and complexity and that this is going to accelerate further. And brands are looking for solutions and partners that can help them navigate this complexity. So to illustrate the density of the fragmentation, what you see on this infographic here, all of the major players of -- companies that offer solutions in our space, and there are over 8,000 logos on this slide. And so it's essential that advertisers are provided with a path beyond the fragmentations of channels and technology silos to reach consumers. And as Nicolas has outlined earlier, our 5 core pillars helped set Xaxis up to success in this otherwise complicated and fast-changing environment. And 2 of these pillars are our own AI and our approach to agile creatives. And Sara Robertson will talk you through this in more detail now. Over to you, Sara.

Thank you, Silvia. So when we started building Copilot 5 years ago, we set out to build an artificial intelligence platform. What we quickly learned was that the words artificial intelligence scare people. They hear that, and they immediately think that we're trying to replace their jobs and automate them away. And what we landed on through this process is the goal of co-pilot is not to replace people, but really to give them amplified intelligence. To make you smarter and sit next to you and help you with your path of extremely difficult programmatic activities. Copilot solves a lot of problems for our clients and our traders. I'm going to walk you through a few of them right now so you can get an idea of what we're dealing with. If you remember the slide that Sylvia just showed with 8,000 logos on it, you can imagine the difficulty that people have in trying to find value in all the data that all of those 8,000 little systems are generating. Copilot's AI helps you leverage that data by searching through it for the strongest signals that map the closest to your business goals. Another challenge is that online KPIs don't really reflect real-world business goals. What does a click have to do with selling a car? Or what does an impression have to do with getting more foot traffic into your restaurant? With Copilot, we can take multiple online metrics as well as off-line ones and combine them together into what we call a COI, which is a composite index that more closely maps to business values. Another thing brands are looking to do is increase efficiency even as we can see our ecosystem is getting more and more complex. This is a great problem to apply AI automation to, to help you increase efficiencies where systems are too complex for humans to deal with. And finally, everyone wants to improve performance and save time. What Copilot lets us do is dynamically predict where the performance is going to be so that we can get warm starts in our campaigns and get a head start on performance when we first start spending our budgets. So this slide kind of helps demonstrate the comparison between 3 different approaches that can be taken on programmatic. The first is just a straight up manual human approach. Humans are really good at interpreting data and can come up with the best insights, but they're also kind of slow. So when we try to do test and learn with a manual strategy, we have to come up with the test, run it for an amount of time, stop the test, evaluate the results. So there's a limited amount of test and learn cycles that we can do when we're doing everything manually. The DSPs have hyper-optimized this process, and they will very quickly zero in on a laser-focused amount of performance. But that tends to overlook changes in the ecosystem, changes in the campaign and eliminate some of the prospecting value that we try to deliver for our clients. What Copilot does is create a tree structure with a lot of branches. And at the bottom of every branch, it optimizes the bids for that particular pocket of performance, which can outperform both the manual and the DSP performance strategies. This screenshot is an actual screenshot of our live Copilot product. I know that might be hard to believe because it's so beautiful and people are used to seeing vaporware in presentations. But I really want you guys to know that this product is mature. It's beautiful. It's useful. This is an example of a visualization that we created to help our traders understand all of the little branches that we make when we're doing a bidding strategy. And you can see we make thousands and thousands of branches. The product is live in over 43 countries. We run more than 1,400 models every day. We've served over 2,000 advertisers by now, and we're constantly generating new customized algorithms for our clients. A little bit about Copilot history. As I said, we've been in development for over 5 years. When we first started, we had this idea of creating an easy button. We thought we would make one algorithm that would just do everybody's jobs and make life so easy. What we quickly discovered was that, that's not how the world works. There's different use cases, different problems to be solved, and it's better for us to create different algorithms to solve different problems. So we expanded into multiple algorithms on a single DSP. By 2019, we had multiple algorithms across the top 3 major DSPs, which is where we're at right now. In 2020, we launched our first beyond bidding algorithm, which is a budget allocation strategy. Budget allocation sits above the bidding algorithms that we had created previously. So now we have this stacked approach to optimization, which is really elegant. And finally, the place where we're going right now is we're starting to build an algorithm marketplace. Our goal to algorithm marketplace is to make activating these algorithms as easy and commonplace as activating audience segments. Everyone in the programmatic space knows how to activate audience segments. We want them to know how to activate our algorithms as well. To this end, we've secured a significant additional investment from the WPP product board, and we're under heavy development for this idea right now. So Copilot at its core is AI. Everyone wants AI. But AI is difficult to do. I'm going to talk to you right now about 3 main areas we're focusing on here. The first is, to have good AI, you need good data. The machines just want to consume as much data as possible. So we spend a lot of time looking for new signals that we can feed through our algorithms that will help it perform better. That might be a customer's first-party data stores. It could be integrating with some of these clean rooms. It could be looking at off-line data. Basically anything that we can find that's a signal, we can feed to our algorithms and start using it for optimization decisions. Secondly, we're looking at expanding the applications of our AI solution. Right now, we're heavy on the programmatic space, but we're also looking at expanding our AI optimization into search, social, other channels, e-commerce and things like that. It has already proven its value in programmatic. I think it will be easy to prove its value in other channels as well. And finally, anyone that's building AI needs to be a good steward of the technology. That means paying attention to explainability, making sure that we know why the AI made decisions, being very careful with our data, providing consumers with dignity in our targeting approaches and basically being good partners to the industry as everyone is trying to figure out what ethical AI looks like. So Copilot's AI focuses heavily on optimizing the media spend. But we all know that you can optimize media until the end of your days and if you're creative, isn't doing what it needs to do, you're not going to get the performance you're looking for. Creative has a nearly 50% influence on a decision every time we serve an ad. And we tend to have very little impact into the optimization levers available on creative. So what we did was launch Xaxis Creative Studios. The idea here is that we can take the creative assets that we get from our agency partners and we can start to reorganize them into formats that are more applicable and more optimizable in the programmatic space, whether that is dynamic creative, conversational ads, shoppable ads all kinds of exciting new innovative formats. We are now able to deliver this as a service to our clients and in partnership with our creative agencies which gives us the opportunity to apply our AI to that extra 50% of optimization that we had previously been overlooking. To talk to you a bit more about how we apply these differentiations, here's Arshan.

Thank you, Sara. Let me now just walk you through where we see some of the biggest opportunities for growth. And again, how we have positioned ourselves to fulfill this. So starting with integrated video. There are a variety of video advertising methods and formats out there today within a typical user's media consumption journey. And this is especially so in Asia as well with the proliferation of high-speed Internet. If you think about the many screens you face in a day, more often than not, video investments end up fragmented. Fragmented video buying decreases the overall effectiveness of running a good video campaign and often leads to waste the media funds and suboptimal ROI. So therefore, what we did was build a solution to bring 3 things, which is reach, performance and savings. Our cross-screen video solution drives business outcomes across all stages of the consumer funnel and across the plethora of content from broadcast right down to user generated. By leveraging the different types of video formats that be short form, native, interactive and in-stream video. Advertisers can run an always-on video campaign in a conversational manner to build excitement, remain top of mind and also from use action. Integrated video very simply allows an advertiser to maximize the video investment with the right mix of data, technology and creative. If we go to the next slide. The second is on digital out-of-home. Firstly, I think the last 18 months has had a massive effect in terms of digital transformation, as we know. And out-of-home advertising has perhaps been one of the beneficiaries as we've seen the acceleration of the fiscal world getting digital and with digital than addressable. Out-of-home has simply today become digital out-of-home in terms of infrastructure, and that paves the way for scale and addressability. Out-of-home complements other digital channels like TV, audio, as it engages with an audience on the move outside of the noise of digital devices that they're already exposed to. So sideline, as you can see, is our advanced programmatic digital out-of-home solution and technology suite. It helps plan execute and measure digital out-of-home campaigns as well as integrate them into a holistic omnichannel programmatic strategy. Sideline essentially does 3 things: precise media planning and buying on the basis of people, not panels using real-time data; agility and flexibility in execution utilizing departing specific day period booking; and thirdly, advanced measurement allowing attribution to business impact, such as footfall or brand lift. Go to next slide. This capability showcased in this case study for Domino's Pizza in Malaysia, we were able to maximize reach and scale, whilst also still having the flexibility of serving multiple creatives. Domino's was able to build connections in real time media target audience and understand the impact of out-of-home advertising on their consumer journey. Digital out-of-home formats were tailored in this specific instance to audiences in real-time basis on variables such as vehicle type, weather and traffic condition. This led to a highly effective reach and they were also able to measure an attribute that is creating a footfall uplift of circa 10%. 65% of all Facebook revenue and 55% of Google ad revenue are self-served proving automation is the key for greater scale and efficiency. To fulfill this promise in this segment, we've created a self-serve marketplace. It's primarily a platform that provides an end-to-end solution for medium-sized enterprises to tap into the vast competitive media solutions and further amplify their campaigns. Through the self-serve platform, medium-sized enterprises are able to do 3 things, gain control of the media plan, have access to the best curated media deals and also understand which ad channel and creative works for them. Go to the next slide. The focus here will be on our high LX markets in APAC, India and Indonesia. And if we zoom in to just India itself, where I'm looking at an example of in-target market opportunity within cities and sectors to be that proved to be global exporters. So just for example, if we look at Hyderabad, which has got a lot of exporters in the space of software solutions, retail and real estate, or Chennai, which has got a lot of focus on medical tourism and textile. We go to the following slide, please. And lastly, on commerce. Xaxis commerce connects brands and shoppers across the retail media spectrum. Retail media essentially provides a huge growth opportunity to help brands acquire new customers, win market share and drive repeat purchases. Our commerce excellence solution helps brands better understand consumer behaviors by connecting online and offline data points, create more accurate target audiences and finally, drive consumers towards purchase across the retail media spectrum. How we enable this solution is by our property build capabilities in Shopbox, firstly, which is a commerce intelligence engine, that identifies local factors that influence shopping behavior and helps brands activate those insights. Copilot, which Sara talked about, our amplified intelligence, making thousands of data-driven decisions that optimize the business outcomes that really matter. And lastly, creates -- Xaxis Creative Studios that enables impactful programmatic creative execution across multiple digital environments. So very much bringing the best in breed from our tech tools and capabilities to purposefully address client needs in commerce. That's all for me. Over to you, Nicolas.

Thank you for that. Thank you, Arshan. Just to wrap up. I think if you look at the last 10 years, while it has been a story of transformation, there is a blueprint, which I think can be repeated and has been repeated throughout our history. And this is what I call the Xaxis flywheel. To look at the reasons of our success, it is about creating an attractive place for attracting, developing and growing world-class specialists in the area of programmatic, historically, but more and more in the area also of AI, machine learning and data science. This allows us to create differentiated technology, Copilot being the best example of this. And then when we apply this talent and this differentiated technology to the problems that the wonderful brands that GroupM agencies work with, we're able to come up with very innovative solutions, which can derive better outcome for clients. Therefore, fueling their growth and scale and ours, which allows us to reinvest into the business and continuing that virtuous cycle. So this is very much the model that has worked for us, and I think it will be the model that also works for us in the future. But the type of skills might be slightly different. So if you look, for instance, in the specialisms that we are developing, programmatic, of course, remains the core of who we are and what we do. But as we discussed, as the ecosystem continues to change, new skills like the ability to be comfortable about operating large data clouds, for instance, or applying machine learning algorithm to client-specific issues will become more and more important to be successful in the future for our clients and for Xaxis. Likewise, this ability to deliver creative ad scale in a very agile manner to be the biggest level of performance will remain key. And so I think we will apply those skills and this type of talent with our technology platform and initiatives to continue to innovate and deliver better outcomes for clients. As Sara alluded to in our part, our ambition is to go beyond traditional display and programmatic media and try to bring these competencies and skills to all digital media, which increasingly in most countries means all media. So that's all we wanted to take you through today. And I think we'll move on now to the Q&A section.

So I'm going to look at what those questions are, and I will make sure that I share them with the team so that we have a chance to hear from everyone. So I think one of the question is regarding the implication of Chrome's decision to delay the Privacy Sandbox effort. And I think more broadly, Richard Kramer asked about how the inevitable -- I'm reading the inevitable signal loss from the demise of mobile ideas and third-party cookies affect or will affect the mapping files and bridges, IDs that Xaxis uses for targeting and activation. So it's a very broad and complex topic, but probably one for you, Silvia, to give a few elements of response about.

Okay. No problem. I'll start with the question about the signal loss. So I think that there are several elements to this. Our demand-side platforms, the DSPs that we work with are developing alternative -- either alternative identifiers or alternative means of measuring and targeting. So for example, instead of using an ID, I am maybe using a user cohort that combines several IDs into an anonymized group of users. So my ability to target is still not diminished. I'm just using a kind of different method in order to achieve this. And there are also other interesting signals that we can use as a targeting input. So let's say, a client's warehouse stock market levels, that is something that I can utilize as a data input for Copilot and then activate against because I can target the campaigns with more volume in the areas where those stock levels are high. So there are plenty of opportunities for me to still map other identifiers and translate that into an effective campaign. Talking to the delay in the implementation of the third-party cookie duplication and that being moved to the end of 2023. I think that's a much needed opportunity for the industry to properly test out these new targeting methodologies. They need to be properly implemented by publishers, by SSPs, by DSPs. They need to be tested by individuals like us from a performance delivery perspective. And it gives us the opportunity as an industry to move forward in a constructive manner that allows enough time for the bedding in of these methodologies. So it's a, I would say, a much welcomed and much needed step that Google has taken yesterday.

Thank you, Sylvia. I think there's one specifically to digital album from Julien, which basically says, you have been developing your own solution sideline, but most out-of-home providers, they go to have their own tech stack, for instance, VIOOH for JCDecaux. How the site line interact with this tech stacks. Maybe Arshan, that's one you want to take? And if not, I'm happy to take it as well. Just let me know.

Sure. I can do some of it, and you can probably chip in. So I think the way that we look at sideline is in order to organize structure and simplify the entire value chain of digital out-of-home. And within that interaction itself is also then working with companies like JCD, for example. So what we would do with a company like JCD is by our proprietary supply onboarder, we've been able to take on all of that inventory that a partner like JCD would have. but also simplify and also onboard a lot of the inventory that currently might be fragmented on a local market level by numerous vendors. So we would aggregate all of that supply into a proprietary onboarder from which where the differentiator that we bring into the mix is a proprietary audience planner, which can help to plan based on people and not on panels. And that's where the secret ingredient or the differentiator or proprietary tech comes in. I hope that answers that.

Thanks, Arshan. I think that was quite clear, so I'm not going to add to that question and move on to the next one. Tim Nollen asked, can you please talk about programmatic CTV? Is this still really done via direct deals, i.e., the ads programmatically but not bid through automation? How much is done through private marketplace will open exchange are to be ever really become a meaningful method in programmatic CTV? So maybe I'll take that one to give you guys the rest. I think connected to this is obviously a very growing part of the ecosystem when it comes to programmatic. And I think in media, there's a saying that the money follow the eyeballs, right? And as the consumer behaviors continue to change and of course, the lockdowns in the world have been an accelerator of this trend and people watch TV differently through connected device and connected TVs, the audiences on those platforms tend to grow. I think that's the first bit. The second bit, which is fueling that growth is I think advertisers have come to now expect the kind of ability to measure, to target and to reach their consumers in a more efficient way than maybe they have done in the past with linear TV. So they want to take very much that mindset that comes from digital and apply it to TV, which, of course, has extremely engaging content and is a great medium. So this fuels the growth of CTV. This being said, I think there is today a constrained supply when it comes to quality CTV environment. And as a result of that, the primary methods of trading tend to be, as I think, was highlighted in the questions, more focused around things like private marketplace where, again, you can apply data and technology to, for instance, choose what creative message to show or even whether or not to show an ad at all. So for instance, you can do very interesting things around geo targeting, focusing on leveraging geo data on that kind of CTV inventory. So you still use a programmatic way of trading, if you want, but it is not real-time bidding methods. And I think that speaks about the fact that today that quality supply is constrained. I think this may change as consumers sort of behaviors continue to change. And I think more players emerge in this space, and we've seen already a few emerge. If you look at, for instance, IMDB, which is the offering from Amazon, which now has a lot of interesting ad-supported content, which is accessible in a programmatic fashion. And again, that may actually create enough scale for it to make sense at some point to be traded on a real-time bidding method. So I think it's a fast evolving space for us. We really see this as part of this integrated video approach that -- or consolidated video approach that Arshan took us through. So we see connected TV as just one more channel where you can apply the same sort of programmatic methods as you would do on YouTube or on digital video or on short-form video, for instance. And so this is the way we look at it. I think Finecast, which if some of you were part of last month's deep dive from GroupM was mentioned a couple of times, their approach is very much to look at it as an extension of a linear TV buy, and they are 100% focused on delivering that for clients. So that's the various approach we have on that. But it's a fascinating space. So I think we have about 5 minutes left, so I'm going to go and look for more questions. Sorry, I'm having trouble actually, getting back to the beginning of the questions. I think that one, maybe I'll leave to you, Sara. It's about our relationship with DSPs and SSPs versus direct to publisher. And the question from Tim goes, any sense of what proportion of your ad buys go through each and relatively who is gaining in relative importance among agency media buyers Xaxis between DSPs and SSPs? For me, I see them as 2 sides of the same coin because they basically are different part of the ecosystem, but, Sara, over to you to give us more on that.

I mean I think I'll give the same answer. The short answer is that almost 100% of our spend goes through the DSPs, and that's because the SSPs are listing their supply within the DSPs. So there's no real reason for us to go direct to the SSPs for those buys. We do most of our work through the demand side platform. In terms of gaining relative importance, Again, I'm not going to say that SSPs or the DSPs are gaining more relative importance. I think for us, it's more which of the DSPs is gaining more relative importance? Which of the demand platforms do we need to be on so that we can deliver that demand to our clients? Generally, the supply is always available in those demand platforms. In instances where it's not, then we can go and try to get it in there. But generally, we're just operating on the demand platforms as hard as we can.

And I think if I can add one component to this. I think for us, it's very important that we want to be agnostic when it comes to DSP choices. As I alluded to earlier in the presentation, we find that more and more of our clients want to own their DSP contracts directly with those tech vendors and often make some sort of standardization decisions around which DSP they want to use on a global level. So it's important to recognize that all of the differentiators that we've created and continue to create to innovate things like Copilot are applicable regardless of the DSPs that our clients choose so that, again, we can bring the value of machine learning and AI regardless of whether the clients want to consolidate maybe their DSP spend on Trade Desk versus Xandr versus Google, for instance. And the reality is that for most clients, you still need to use multiple DSPs because they don't give you access to the same supply, as Sara alluded to. So I hope this answers the question. So I think we have 2 minutes left. So I'll take -- this one is an interesting one. The question, anonymous is, why is Xaxis not an agency or being positioned as one? Is it a plan of WPP to keep it hidden behind the mainstream agencies? So I'm not going to answer the WPP part. But I think when you look at our business today, I think 90% of our clients are clients of GroupM media agencies. So their clients of MediaCom, Mindshare, Wavemaker, Essence, M6, et cetera. And frankly, that allows us to have, I think, a very symbiotic relationship where they can focus on the bigger picture and the macro decision of helping clients with overall media investment strategy between off-line and online. Also things about transformation and the evolution of that media strategy to become more and more a data strategy conversations. And I think for us, we can focus on being cutting edge when it comes to activation across all programmatic channel with the intention, as Sara, I think, alluded to as well to increasingly go beyond pure programmatic channel on display to other type of channel like search and social. So that, I think, has proven to work well for our clients and for Xaxis and for our agencies. I do feel that as more and more brands will maybe look for a different type of agency services. So this could be direct-to-consumer brands as an example or just maybe this kind of a scale-up type of business that are maybe too little to basically do a TV at scale, but are actually big enough to need help when it comes to digital investments in general. I think that there may be an opportunity for Xaxis to grow our involvement with this type of clients. So maybe something to look forward to for the next 10 years. And with that, I think we're actually out of time. So I just want to thank WPP for giving us this opportunity and thank all the speakers, Silvia, Arshan and Sara. And I hope you enjoyed the conversation.

Well done, everyone. Great job. Great job, Nicolas and team.