VIDEO TRANSCRIPT:

Good morning, good afternoon, good evening. Thank you for joining us today. My name is Ray Velez. I'm the Chief Technology Officer of Customer Data Solutions at Publicis Sapien. Today's session, we're going to be talking about AI-powered paid media measurement and optimization to drive growth. And what we mean by that is we're going to be explaining how we help our clients in partnership with our friends at Amazon Web Services optimize paid media. your strategy and tactics, both across channels, audience and campaigns to help drive more accurate granular and timely measurements. And these measurements are enabling powerful outcomes for our clients. And we'll talk about one of the case studies where we have done that for a large global QSR. So today I'm joined by Andre Engberts, Vice President of Technology, Pubicis Sapien.

Thank you, Ray. Very nice to be here.

And Amit Kumar Sharma, Senior Director of Engineering, Publicis Sapient, and

And great to be on.

Tim Barnes, GM of Advertising and Marketing Technology Vertical at Amazon Web Services.

Hi, Ray. So very excited to be here with you today.

Awesome. Well, that thanks everybody for joining. Let's start the conversation with our first topic. So why is paid media measurement and analytics becoming a priority for advertisers?

Thank you, Ray. Paid media is key to acquiring, growing, retaining and regaining customers and advertisers have many media options they can use for ad copy and many creative concepts to choose from. But ad copy to select, which audience to target, ad match to target, that's all becoming critical decision points in driving better ad effectiveness. better return on ad spend. So to do more of what works and to learn from what does not work, that's the engine that's driving better ad campaigns and AI-driven analytics is essentially an integral step in this process for measuring attribution and lift by channel audience and creative and optimizing campaigns for the adjustment of these creative audiences and their placement. So there are several methods that are traditionally applied to big media measurements between media mix modelling, attribution modelling, incremental testing, lift study, and our client, this large QSR that Ray was mentioning, already had an MMM solution in place for overall channel performance analysis, but they needed a more granular, higher frequency method to measure attribution. that would incorporate lift in order to provide quick feedback to campaign strategists and to optimize campaigns toward lower operating cost and higher gains. The methodology and algorithms that we put in place in this solution address these needs and we leverage AI for higher speed analysis and more precision at scale. And Ahmed, from a solution standpoint, he's going to talk to you about why AI was more powerful than traditional queries. Ahmed?

Thanks, Andre. Yep, thanks, Andre. Think of traditional queries a little bit of like a flashlight in a vast warehouse. It is pointing to only what you can see. But the approach is too narrow for a large complex data which Andre explained about. Train media comes with large swath of data, partner data, first party data, zero party data. So we needed an approach which could solve for this. And AI is like turning on all the lights and revealing all hidden patterns for your exposure, demographic and conversion data audit all at once. And generating AI on top of it. It goes a step further. It's like having an expert guide you to talk to allowing analysts to ask questions and test their hypothesis. And when we looked at our partner AWS and they had all these hidden gems within them and Tim is going to talk a little bit about it.

Yeah. You know, Ray, I would argue that paid media measurement has always been a priority for advertisers. The problem is it's just become increasingly challenging to pull together disparate channels and signals into a cohesive measurement platform. So for this QSR, AWS worked with Publis' Sapient and Starcom to build a solution that combined the power of data engineering, AI and Gen AI on the AWS cloud. We combined that with AWS clean rooms and identity resolution services to create a unique solution that quantifies the impact of media, creative exposure and sources like programmatic display, video and TV channels. And we use high measurement frequency. for campaign planning and flight optimization.

Thanks everybody. And I think what's super exciting to underscore here is we're actually democratizing access to the data. So like we were talking about by leveraging AWS clean rooms by leveraging generative AI capabilities, it truly makes that data available to anyone not just a data scientist who's writing a Python model to execute on a measurement report. But anyone who can build a prompt and interact with the data with a conversational experience. And that's super exciting because as data engineering team, we've been getting that democratization data of data requests for years and years and years. And it's really great to see that result in a solution that drives that value output. So as we move on to the second question, I think one of the... The key drivers is what's the value proposition for doing something like this? So why did this QSR decide to invest in this particular solution? I know Andre talked about they already had a marketing mix modeling capability, but what made this solution different?

Sure, thank you, Ray. So in a competitive QSR industry environment, the main path to growth is through increased traffic coming to restaurants, because there is just that much that you can increase check size in QSR. Because people just want to eat significantly more and frankly at QSR, if they come to a QSR, it's also because their meal budget is limited. So within this context, incremental in-store guest visits was established as a primary metric as a campaign objective. And so planning efforts for the design of a solution were aligned to media channels, partner selection, and tactics that are the most effective drivers of incremental store visits. And these methods in turn were applied to the relentless evaluation and optimization of in-market activities to deliver the strongest performance. So our client wanted to have a strong data solution that would be built independently from their activation platform. And this, the reason for this is it was to provide an objective measurement methodology. So you don't get to grade your own homework essentially. But it's also to be able to bring data from measurement partners where that data would not have been necessarily allowed in an activation cleanroom. So the separation is also a question of data privacy, not just objectivity. The client had been already operating a SaaS environment for measurement for a while, but that original solution, even though it was meeting the independence criteria, was a SaaS product that was working more like a black box with heavy costs associated with data movement. So our clients wanted an environment that would give them more transparency on the use and the quality of the data, on the design and performance of algorithms, and with a lower total cost of ownership. And so the AWS platform met all these criteria and it was very quickly agreed by the client to go with the AWS environment for these reasons, because it had the right capabilities and also a much lower cost of data storage.

processing and data transfer costs. So I will hand it over to Andy to actually go further into the explanation of the solution details. Yeah,

Yeah.

I think when Klein came to us, the situation presented a clear challenge for this measure and media optimization efforts. We have to lay down some markets and mandates for working on this solution for long term. So first was speed and granularity. We must deliver frequent and incremental performance inside at the individual audience and creative level that requires that we need to be. working at complex data but at speed and looking at all complex and possible scenarios possible at any level. The other thing that Andre touched upon this as well, client was very clear that we wanted ownership and agility with this, right? So it must be a proprietary customer solution because they already burned their hand with gas. They wanted to be cloud native. AWS was great partner in these conversations and they wanted this to be nimble and they didn't want to be a complex solution. They wanted to be future proof and free from vendor constraints and we delivered that. with this solution. The most important and the final mandate was secure data integration. When we talk large complex data from third party or from the paid channels, they presented a challenge of how do you bring a data somewhere where you can apply the lens of complex data warehousing technology on it, right? And answer was very clear with us talking to AWS, AWS clean drone. There was no other matching solution out there in the industry. street and then Tim will explain a little bit about why clean rooms are shining light in this area.

Thank you, Amit. We work very closely with our partners like Google to SafePN to build a measurement solution for this QSCO that was flexible, extensible and met their privacy and security requirements. Additionally, this solution brings full transparency on paid media, measurement operations from data acquisition, analytics processing, and ultimately results publishing. This solution is seen by this large QSR as a really good fit to meet their overall requirements and help this QSR, who by the way was already an AWS customer. alleviate some of the integration and migration hurdles, streamlining that process for integration. And I think importantly, we see this as scalable across many customers, especially as we integrate with partners like Amazon Advertising through their Amazon Marketing Cloud solution.

Thanks, everybody. One thing I want to underscore that's important about this solution is that is it 100% legal and privacy safe, right? It's a little bit of anarchy with all the different regulations and rulings around how to treat your data by state-by-state level, especially in the U.S. And then when you start to think about going global, it gets more and more complex. So a platform like AWS Clean Rooms enables you to stay within the guidelines of legal and privacy safe requirements. And starting from this point where we want to move past the black box solution. And when you think about the legacy black box measurement solutions. that have grown up across the industry there's a lot of risk you're giving your data away you're getting a result back and you can't interrogate what's in between so a legal and privacy safe claim room is going to help you drive that interrogation it's going to help you unlock a whole bunch of new use cases which also become possible when you start to use a claim based platform like AWS claim rooms which kind of leads us into that next question which is for the team can you give us a peek into what What was actually built and then what was innovative about it as well.

Sure. So Aperture is the name of a platform that was built for this large QSR on top of AWS with the AWS Clean Room and the AWS AI engine, SageMaker, as it's called. And it's featuring person level analysis. I'm providing very high precision in measurement, hence the name aperture, and the platform is enabled by a collection of capabilities of data and technology partnerships. So what was in there was a data platform, the clean room, the AI engine and the identity resolution service infrastructure from AWS, all stitched together using the regular AWS architecture framework and data included paid media exposure logs from publishers, identity and data connectivity from live ramp, demographic data from experience, geolocation data for food traffic and public space sapient provides AI modeling and platform support services. So Aperture uses a specialized custom-built test and run approach powered by unique AI-based algorithms that estimate incremental contribution of media and creative at a very granular level. So the decisions can be both broad and detailed, right? Broad because of a scale that's coming with AWS platform and with AI. and detailed because we can go all the way to the level of a creative. And after a first year of operation, we are starting to see around 5% gains in investment optimization. And how this is put together, stitched together, let Amit get into more details on the components of this solution.

Yeah. Thanks, Andre. I think the main focus of this aperture platform was analyst-centric and benefit-focused. And I think Ray and Tim already all talked about how this QSRVP benefit today about this. But the key innovation in the solution was we fundamentally changed day-to-day work of market analysts. Instead of being slowed down by manual queries, they can now use single prompt explain. explore complex campaign scenarios predict our wise and discover new audience and icing on the cake at lightning speed using our product processing in the aperture platform which is anchored in highly scalable and secure AWS cloud infrastructure by leveraging technologies like Elizabeth cleanroom we are not only solving today's challenges but we have opened the door to our next frontier like deploying agentic AI to truly We make this as a conversational Ask Me Analytics solution. For example, you're looking for information without having a big backlog of technology team supporting you for new queries or new dashboards and new thing. No need. Just ask using a natural language and their platform will give you the results. All of this was possible because of the underlying AWS infrastructure and their AI capabilities. And Tim will talk a little bit about how AWS has... investing more in this area.

Thank you. Yeah, AWS is actively working and evolving our clean room technology and the underlying AI to enable more advanced use cases of third party and first party data collaboration between all parties. An example is our recently announced AWS clean rooms custom models. This feature allows AWS CleanRoom's collaboration partner to bring use case specific AI models into the CleanRoom collaboration and securely train the model with two or more data sets without ever revealing the underlying content of that data to others. This AWS CleanRoom feature is being used by Amazon Ads to power the Amazon Marketing Cloud custom model offering to create tailored audiences, measurement solutions, etc. using Amazon signals. I think it's just a great example of bringing the power of AWS Cloud, AI, and other tools and great partners like Publis to Sapient to solve large-scale ad tech marketing customer challenges.

Thanks, everybody. Hopefully that helps to give some insights of what the innovation is that we can unlock by leveraging these technologies and building these platforms. There's a couple of things I want to underscore, and these are topics for further research and I think helpful to understand how this is driving innovation. So number one is this concept of zero copy that's built into the clean room, and I'll talk about that. The future use cases, right, we've covered what's already out there and in production and we're helping to bring to brands on a global basis. But as you think about the proliferation of newer capabilities that are being unlocked with generative AI, like the Ask Me solution that Amit talked about, what really makes them possible and allows us to move faster than ever before is an innovation that really wasn't necessarily. be critical from an engineering perspective was more critical from a legal process perspective and that's called zero copy so underpinning these capabilities is the ability to do a zero copy match with your legal and privacy say first-party data and and the partners that you're involved with so that zero copy takes what used to be a month and in some cases we've seen year-long legal process down to weeks and so now this is getting momentum It's well understood and very trusted that you're protecting your data in a way that allows you to get the results that we've talked about. And then looking ahead, you could see those concepts of clean rooms and zero copy powering your agents in innovative and new ways. Right. So I like to talk about this concept of using your data in a very ephemeral way, meaning, hey, maybe I'm going to do a zero copy. Copy match, run a just-in-time personalization, audience optimization, and never move any of that data. Just use it to get the outcomes done while still maintaining that privacy safe and legal and regulatory compliant approach. And then lastly, when you think about the power of all the data that we're putting into this platform, the future we also see is the ability to to start to think about this as a digital twin. And so what does that mean? I've got so much data on a customer 360 basis, whether it's an off platform interaction on a media platform, or it's a visit to a restaurant or a store. When I bring all of that data together in a single place, you can start to make really powerful enterprise optimization prediction. So think about now I know the demand side. inside better than ever before because I'm capturing data that's compliant and within the guidelines of advertising platforms, but I'm also capturing my first party data. When I combine those two together, I now can make better predictions on supply chain. I now can make better predictions on staffing. Just thinking about what this new data provides is incredibly powerful. incredibly exciting and that's really the next frontier of innovation that we see coming from paid media measurement platforms and leveraging clean room capability more broadly. So I want to thank everyone for joining us today. Thank you, Andre and Tim for having such an amazing discussion. Thank you to all our live viewers for tuning in and engaging with us today. I hope you can see how excited we are about these new technologies and the positive impact. impacts they can have on your businesses, on your customer experience, on your media performance, and ultimately on the way you leverage this data to run your enterprise and optimize your enterprise. Overall, Aperture on AWS has helped us transform our client's go-to-market approach by enabling agility and optimization and paid media measurement activation in support of their business objective. I think what's really powerful about Aperture is it brings together There are traditional services with the capabilities that clouds make possible to help accelerate our time to value with our clients. That's the core value proposition that a partnership like Amazon Web Services and Publisys Sapient can help you drive better outcomes.

Relevant Links

• AI-Powered Paid Media Measurement and Optimization with AWS Clean Rooms
• Privacy-first audience collaboration for QSR marketers
• Global quick-service restaurant brands rarely struggle with the idea of local relevance. They struggle with the mechanics of delivering it at scale.
• Quick-service restaurant brands do not need more marketing reports.
• From Paid Media Measurement to Enterprise Optimization in QSR
• Restauration rapide en Europe : relier mesure média, données first-party et création de contenu pour accélérer la croissance locale (Europe)
• México: del gasto en medios a una máquina de crecimiento para restaurantes de servicio rápido (LATAM)
• FAQ (FAQ)
• 12 Things Buyers Should Know About Publicis Sapient’s AWS-Powered QSR Marketing Solutions (LIST)