Freddy Clapson is the Managing Partner, Global Programmatic & Innovation Lead, team x (the integrated creative, network, and media agency, built for Mercedes-Benz).
KEY TAKEAWAYS:
Jonathan Elliott: Tell me about your work, team x, and what you're doing for Mercedes-Benz.
Freddy Clapson: team x is the integrated creative, network and media agency built for Mercedes-Benz. In most cases, clients have separate creative and media agencies, but we combine both to provide end-to-end solutions. team x handles everything from initial creative ideas and storyboarding through to translation and asset production through to campaign planning, buying, optimization and reporting.
My team oversees the global strategy and best practices for programmatic use across our 39 markets. While our local teams plan and buy media based on the needs of their local Mercedes-Benz branches, we support them with things like maturity benchmarking, luxury strategy (identifying luxury users and the best websites, partners, and publishers to reach them), and emerging channels. This includes optimising connected TV, audio, and digital out-of-home for a seamless omni-channel experience.
We also do stuff like mixed reality and dynamic experiences. Given my role in innovation, whenever something interesting comes up, my boss says, "Freddy will handle that." And I love it.
JE: You recently said that advertising has been allowed to become "creepy because we've not educated people for the last 50 years." Can you tell me a bit more about what you meant by that?
FC: In any technology-intensive industry, there are two main components or categories: what the end users see and the machinery operating under the hood. Take airplanes, for example. Passengers see the big engines but have little idea of the extensive engineering and the thousands of people working behind the scenes to ensure these planes are safe and ready to fly.
It's the same with ads. People see ads on websites and think it's just a simple handshake deal that places them there. However, they might not realise that four of the five largest companies in the world are primarily funded by advertising technology that makes these ads work. Google, Facebook (now Meta), and, even increasingly, Amazon are driven by advertising revenue. These companies have built vast advertising and data empires without the required education going to the end users on how their data is being used for targeting. Google’s announcement on 3rd party cookies is even being referenced by their CEO in earnings calls, with their stance on user privacy leading the way in how they are talking to their shareholders on revenue projections.
This tech has evolved to the point where personalised ads are so accurate that people sometimes feel like their devices are listening to them. That's what I meant by creepy. You often hear people say things like, "I talked about this product, and now I see an ad for it. They must be listening to me!". So, there's a disconnect between what people know about advertising and how it works in their lives. In reality, these people have likely been served thousands of ads today but remember the one related to something they recently talked about because it's at the top of their minds. As people become more educated about what happens with their online data, they've started to find it unsettling or even creepy. But the industry has been one step behind in educating people on how their data is used and how it’s connected to advertising.
This shift in user’s awareness is a major contributing factor to the deprecation of cookies. The ability to track a person under a single identifier is diminishing, thus making it harder to do. Now, companies have to rely on individual signals or alternative IDs. Ironically, the move towards a post-privacy world means some new identifiers, like hashed email addresses, are far more personal than the anonymous cookie IDs they replace.
But this is all just a part of the broader trend. The technology used in advertising has become increasingly sophisticated in its ability to predict user behaviour. This intersection of advertising and behavioural science has led people to believe they are being more widely tracked due to its accuracy. Which is pretty much your working definition of 'creepy'. If the crowd gets so big, it turns into a singular entity, and then if people start thinking collectively that they're being tracked or followed, it's terrible news for the industry. As this technology develops, it poses new challenges and, inevitably, discussions around privacy and data use. This perception, whether accurate or not, has significant implications for trust in digital advertising.
JE: Let’s talk about Google’s decision to not completely deprecate 3rd party cookies at all. What impact do you think this has on the industry?
FC: Whilst Google’s decision to not completely deprecate 3rd party cookies is shocking to some, its short-term impact is minimal. The rest of the industry has already moved to a post-cookie world, with many browsers and tech partners already moving beyond it. Whatever Google decides to release with this pivot to “elevate user choice”, the third-party cookie will not return to its previous place of ubiquity. Thankfully the hard work that the rest of the industry has done to move forward is not in vain. The longer-term impact returns the same: that we have a more fragmented digital ecosystem than before which will make measurement and targeting more important. Alternatives are being sought after to make previously seen granularity in targeting fit that standards that users have for their privacy. However there is a long journey here if one wants to be a true, all-encompassing replacement for the 3rd party cookie.
JE: What do you think are the most promising strategies being adopted by businesses to adapt effectively to the post-cookie era?
FC: At team x, we're seeing quite a range of responses - from replicating cookies as closely as possible by replacing them with new identifiers, and then to completely shifting back to contextual buys and inferring data from publishers.
Looking at some of the alternative IDs available, we have options like ID5, EUID/UID2 from The Trade Desk, Google's Privacy Sandbox, and Lotame's Panorama, which allows for the interoperability of different alternative IDs due to their typically low match rates.
It's fascinating because some people are aiming to create a "cookie 2.0," focusing on tracking user behaviour to drive lower funnel activities like e-commerce and performance-based measurement.
Others are shifting back to contextual buys, where decisions are based on the context that users are reading from rather than the user data. For instance, advertisers might opt to buy space in Conde Nast during London Fashion Week to align with the event's context. This approach is contextual-based and infers audience data. Similarly, there's moment-based targeting, where companies buy ad space around significant events, such as The Euros or The Olympics, to associate their brand with these moments, replicating event sponsorships.
What seems to be enduring is the use of contextual and historical data to create audiences. The creation of look-alike audiences is a nuanced but critical part of the transition away from cookies. Previously, we expanded audiences built from first-party cookie data by using third-party cookie data to find users with similar online behaviour. Without third-party cookies, the challenge is finding new ways to produce equally powerful audiences.
And some of our partners are using contextual data, like 'device + URL' or 'device + category history,' to create look-alike audiences. For example, if certain device behaviours and category interactions lead to a purchase, we replicate that pattern to build our own look-alike audiences. This method is more sustainable compared to relying on various alternative IDs with limited reach.
If multiple alternative IDs are each capturing only 10-12% of the total audience, none of them can achieve the scale needed for effectiveness. Instead, if targeting based on other signals, like a combination of 'iPhone 15 + tech + auto' gets higher conversion rates, that's way more effective.
JE: What roles do emerging channels like digital out of home, Connected TV and audio play in the context of cookieless planning and buying?
FC: Emerging channels like digital out-of-home (DOOH), connected TV (CTV), and audio are inherently cookieless as they aren’t browser based. But you have to ask yourself what third-party cookie deprecation actually affects. We’ve seen two main outcomes from third party cookie deprecation: it reduces the scale of retargeting strategies due to consent layers placed on websites and removes our ability to target traditional upper and middle funnel audiences built from third party cookies. This predominantly impacts display and online video campaigns sitting outside of walled gardens.
The rest of programmatic activity isn’t impacted too much. Audio is largely served through apps like Spotify, which are cookieless. Connected TVs don't have browsers and digital out-of-home advertising consists of screens placed in physical locations and does not rely on cookies. The deprecation of third-party cookies mainly impacts a proportion of available inventory opportunities (upper and middle-funnel display and standard video ads), however constitute most of the inventory availability and therefore hugely impact the industry.
While this is significant, the emerging channels are cookieless by nature, offering a richer, more integrated audience experience.
The challenge lies in connecting these disparate channels. For instance, how do you tie Spotify audiences, which use single sign-on, with digital out-of-home billboards that have no direct data connections? Maybe mobile tracking data can help bridge this gap.
For us, this is great. If we identify planning strategies that work across various data sources, these emerging channels become more viable and impactful than simply relying on standard display and video ads using third-party cookies.
Basically, as long as plans adapt to a cookieless environment and focus on the right metrics and ask "what are we really looking at?" rather than "what can Google give us in DV360?", especially considering their recent announcements, we should be able to plan any emerging channel.
If your planning approach gives you that flexibility, emerging channels become far more prevalent and provide you the space to be impactful with your audience rather than just a generic display with open web video. This shift in the landscape gives us the chance to be more creative in our planning and execution rather than go with old faithful that we’ve seen previously.
JE: What do you think is significant about Google's launch of open-source MMM Meridian?
FC: Thanks to third-party cookies, we could track ad performance across multiple channels and de-duplicate conversions. For example, Meta might claim they drove 60% of your total sales, while Google claims 70%. Obviously, these numbers don't add up.
With the decline of third-party cookies, de-duplication has become nearly impossible and so we're seeing a resurgence of econometrics and marketing mix modelling (MMM). Interestingly, major tech companies are now offering their own MMM services: Google has Meridian, Meta has Robyn, and it's likely Microsoft will do something now that they've got Xandr and LinkedIn and everything else.
Owning the econometrics methodology and the source data behind it gives you control over your performance scorecard. This allows you to improve your performance based on your own model. Although these models are open source, if it's your code, any changes have a history, similar to the concept of "marking your own homework." This is why we have brand safety providers—those selling video or display inventory might relax their brand safety settings to maximise sales, so you need counterbalances.
If a platform is assessing its own performance metrics, it significantly influences the media plan. Meta might suggest one method for measuring conversions and revenue, while Google suggests another. The truth most likely lies closer to a neutral econometric model, with differences highlighting what each company deems important to enhance their appearance.
This becomes a new battleground of like "where is performance better?", and everyone will obviously say that their product is a little bit bigger, they'll be better than everyone else's. So yeah, it'll be interesting to see if you take these source codes from the Meridian model, and then the Meta model, plus any other models that are available, run the same data through it and see exactly where these things change.
JE: How do you think the role of publishers will evolve in response to the changes in advertising methodologies, and which ones will win?
FC: Publishers will become more important because that single identifier that we have with third-party cookies will diminish, and how they collect data, or what their audiences look like, will become far more important. Those who don't understand how their audiences work will lose value. If you understand more about your topic, you are seen to be more valuable. And those that don't understand what their users look like or behave will end up seeing their inventory being of lesser value because advertisers will not be able to use it for as many things. Or they may struggle to integrate it into their overall planning strategies.
