← AdTechGod Pod
AdTechGod Pod
The Refresh News: April 7 – TikTok Bidding War, Ad Fraud Drama, Spotify’s SAX Debut & WPP's AI Bet
00:08:15
Loading summary
Transcript2 lines
- [00:00]
A
This episode is brought to you by Zeta Global. Do you know what it takes to transform marketing into a data driven profit center? Are you able to align the C suite around your AI vision and strategy? Zeta Global has the Playbook to help you get started. Download Driving Growth in the AI era today at ZetaCMO AI book again, that's Zeta CMO AI slash book.
- [00:32]
B
@ this rate, it's looking like we'll get a TikTok decision before reputation. Taylor's version and Taylor, if you're listening, that is a dare. I'm Kate with marketecture and this is the Refresh your weekly download on what went down in advertising. Today is Monday, April 7, 2025. This week we're covering new bids for TikTok analytics, latest report Spotify's ad exchange launch and WPP buying Infosum. So let's get into it. Unfortunately we are still talking about TikTok. The saga continues as both Amazon and Applovin made last minute bids for the app. And I told myself I wasn't going to talk about AppLovin this week, but here I am talking about them. Just not for the reason I thought. Applovin throwing their hat in actually makes a lot of sense here for a couple reasons. The first is that the app is said to be aiming to take a major slice of the mobile pie from walled gardens like Facebook, particularly when it comes to commerce. And what better slice of the pie to take than one of the most popular social video apps in the US that also happens to have a major shopping focus. The other reason has to do with app's recent business pivot to a self serve platform, one that is built on the backbone of an algorithm that was trained on user data from in app gaming environments. These environments facilitate instant gratification rewards based mentalities not unlike what you would find within TikTok shop. Conveniently, if TikTok is sold, it's going to need a new algorithm and applovins is said to be pretty powerful, particularly when it comes to driving performance. Now when it comes to Amazon, this deal is less likely because antitrust if they were to buy TikTok though, it surely would only be so that they could plug into TikTok shop and then when you open the app it would just be TikTok shop. Inescapable amounts of TikTok shop the final latest rumor is that the Trump administration is compiling a consortium of billionaires to take ownership of TikTok and then lease the algorithm from China. This solution is pretty much as bland as it gets, but does seem to be favored by Trump, so we'll probably end up with some iteration of this. TikTok will more than likely continue to exist in some form, but its next iteration could lead to a serious displacement of creators and advertisers as a result. Regardless of what happens, this has surely been the weirdest mashup of foreign policy and Internet culture we may ever see. Moving over to who strikes again? They quietly released a new report last Friday giving the Wall Street Journal exclusive coverage. And since this report was 240 pages long, here's the TLDR it focuses on general invalid traffic and industry leading verification vendors like IAS, DoubleVerify, and Human failing to block ads from being served to these bots. Basically, these vendors aren't doing what advertisers are paying them to do. General invalid traffic is bots who are making it known that they're bots. Think web crawlers for Google to index new web content, or large language model crawlers to pull data for training. They're tied to an IP address and have what's called a user agent to help identify them. Atlantic said they identified tens of millions of instances over seven years in which ads served to bots across thousands of websites. But to give this a little bit of context, billions of ad impressions are served across the programmatic ecosystem every single day. I'm not saying that excuses these verification vendors for not catching what seems to be pretty low hanging invalid traffic fruit, but just trying to put things in perspective a little bit. IAS and Verify have given their rebuttal, releasing follow up blogs that condemned the report as misleading and unsubstantiated based on flawed assessment techniques. Doubleverify statement carried some not so subtle I'm getting real sick of your shit Nanigans energy and regardless of what side you're on, we should be able to understand their frustration. To date, these reports have always come with a degree of exclusivity. The accused often find out about the reports when everyone else does, and there's a lack of collaboration with those accused parties. This has often led to an incomplete picture around how this is happening and why. There were a host of factors that were identified as to why these impressions were being missed. It sounds like a good chunk of the error was tied to specific platform failings like user agent not being passed to the verification vendor by the DSP or SSP or the IP address. Another bot identifier could have been inaccurate or outdated. On the flip side, it was also called out that known persistent G bots were still being allowed through as biddable impressions. To further complicate things, it was also speculated some of this traffic may be intentionally allowed through for training purposes. Another myth is that verification vendors often only run bot detection code on a percentage of traffic, and then there's a potential for human error not applying the controls correctly or at all. To their credit, analytics consistently helps to identify and generate conversation around various components of the system that are not working as they should, irrespective of the fact that the system may actually be working as designed. It's it's just that the current design leaves some room for improvement. We've seen a lot of these reports at this point with similar storylines and characters like a Scooby Doo episode or Nancy Drew novel. We know how each one plays out. However, each one has identified slightly different issues in media quality, which illuminates the fact that we've got a lot of distinct but interconnected problems to solve. All in all, I think it's important to acknowledge that multiple things can be true at one time. AdTech has a media quality issue that is deeply complex. Inaction and a lack of accountability means we'll see these issues persist. Third parties like Adalytics are necessary for holding up the mirror and calling out flaws. At the same time, verification vendors are uncovering large, often sophisticated ad fraud schemes. Finally, the system needs a reboot and legacy methodologies need to be overhauled. My prediction is that Adalytics will continue to publish reports as long as the industry continues to squabble over these issues, and the open web can be monetized by marketers and bad actors alike. Next up, Spotify is betting on Programmatic. They just launched their Ad Exchange Spotify Ad Exchange or Saks for short, after announcing plans for it last October. With this release, Spotify wants to make buying, activating and creating ads a breeze. The rollouts bring improvements to buying accessibility, particularly the ability to activate ads across multiple devices and ad formats within their ecosystem, and effectively measuring performance across the entirety of your buy. They also mentioned improved audience targeting capabilities and maybe most intriguing of all, the ability to create AI generated audio within the platform itself. Spotify has gradually opened up more and more programmatic buying opportunities over the years. But if you've ever had to buy and play in media for the non whale clients, you know that budgets were a major limiting factor in being able to offer your clients the same opportunities as the big guys. So naturally, my little programmatic buyer heart is singing. As we continue to move out of the age of exclusivity, we've seen several major players move toward a self serve model recently, including Applovin and Comcast with universal ads. This highlights the increasing focus being placed on small to medium sized businesses as an underserved segment of the market that represents meaningful revenue being left on the table. Finally, WPP just announced that they're acquiring Infosum, one of the leading data collaboration companies for advertisers. This move highlights two things. The first is that data and having access, control and ownership over it is very much in vogue. The second is that WPP is moving away from cookie and or identifier based targeting to federated or cohort based audiences. Brian Lesser, CEO of Group M and former CEO of Infosum, said the goal is to move from ID to AI, meaning a more AI focused predictive model. That statement solidifies their intentions as well as the fact that we continue to see AI being used as shorthand for machine learning or algorithmic learning since ML just never really caught on. Anyways. If anyone can make Fetch happen, it's probably going to be holding companies and their massive influence. I'm actually not knocking this approach. I think you can reach someone with a much greater degree of relevance if you're targeting based on demonstrated behaviors or preferences rather than what their email or phone number is or what a cookie has captured about them. And it's highly likely that much of media buying, including audience targeting, is going to be algorithmically driven in the nearing future. The purchase journey is multidimensional because people are multi dimensional and we should start viewing them as such. Big picture we see an increasing trend toward the ability to own and control your data and avoid moving it between various platforms, which increases its vulnerability and risk. WPP is betting on this trend and its acquisition of Infosum to safeguard its future. That's what we have time for today. And just remember, if Cory Booker can give a full blown speech for 25 hours, you can respond to that client email. Thanks for joining us for the refresh and we'll catch you next week.