Understanding Ad Tech Transparency

AppLovin's ad spend quadrupled since Axon 2 launched. 🤯 AppLovin's CEO Adam Foroughi & CTO Basil Shikin recently shared detailed posts explaining their tech and data practices. Key takeaways for marketers & advertisers below. 👇 🧠 It's About the AI, Not Hidden Data: Axon Engine: They emphasize their AI (Axon) is built from scratch, using standard inputs like MAX loss notifications, advertiser data, and ad engagement. Growth comes from model sophistication and a reinforcement learning loop, not a secret data trove. Growth Metrics: Since Axon 2 (2Q23), ad spend quadrupled, unlocking ~$10B annual run rate for gaming clients. Their newer web advertising product hit a $1B run rate in months. Mission Focus: The core goal highlighted is driving incremental revenue for advertisers. 📊 Transparency on Data & Attribution: iOS/ATT: They state compliance with Apple's ATT, using general signals and model learning when IDFA isn't available, without fingerprinting. Web Pixels: The CTO detailed how advertiser tools (like Elevar, Intelligems) might append extra IDs to AppLovin's pixel, but AppLovin purges data they don't request or expect. They point to their dev docs for clarity on requested data. No Data Brokering: Explicitly stated they don't buy/sell data from brokers. Attribution: Rely on MMPs (like AppsFlyer, Adjust) for app attribution. For web (still early days), they use an internal system with 1st-party cookies/transaction IDs, noting clients ultimately rely on their own attribution tools for decisions. 💡 The Narrative: AppLovin positions its success as a result of world-class tech built by a lean, focused team, operating within current privacy frameworks. They're directly addressing how their systems work to build trust with partners. Definitely an interesting read for anyone in the ad tech space trying to understand the mechanics behind major platforms.

What happens when certain attribution methods inflate conversion rates, giving an illusion of better performance? 🚨 Some platforms rely heavily on IP-based attribution, which can overstate conversions by 5–15x compared to cookie or device-based tracking. IP attribution matching is okay for specific scenarios. Eg. CTV to cookie/ifa (mobile device) but for other scenarios like we found in a recent competitor report (mobile ifa to cookie) it makes zero sense. An example would be the iP of a cellular phone is the cell tower when it's not attached to wifi. Why does this matter? 🔹 Lack of Precision – IPs are shared across multiple users, leading to misattribution. 🔹 Inflated Performance Metrics – Clients may unknowingly compare inflated numbers to real results. 🔹 Opaque Reporting – Without clear transparency, advertisers might not realize their "success" is built on unreliable data. At Simpli.fi, we take a different approach. We prioritize deterministic tracking (cookies, device IDs) to ensure real, verifiable conversions—because integrity in measurement drives real business outcomes. Ask: How are my conversions actually being tracked? The answer makes all the difference. #AdTech #Attribution #DigitalMarketing #Transparency

I can’t recall who first said it, but I remember toward the end of 2016 when I first heard the phrase “tech tax.” It was an unflattering term for the hidden costs buried in the programmatic supply chain—costs that no one could fully quantify or attribute. So it was no surprise when the ANA’s 2017 "Seeing Through the Financial Fog" report exposed that 60% of media budgets never make it publishers. The data point igniting a push from buyers and sellers for real adtech fee transparency. Here we are almost a decade later and it feels like Déjà vu all over again. Hearing the calls from across the industry, my colleagues and I at TRUSTX took an unorthodox step to address the issue. We flipped a switch and started bidding gross. A $5 DSP bid through TRUSTX bought $5 of premium inventory. Period. Why? Because transparency shouldn’t be optional. 1 - Eliminating the waterfall of gross-to-net fees makes it clear exactly what you’re paying and to whom. 2 - Bidding gross lets buyers and sellers validate honesty by comparing the input (in our case, DSP bids) to the output (publisher logs). 3 - It's our view that supply-side tech should be funded by publishers. SSPs that extract a fee *before* buying the media means that advertisers are actually paying. I get why its easier to load fees into the supply-chain. But the complexity makes it really hard to know what you're being charged. As my good friend, Bob Liodice has said many times (quoting Dave Berkus), "where there's mystery, there's margin." Indeed.