

The first criminal case of streaming fraud where a North Carolina musician who used AI to make songs, then streamed them billions of times himself making $8 million
casey hamlyn
2.1K posts



The first criminal case of streaming fraud where a North Carolina musician who used AI to make songs, then streamed them billions of times himself making $8 million

Most people think $APP is just a heated ad tech stock riding a gaming cycle but the reality is much more interesting. I just analyzed this interview with their Chief Product Officer Giovanni Ge and the signal to noise ratio is wild. The market is completely missing three structural realities that make this company an anomaly. 👇 First is the headcount leverage. They are running a hundred billion dollar plus enterprise value engine with less than 500 core engineers. This lean culture allows them to iterate fast without bloat and bureaucracy. Second is the technical moat regarding time horizons. While the industry largely optimizes for seven day attribution windows AppLovin has successfully deployed a 28 day model. This sounds like a minor detail but it is actually the entire ballgame for ecom because it allows them to identify high value whales that shorter models miss which explains why their non-gaming revenue is exploding. Third and most importantly is the roadmap nobody is pricing in. They are explicitly building their own social media platform. They know their biggest weakness is a lack of owned inventory so they are building a walled garden to close the loop. They sold their gaming studios not because they failed but because their AI proved it didn't need first party data to work which was a massive flex that the shorts misinterpreted as a weakness. "We aim to build a completely new next-generation social media platform. This platform will bring valuable organic traffic to AppLovin." They are building a sovereign AI infrastructure that operates with better unit economics than almost anything else in the market. The interview is not in english btw 🤣: youtube.com/watch?v=18qSqz…





Retweets = votes. I appreciate YINZ in advance 🖤💛 #HereWeGo #ProBowlVote Ben Skowronek #ProBowlVote @BennySkow





true special teams grit

Why is Incrementality Testing needed? Platform reported revenue can be WAY off. The spread of results we've seen at @CommnThreadCo shows how wide a difference we see from that an Incrementality test reports vs what the platform reports. Incrementality % in the chart = Incrementality ROAS / Platform ROAS



⚡Video Action campaigns will be migrating to Demand Gen campaigns next year. What to know about Demand Gen campaigns: - They serve image and video assets on YouTube, Gmail, Discover, and will soon have the option to serve on Google Video Partners inventory. - Demand Gen campaigns have access to Lookalike segments, creative A/B experiments and AI-powered tools such as text prompt image generation. Results we’ve seen so far: - On average, advertisers who run video ads with Demand Gen saw similar conversions to their Video Action campaigns (VAC) at a similar cost per action. - And those who uploaded video and image assets to Demand Gen saw 20% more conversions at the same cost per action than those who uploaded video assets only. Next steps for Video Action campaign users: - We recommend experimenting with Demand Gen well ahead of the auto-migration which will start in Q2 2025. - You can create new Demand Gen campaigns from scratch now. - Additionally, a new copy-and-paste tool will be available in the coming months that will duplicate your VAC settings into new Demand Gen campaigns. - In March 2025, you’ll no longer be able to create new VAC campaigns. See the blog post for more details: blog.google/products/ads-c…