We started by tearing the acquisition funnel apart to find where paid installs were leaking before they ever became orders. From there we built a content engine around creators ordering, unboxing, and eating in real neighborhoods — the kind of format-native food content that makes people hungry mid-scroll — with a first-order incentive baked into every hook. On paid, we optimized daily against cost-per-first-order across Meta and TikTok, cutting anything that drove installs but not orders. Then we layered in a retention loop to turn that first meal into a habit. Every play was judged against one number: repeat orders. Downloads that never came back didn't count.