Recent content by davamosk

I never found good performance with playables in a non‑gaming app using AppLovin. In fact, I haven’t had any good performance with AppLovin recently. Moloco is now my first option for DSPs.

The structure of your campaigns depends a lot on your budget capacity. The more granular you want to go, the more money per day you will need as you need a minimum budget per day to exit the learning phase and feed the ML with enough signals. And yes your deduction is correct. Meta’s total...

Sorry for the late reply. I was in New York! I see +20% underattribution when using AEM as well. You will clearly see that by looking at App Store Connect in the app, referred events are coming from Meta and Instagram. If the brand has no presence and it's 100% relying on paid, it's very easy...

cool update: I have recently tested ICM+ODM (Google granted a coupon) and the results were horrible, with Google making up +350% conversions compared what we were seeing in ASC hahaha

Meta’s AEM data reaches MMPs at install level but it’s Meta-attributed, not IDFA-deterministic (that's what i meant). For Google UAC iOS there’s no equivalent: non-ATT users are only available aggregated (SKAN/ODM/modeled conversions via the Google Ads API), not user-level in the MMP.

no, you just play with probabilistic (same as Meta)

I am not running applovin actively - very bad performance over the last 2-3 months across different apps and webs. Moloco is lately bringing much better performance for me (also with different products). Regarding your question, I would just run parallel campaigns if you can afford that and...

I'm more a fan of decreasing low‑performing ones than boosting the good ones, but I lack a fixed system because my value rules focus on placements rather than age groups. Combining both has never worked for me because I end up paying too much in CPM due to excessive filtering.

1. Because these unified views are relying on SKAN and nobody uses SKAN in 2026 2. Because MMPs are just mirroring what SRNs send them via probabilistic attribution 3. Because iOS underattributes by default in app campaigns, and the unified view doesn't consider any part of the organic in mind...

On paper, you can work with any MMP if implementation is executed properly. On practice, there's a huge difference between working with appsflyer vs Branch for example. UI & UX are so much different between them and than can speed up or slow you on your reporting, partern setups,etc. Regarding...

I hope he publishes about that! do you have any reference?

these guys were doing OEM advertising (pre-install campaigns) - CPIs were super low but retention and activation super poor the last time I ran this

it keeps being poor from what I know. I haven't run it in a while but i would only go for something like that as an alternative channel to test,

i dont know any specific courses about that but I'm sure you can hire any UA freelancer and have some consulting calls (i'm not available)

I have tested incremental attribution and havent seen any incremental uplift or better performance at all.