You're about to spend on paid acquisition.Make sure the measurement is ready before you do.

Our side takes one week. Dev implementation typically adds 7-10 days depending on your team's availability and release cycle. The configuration work is never the bottleneck. Getting SDK changes into a build is. If you have a release coming up soon, that's usually the best time to start.

Very much Involved. We handle every configuration decision, write the implementation guide, and review their SDK integration directly. What we need from your developers: codebase access for code review, and availability to push the SDK changes in their next release. We don't hand over a doc and disappear, we stay in it until the implementation is clean.

The event taxonomy, tracking plan, and SKAN schema we build are designed to be MMP-agnostic wherever possible. If you migrate from AppsFlyer to Adjust, or to any other provider, the underlying instrumentation stays intact. The MMP-specific configuration needs to be rebuilt, but the hard work is already documented and carries over.

The engagement ends with a validated, documented stack. Not an ongoing retainer. That said, if you want someone to keep reading the data after the infrastructure is in place, that's what the Mobile Growth Analyst service covers. We can talk about that once the foundation is built.

Usually neither is right on its own. They're measuring different windows, and the gap is reconcilable. The most common cause is a window mismatch: Meta counts a 7-day-click, 1-day-view conversion while the MMP is on a different default, so Meta claims credit for conversions that happened weeks later and every other channel looks worse than it is. I line up the attribution windows across Meta, Google, and the MMP so all three measure the same event the same way, then use Meta CAPI with matched event identifiers to remove the double-counting. You end up with one number every platform agrees on, instead of arguing about which dashboard to believe.

Yes, web-to-app is one of the most common gaps I close. When ads point at a website and the app install happens after, click data has to survive the jump from the marketing site to the app domain, or every install looks organic. I set up AppsFlyer SmartScript on the landing page through Google Tag Manager to capture the click IDs, configure cross-domain tracking so those IDs survive the handoff, and wire the postbacks back to Meta and Google so they can optimize toward installs. For HappyShyPeople this took every paid channel from showing no attribution to fully attributed in two days.

Yes, and pre-launch is the cheapest possible time to do it. For a fintech remittance app I rebuilt the SDK on web and app, built the event tracking plan covering KYC, registration, and first transfer, configured SKAN 4.0, and integrated and validated all four planned channels, Meta, Google, X, and TikTok, so each was confirmed passing data correctly before the first campaign went live. The whole launch-ready stack took two weeks. Fixing measurement before spend goes out is far cheaper than discovering it was broken three months into a budget.

Sequencing. A surprising number of stacks fire the ATT prompt before the attribution SDK has finished initializing, or only initialize the SDK after a user consents, which means anyone who declines is never counted at all. At Bridebook that produced a 100% consent rate that looked great and was actually hiding every non-consenting user; the real rate was 26%. I check where the prompt fires relative to SDK init, fix the sequence so every user is measured from install regardless of their ATT choice, and only then rely on the consent numbers.

Scope varies depending on your current stack, which MMP you're on, and how much needs to be built versus configured. Email us your requirements and I will share the detail proposal with scope and timelines