I’m working on setting up iOS app installation tracking with Firebase and Google Analytics and want to know which marketing channels are driving my app downloads (Facebook ads, Google campaigns, website traffic, social media, etc).
After going through Firebase docs and GA setup guides, I can’t find a straightforward method to do this properly. iOS makes this really challenging compared to Android because there’s no install referrer data available. When users click an ad and go through the App Store, I lose track of where they originally came from.
The main iOS tracking issues I’ve found are:
- Install referrer data doesn’t exist (unlike Android)
- SKAdNetwork in Firebase works only with Google Ads campaigns
- UTM tracking gets lost during App Store redirects
Because of these problems, I’m thinking about recommending a Mobile Measurement Partner (MMP) tool like AppsFlyer to my team. These platforms seem to offer:
- Statistical modeling for attribution
- Connections with Facebook and other ad networks
- Device fingerprinting methods
- Full SKAdNetwork coverage
However, I need to justify this decision and prove that Firebase plus GA can’t handle comprehensive iOS attribution for multiple advertising platforms.
Has anyone dealt with this challenge before? I’m hoping to get insights on:
- Does Firebase with GA actually work for iOS attribution beyond Google Ads?
- Is using an MMP the only realistic solution for accurate cross-platform iOS tracking?
Any experience would be helpful!
Been down this exact rabbit hole last year when we launched our fintech app. You’re right about Firebase + GA being pretty limited for iOS attribution.
We tried making it work with just Firebase for 3 months. Reality is you’ll only get decent attribution data from Google Ads campaigns, and even that’s not perfect. Everything else becomes this black box of “organic” installs that aren’t actually organic.
The UTM parameter issue you mentioned kills it. We ran Facebook and TikTok campaigns and basically couldn’t prove ROI because all those installs showed up as direct or organic in Firebase.
We switched to Adjust (similar to AppsFlyer) and the difference was night and day. Suddenly we could see which creative assets actually drove installs, optimize our Facebook spend, and kill underperforming campaigns.
The statistical modeling these MMPs use isn’t perfect, but it beats flying blind. Plus the device fingerprinting catches a decent chunk of users who don’t go through SKAdNetwork.
Cost is worth considering though. These tools aren’t cheap, especially if you’re just starting out. But if you’re spending serious money on paid acquisition, the attribution data pays for itself pretty quickly.
My recommendation: if you’re running campaigns on multiple platforms beyond Google Ads, get an MMP. Firebase alone won’t cut it for serious iOS attribution.
Firebase and GA work fine for basic iOS tracking, but you’ll hit walls fast with multi-channel campaigns. I spent 6 months trying to make it work for our SaaS app before accepting the limitations. The real problem isn’t just missing referrer data - Firebase treats everything as organic when attribution chains break. We were spending heavily on LinkedIn and Twitter campaigns that looked worthless in Firebase. Almost killed our best performing channels. MMPs handle this better, but there’s a middle ground. We built server-side attribution by storing click IDs in our backend when users hit campaign links, then matching them with Firebase first-open events using device timezone and rough timing correlation. Not perfect but caught about 60% of our lost attribution. This worked well enough to justify MMP costs to leadership with actual data showing the attribution gap. Now we use AppsFlyer but the custom solution helped bridge the gap initially. Bottom line: Firebase alone will underreport your paid acquisition significantly on iOS. Whether you build workarounds or buy an MMP depends on your budget and technical resources, but doing nothing means throwing money at campaigns you can’t measure properly.
Same boat here with our e-commerce app. Firebase and GA only show maybe 30% of your real iOS attribution - makes it impossible to optimize campaigns properly. We wasted months trying to fix it with Firebase dynamic links and custom URL schemes, but Apple’s privacy updates and App Store redirects kept breaking everything. Data was so bad we were basically guessing with our ad budget. What sold our team on switching to an MMP was running both systems for two weeks. Kept Firebase going while testing Branch. Turns out our Instagram campaigns were performing 40% better than Firebase showed, and some channels we thought worked weren’t bringing quality users at all. Yeah, MMPs are more complex to integrate and you’ll need dev time for their SDK. But the attribution accuracy boost is huge, especially for probabilistic matching on iOS 14.5+ where deterministic tracking dies. Firebase just wasn’t built for multi-channel attribution complexity. The MMP cost pays for itself fast when you can actually optimize ad spend with real conversion data instead of flying blind.
Firebase and GA for iOS attribution? It’s like solving a puzzle with half the pieces missing. I’ve been through this with multiple app launches.
What changed everything wasn’t jumping to expensive MMP tools. We built an automated attribution system using Latenode that connects different data sources.
Here’s what we did: Created automated workflows that pull install data from Firebase, grab click data from Facebook Ads API, Google Ads API, and other ad platforms. Then we cross-reference timestamps and device data for probabilistic matches.
Runs every hour, updating our dashboard with way better accuracy than Firebase alone. We’re catching 70% more attributable installs now.
It’s not as fancy as AppsFlyer’s modeling, but it fixed our biggest problems without monthly MMP costs. Plus we can tweak the attribution logic for our specific needs.
Took 2 days to set up initially, but now it runs automatically and feeds clean data back to our analytics. Much cheaper than paying thousands monthly for an MMP when you’re still testing product market fit.
If you want to try this before committing to expensive tools, automation gets you surprisingly far: https://latenode.com