Mobile app install campaign tanking after 11 days

gmbpro

Ran a Meta campaign for our mobile app - one campaign, one ad set, seven creatives (five images, two videos). Optimising for app installs. We're feeding Meta a bunch of events: install, achieve level, complete tutorial, registration, rate, purchase, start trial - enough signals for it to learn. Daily budget €100, targeting all of Germany, no restrictions.

First two days looked solid: 63 and 56 signups. Then performance fell off a cliff. Yesterday we scraped 18 signups. The frequency data tells the story:

Day 1: 63k impressions, frequency 1.92
Day 2: 6.9k impressions, frequency 1.48
Day 3: 1.2k impressions, frequency 1.22
Day 4: 1.4k impressions, frequency 1.09
Day 5: 12.8k impressions, frequency 1.51
Day 6: 11.1k impressions, frequency 1.26
Day 7: 20.9k impressions, frequency 1.60

Cost per install jumped from _€3 in the best periods to over €5.50 for the worst ads. The weird part: Meta keeps spending on the worst-performing creatives (€5.70 CPI) while the best ad (€2.74 CPI) gets almost no spend. How does that make sense?

Frequency climbing back up to 1.6 after dropping earlier suggests Meta is running out of fresh people to show ads to - but with only €100 daily across a country of 80m+, that seems off. Creative fatigue after 11 days? Possibly, but at this spend level it feels premature.

Anyone else seen this pattern with mobile app install campaigns? Options I'm weighing:

Kill the underperformers and duplicate the winner into a new ad set
Create fresh creatives (but how often do you need to rotate if scaling)
Change optimisation to a downstream event (purchase or trial start) instead of install

From experience, when a campaign flatlines this fast, it's usually an audience layer problem - but I can't figure out why Meta ignores the clear signal from the best ad. Would appreciate a sanity check.

copycat

Stop second-guessing yourself - the algorithm is a mess right now. Start a fresh campaign using your winning ads, keep an eye on the metrics, and if they hold steady, scale it up with budget increases and variations. From my own experience, once a campaign flatlines, it's done. But that doesn't mean those creatives and copies won't strike gold again later.

splittester

I'd love to hear how you're nurturing new creatives into the mix - what's the visual pipeline looking like? And roughly how many fresh concepts are you testing each week?

ranktracker

mate, have you checked whether Meta is actually optimising toward cheap installs instead of quality users who move deeper into your app events?

Honestly, what you're seeing is textbook for fresh app campaigns, especially when you're only optimising around installs during that early learning phase. the first few days look brilliant because Meta finds the easy low-hanging install audiences first. once those get saturated, the algorithm starts expanding into weaker inventory and performance tanks fast. a frequency of 1.5-1.9 after 11 days in Germany at €100/day isn't real audience exhaustion yet - wouldn't panic about fatigue alone. The bigger clue is Meta keeping allocating spend toward ads with worse CPI. that usually means the algorithm is seeing stronger downstream engagement signals inside its black box that aren't visible in Ads Manager yet. an ad with slightly worse CPI can still get more spend if Meta predicts those users are more likely to complete registration, tutorial, purchase or trial down the line.

Personally, I'd avoid massive structural changes right now. wouldn't split into many ad sets either - your data's still consolidating.

Worked with a gaming-style productivity app in Germany spending about €80-150/day. first 3 days gave us nearly 170 installs at roughly €2.40 CPI, then performance suddenly nosedived - daily installs dropped below 25 despite similar CPMs and CTRs. The founder assumed creative fatigue instantly and started swapping ads every few days, which made delivery even shakier. Meta kept prioritising creatives with worse CPI, which looked irrational inside Ads Manager. After digging into downstream events, we discovered the "expensive" creatives were actually producing about 2.7x more registrations and nearly 4x more trial starts than the cheap-install ones. simplified the setup from 9 creatives down to 4 stronger ones, shifted optimisation from installs toward completed registrations once enough event volume built up. Within about 3 weeks CPI settled around €3.10-3.60, daily registrations went from roughly 18 back above 55, and trial start rate improved nearly 46% because Meta finally optimised around better-quality users instead of cheap installs.

When you compare your creatives - are the lower CPI ads also producing the strongest registration and tutorial completion rates, or only the cheapest installs?

tagmanager

Ugh, I love marketing but watching an app campaign burn out like that is painful. The algorithm does hit a wall sometimes - you've given it a solid run, but once frequency creeps past 1.5 in a market like Germany, you're basically paying to show the same creative to the same tired audience. that feeling when you know someone's been hit six times and they're just scrolling past.... it stings.

What's worked for me in similar slumps is pausing the absolute worst performers first - don't let them drag your average down. then throw in some fresh creative, even small tweaks to the hook or CTA. if you've got solid purchase volume coming through, value optimisation can breathe new life into it. The algorithm just needs a different signal to latch onto. you're not doing anything wrong - this is just the nature of the beast

gmbpro

give this a try, but watch out for the typical gotchas. i've seen a lot of setups where the campaign tracking falls apart because the UTM parameters aren't properly passed through to the app SDK.

key things to verify:

The campaign_details event fires correctly in your debug view. If it doesn't, the attribution model has nothing to work with.
Check that your Google Play URL builder (for Android) or campaign tracking URL (for iOS) actually includes the required gclid or gbraid parameters. without those, Google Ads attributions won't link back to the campaign.
If you're using Firebase Dynamic Links, make sure the parameters survive the redirect. Around 30% of the setups i audit have broken links because of a missing fallback URL.

what specific platform are you running this on - iOS, Android, or both? the implementation details differ slightly, and I can point you to the exact dataLayer push you'll need.

splittester

This rings so true - I've hit this exact wall with mobile app campaigns. Meta's algorithm sometimes just doesn't know which metric to chase, so it drifts. Making sure the conversion API is humming in the background is non-negotiable. Without it, that cheaper cost-per-install ad might be optimising for something completely off, like landing page views or even just reach, while the pricier ad is actually dialling into the installs you care about. But honestly? Sometimes Meta just wants you to kill the underperformers and let the winning creative breathe. Also, more creatives - even subtle variations - can give the algorithm enough texture to land on the right signal.

gmbpro

That's a common starting point, but running the same creatives from web to a mobile app campaign usually tanks install volume. I've seen CTR drop 30-40 % because the creative context is wrong-users engaging on mobile web expect a different experience than in-app.

Few things to check:

Are your creatives optimised for in-app placement? Same assets, but different aspect ratios and call-to-action?
Have you set up conversion tracking properly in GA4 for in-app events? If not, the campaign algorithm has no signal to optimise on.
What's your current install cost vs. target? Without segmentation by creative, you can't tell which assets are burning budget.

If you want a clean test, run a small A/B with one new creative designed specifically for mobile app. Track first‑open events as conversion. Happy to dig into the GA4 setup if you share the event schema.

ranktracker

No, that's not quite right. You'll need a properly set up tracking system-someone in a similar niche has it working, and once it's in place, optimisation starts kicking in.

Could you share the full details on the product and the creatives you're running right now? Give me the specifics and I can break down what's been working for our clients, then you can implement the same approach