Spending 15k/day on Meta - Need Google UAC Hacks

liamcopy

We're spending over $15k a day on Meta for our iOS app. It's working. but Android? that's a blank canvas.

Our vision was Google. No tracking nightmares. clean data. but we can't crack the code.

right now we're just dumping our Meta winning videos (9:16, 5-18 seconds) into Google App Campaigns. some hit. Some miss. feels random.

it's wrong to treat Google and Meta the same. Google is a black box. more placements. weirder algorithm. feels like throwing darts blindfolded.

i searched for in-depth advice on UAC. Found nothing. just surface-level crap.

Someone said repurposing 9:16 only covers one placement. google needs 16:9 for YouTube pre-roll, 1:1 for Display, 9:16 for Discover. if you only feed portrait short-form, the algorithm can't compete across the full network. it runs where assets fit.

Another person mentioned staging bids. Start with tCPI to build install volume. Then shift to tCPA once you have data. jumping straight to tCPA on a new Android campaign starves the algorithm before it learns.

Someone else suggested Apple Search Ads first - low-hanging fruit. for UAC, try images and text first. Best luck on Play Store and Google Search. no video needed. Then switch to optimize for in-app event when you have volume. And turn off view-through conversion optimisation. otherwise it counts bullshit view-throughs as real conversions.

i'm sharing this because i know others are in the same boat. What's your experience? any real hacks? Happy to give more context if someone has specific tips

ctrjunkie

Honestly, it's a trap to just chuck your best Meta creatives into Google App Campaigns and hope for the best. Repurposing 9:16 winners is fine, but that only covers one placement type. Google's UAC pulls from your full asset pool - YouTube pre-roll (16:9), Display (1:1), and Discover (9:16). If you're only feeding it portrait short-form, the algorithm can't compete across the whole network. It basically just runs wherever your assets physically fit, which is a massive waste of budget.

the other thing people seem to skip: bidding needs staging. Start with tCPI to build install volume, then switch to tCPA once you've got enough conversion data. Going straight to tCPA on a fresh Android campaign starves the algorithm before it's even had a chance to learn anything. I've seen it kill performance way too many times.

Would love to hear what asset mix and bid setup you're currently running - happy to compare notes.

jamesaso

apple Search Ads is probably the lowest hanging fruit you're not touching. minimal creative work, straightforward setup - and if you're dropping that kind of budget on Meta, you can afford a few hundred quid on ASA to test.

For UAC, start with just images and text. skip the video until you've got a baseline. the best placements are Google Play search results and regular search - no video required, and those channels consistently outperform the rest. once you hit a decent install volume, switch the optimisation goal to an in-app event like a sign-up or purchase. one single UAC install campaign is fine, no need to overcomplicate it.

If you do eventually run video assets, go into settings and turn off view-through conversion optimisation. Otherwise it'll count those worthless view-throughs as real conversions 1:1 and the algorithm will optimise towards complete garbage. learned that one the hard way

growthgeek

honestly, most of what gets said about Google App Campaigns is garbage. People obsess over creative variation like it's Meta, then wonder why it flops. google's algorithm doesn't give a damn about your hook-heavy UGC - it shoves your ad into YouTube, Discover, Play Store, and Search simultaneously. that's four completely different user mindsets.

i've seen Meta winners bomb on Google because the assets are too narrow. google needs breadth and signal depth. optimising towards installs only is a rookie move. Feed it stronger in-app events - retention data changes everything on Android once the algorithm understands LTV signals. Also, stop blindly reusing Meta-style creatives. Test separate buckets by intent type.

app Campaigns are still a black box, but the accounts that actually scale give Google more conversion variety and stop trying to micromanage placements. Let the machine do its job, just give it better fuel

wordweaver

Oh honey, you've nailed it. treating Google like Meta is like trying to use a spatula as a hairbrush - technically both exist in a kitchen, but you're gonna have a bad time.

the whole "9:16 vertical is king" brainrot from Meta has people neglecting what Google actually likes. World and square? Underrated gems. Meta is about interrupting a scroll. Google is about catching someone mid-search or mid-YouTube binge. totally different vibe.

Here's what actually moved the needle for us: stop overthinking your creative strategy and just dump everything into UAC - text, images, video, even HTML if you're feeling spicy. Let the algorithm do the heavy lifting. i was sceptical too, but the last few months? night and day. Structure is dead simple: one campaign per country, target CPI, then walk away for at least a week. touching it earlier just confuses the machine.

And for the love of god, use your own first-party data. upload a list of your best iOS users who also have Android devices. Google will find lookalikes that actually convert. That alone fixed our random performance spikes.

Also, peek at asset reporting. Most people ignore it, but it'll tell you which combos are driving installs vs just racking up views. Shocking stuff.

so yeah, ditch the "repurpose Meta creative as main strategy" mindset. test horizontal video. Test static images. test boring text ads. You'll be surprised what works once you stop assuming Google is just a less fun Meta

miaaffiliate

Scaling app campaigns from Meta to Google is a different beast - intent signals just don't translate. Straight porting creative is a waste of budget, you've got to match search behaviour instead.

This is exactly why I started using Leadline. It surfaces actual demand - places where people are complaining about the same problems your app solves. Saves you guessing.

dailygrind

Mate, I've spent a fair chunk of time wrestling with Google App Campaigns across a few high-spend accounts, and that learning phase is the single biggest trap people fall into.

you absolutely need at least 50 in-app events per week to exit learning properly. the most common failure I see is someone running a campaign at, say, £200/day when their cost per event is £15 - that's only 13 events a week. You'll never leave learning, and Google will keep throttling delivery because it doesn't have the data to optimise. The algorithm needs volume to actually start bidding intelligently.

here's the process i follow when setting these up:

Figure out your event cost first - either from historical data or a conservative estimate. Divide your daily budget by that cost. If it doesn't give you at least 7 events per day (50/week), increase the budget until it does. I've had to push some clients to £1,000/day just to get the ball rolling.
Use SkAdNetwork events if you're on iOS - they're far more reliable for attribution than what Google reports in the UI. Cross-check event counts between Google Ads and your MMP (Adjust/Branch) to see if you're genuinely hitting 50/week.
Don't touch the campaign for at least 10 days once it's live. Tinkering with bids or budgets resets the learning clock. Set a reminder in your calendar and walk away.

one tool i rely on is the Google Ads Script for App Campaign Diagnostics - it pulls event frequency and learning status into a spreadsheet so you can spot underperforming campaigns at a glance. combined with Screaming Frog (admittedly more for web, but the same principle applies: you want clean data before you scale), it helps you avoid the "low budget, never exits learning" death spiral.

at 15k/day on Meta you clearly know how to spend, so treat the Google side with the same discipline: volume first, optimisation second

liamcopy

Thanks for that.

View-through attribution? Turned it off straight away. No install campaigns running here. Spending roughly 1k a day spread across a few countries-most of it hitting a US tCPA campaign. Directly targeting in-app purchase. That's the real battleground.

growthgeek

Mate, you're going to hate this, but the classic "just throw your Meta winners into one Google App Campaign" is a recipe for wasted spend. Google's auction is not Meta's. I split creatives based on what actually drives the install intent, not just what worked in Reels.

For Android especially, YouTube placements eat up totally different framing than Search or Play Store. Problem-aware hooks like "sick of X?" land well on YouTube because people are in discovery mode, but those same hooks bomb on Play Store where users want feature validation fast. I keep separate buckets: social proof/testimonials, outcome-driven demos, and pure feature walkthroughs. First three seconds are everything, but the context matters more. If you treat Google like another Meta channel, you'll burn through budget on zero-quality installs.

liamcopy

pairing videos with headlines and descriptions is the right move.

spent months tweaking that exact setup.

Still feels like a black box though.

No idea how Google actually stitches them together. you feed in the assets, hit publish, and just pray the algorithm doesn't serve the wrong combo.

Drives me mad.

keep testing - but don't expect transparency.

growthgeek

Everyone talks about App Campaigns like they're a black box, but the real problem is people dumping assets and praying. Google's mixing is a horror show if you don't lock down the emotional tone across each asset group. I've seen a gritty "fix your shit" video paired with soft lifestyle headlines, and the ML just spits out garbage campaigns that no sane person would ever build. My approach: keep each asset group emotionally coherent. Don't let the algorithm Frankenstein together a message that reads like it was written by a committee of robots. It's not about being conservative - it's about giving the system clear guardrails so it can actually learn. Otherwise your reporting becomes a muddy mess and you're left guessing what worked. Also, treat YouTube creative completely differently from Play Store or Discover. Those placements reward different pacing and hooks. Ignore that and you're burning budget.