That May Meta Outage Wrecked Our Campaigns

communityfirst

Nothing tests your resilience like watching a perfectly calibrated ad account implode overnight. We run a mobile app, started Meta ads in March, and by late April we'd cracked it - winning creatives, roughly $500 daily spend, ROAS nearly 2x, and CPI hovering around $2. Everything was humming beautifully for about ten days.

Then the May 5th outage hit. Next morning, CPI had jumped to nearly $10 - a fivefold increase. that week we scrambled: launched new campaigns, added fresh creatives, tried everything. Nothing stuck. Dropped spend to $100 a day hoping to ride it out, but no luck. Now we're back at $1,000 daily with wildly volatile results - some days $3.50 CPI, the next day double that. Our winning creatives are still running, plus new ones, yet we can't break back below $2.

Spun up a brand new ABO (not duplicated from another) and saw $2.20 CPI within five hours... then two days later it was back above $4. Honestly, it feels like the account lost its cheap-inventory pocket. Anyone else been through this? Would love to hear what worked - or if patience is the only real fix.

analyticsjunkie

oh, so you're saying Meta just forgot how to spend money properly after that outage? shocking. 🙄

Honestly, this tracks. The creatives might be fine, but Meta's delivery algorithm has been having a full-on meltdown. the fact those new ABOs hit your old CPI for a hot second before floating away again tells me the system can still find the right people - it just panic-expands the moment you give it a bit more budget. Mobile app auctions have been an absolute circus lately, even on accounts with clean history.

are your retention numbers still holding up, or did those take a nosedive too? That's usually the canary in the coalmine for me.

digitalnomad

big L indeed 💀 Wish they'd sort out their algo. It's been a nightmare since that outage.

roichaser

Honestly, I wouldn't waste time rebuilding campaigns right now. Everyone panics after an outage and starts nuking accounts. That jump from $2 CPI to $10 overnight screams 'you lost your cheap-inventory pocket', not 'your creatives suddenly turned to garbage'.

Stop guessing and diagnose properly. Check if it's CPM, CTR, or CVR that went off a cliff. If CPM blew up but CTR and CVR held, that's an auction quality problem-Meta likely routed you to a different, pricier audience pool. If CTR tanked, either fatigue kicked in or the algo dumped you into a worse segment. If install rate died despite steady CTR, look at the app store page, your tracking, geo mix, placement mix, or OS split-not the creatives.

At $500/day, running a five-campaign mess is just asking for budget waste. Strip it down: one broad CBO with your proven 3-5 best creatives, one tiny creative test campaign. That's it. And actually compare pre-May 5 to now by placement and country-that often reveals the real issue faster than any 're-optimisation' ritual.

ppcnerd

Exactly this. when the algorithm behaves irrationally, you hit a wall. You can't tell if the right move is to optimise harder or to simply stop touching anything. standing still feels as dangerous as reacting

liamcopy

Honestly, i think you're chasing the wrong culprit.

Your creatives proved they could hold a $2 CPI before the May outage. That tells me the product and the ads were fine. the problem isn't the creative - it's the algorithm losing its damn mind after the outage.

I've seen this exact pattern with app advertisers.

One client - subscription app, $700-1,200/day spend - saw CPI go from $1.80-2.40 to $8-11 overnight after early May. the team panicked. new ABOs every day. bid caps swapped like socks. creatives refreshed weekly. targeting rebuilt on a whim.

result? Wild swings. one day $2.90 CPI, the next $9. Same CTR. same retention.

the fix? stop. Consolidate. We killed 7 fragmented campaigns down to 2 stable ones. Paused the duplication. optimised for registrations and trial starts instead of just raw installs.

Within three weeks, CPI settled back to $2.40-3.10. and retention quality jumped nearly 28%. because Meta stopped chasing cheap, low-intent installs.

So my real question: are you only optimising for installs? or feeding Meta deeper events - registrations, purchases, subscriptions? That's where the stability lives.

Stop restructuring. Let the algorithm breathe

communityfirst

We've seen a similar pattern on our side. That first fortnight was brutal-our conversion rate dropped by nearly half. We're hovering close to where we were before the outage now, but still a hair below. It's a slow grind back, but the team's resilience has been incredible.

dusting

The placement landscape has shifted in ways that feel structural, not just temporary. I've been tracking the same trend - stories are now costing nearly five times per purchase compared to feed or reels. It's like the platform's traffic signals have been rewired and stories lost their green light. Worth checking if your budget is draining on a placement that no longer behaves the way it used to.

communityfirst

Absolutely. That outage was a stark reminder of how fragile relying on a single platform can be. It shook a lot of teams-we lost a full day of organic traffic and had to scramble. But it also pushed us to diversify our visual strategy and lean harder into Pinterest. sometimes the biggest disruptions teach the most valuable lessons about resilience

analyticsjunkie

Meta's still not over that May 5th mess, and honestly, it feels like they broke something they can't be arsed to fix. The fact that fresh ABOs still get a cheap CPI honeymoon before the whole thing drifts upwards screams delivery or traffic quality tweak, not us suddenly forgetting how to make good ads.

Have you clocked any weird shifts in country mix or placements since then? my campaigns have been serving traffic like they're on a budget holiday in cheaper regions, and device distribution looks like someone swapped mobile for desktop bots halfway through. curious if that's just my account being cursed.

communityfirst

Would love to message about turning this outage into a team learning moment. Always something to take forward - let's connect!