three days of zero sales - anyone else?

cyberpunk

I've been running ads since 2011, and these last three days are genuinely the worst I've seen. Literally zero sales. Testing new creatives constantly, throwing everything at it, but it feels totally random now.

Two things that are bugging me:

When I've got a winning ad set, do I add new creatives into it or create a fresh ad set? I'm worried about messing up the performer.
I've heard that all ads in a set contribute to the winners - turn off the non-performers and the whole thing collapses. True? How do you handle that?

Someone in the thread mentioned checking EMQ first - account trust decay or pixel tracking failure can cause three-day slumps. I haven't looked at mine yet. What's your EMQ sitting at? Might be worth digging into before panicking.

backlinker

You don't mess with a winning ad set - it's that simple. If it's performing, there's no logic in intervening. Keep your winners isolated, and just clone them into separate campaigns or ad sets rather than risking the algo breaking its stride.

Exactly this. I've seen it dozens of times: an ad spends a decent chunk but the backend metrics look anaemic, while another ad is absolutely crushing it. The platform's attribution window and learning phase can make things look weird. If the ad set is winning on overall ROAS or CPA, leave it be. Poking it only resets the learning phase and tanks performance.

trafficcop

Oh honey, welcome to the Andromeda rollercoaster. Have you checked if the account is actually dead or if Meta's just having one of those spastic exploratory delivery phases where conversion intent vanishes but engagement metrics look peachy?

The real kicker is soft metrics can stay "acceptable" while actual purchase intent nosedives. That's the illusion. I'd hold off ripping everything apart after a few bad days unless the collapse drags on. Meta's been a fickle beast since Andromeda - sometimes accounts bounce back after these chaotic swings.

Worked with a womenswear brand spending around $6k/day. The team got trigger-happy and purged "underperforming" creatives from a winning scaling ad set because only two pieces were driving most buys. Within 48 hours the entire set flatlined - even those winners, which had scaled profitably for weeks. CPA jumped from ~$31 to over $70, conversion volume tanked. Rebuilt by splitting creative testing from scaling and stopped deleting supporting ads aggressively. Took about three weeks, but ROAS clawed back from 0.9 to 2.7, daily purchases stabilised, and scaling got way less volatile because Meta had a healthier exploratory sandbox instead of a hyper-compressed winner structure.

When you say "zero sales" - did CTR and page view metrics hold up, or did the whole funnel crater together? That'll tell you if it's a delivery glitch or a genuine meltdown.

pixelpusher

Everyone says "new adset, always" like it's gospel. Honestly, that's lazy. If a campaign is scaling and the learning phase is solid, duplicating an entire adset creates a whole new learning phase anyway. Better approach: refresh creative inside the existing adset without killing the machine. Signal pooling is real-but turning off underperformers mid-flight? That messes with the algorithm's momentum. My view: don't turn them off unless they've had zero spend for a significant window. replace ads that are genuinely dead weight (sub-10% budget allocation or zero impressions), but only one at a time. let the system stabilise before swapping the next. Half the time people break a winning adset by over-optimising. leave it alone unless it's actually broken.

datanerd

The wildest thing about Facebook right now is watching veteran media buyers - myself included, with over a decade running traffic back when there was only right-hand side ads, feeds didn't even exist - now sitting here asking questions like we just started. The only pattern I've seen hold: the media buyers I'm still in touch with only run lead gen. Zero actual purchases. They might fire a purchase event, but nothing's being sold. So from a 7-day zoom-out, lead gen still mostly holds profitability - just trading 100%+ ROI for 30-40% here and there. Take that however you want.

roichaser

i got hit early too, and everyone and their mum swore my ads were the problem. same old song and dance. turns out it wasn't my setup - the platform was just broken

cyberpunk

I've been in this exact spot - running ads since the early days, then suddenly a three-day wall of zero sales. It's maddening. In my experience, a cliff like that almost always means Meta tweaked the algorithm or attribution broke, not that your creative suddenly died. Don't touch a thing. Panic-editing when the system is already volatile just kicks a stable ad set back into learning phase.

On new creatives: never add them directly to a winning ad set. That's a quick way to trigger a Significant Edit and reset your proven winners. Always spin up a separate sandbox campaign to test. And that theory about non-winning ads "feeding" the winners? Misunderstood. Ad sets collapse when you pause the losers because turning off several at once shocks the system and triggers a significant edit, wiping the delivery model. Best practice is to start with 3-5 ads per set and let the algorithm naturally starve the underperformers - no manual switch flipping.

datanerd

I've been running Meta for B2B lead gen since 2020, and I see this pattern all the time. A three-day zero-conversion dip is almost never creative fatigue - it's account trust decay or pixel/tracking failure. If your EMQ (Event Match Quality) is below 7, Meta can't reliably match conversions to users, so it's optimising blind. The model degrades fast.

Check your Events Manager:

# Pseudocode for sanity check
if emq < 7:
    print("Fix event dedup or CAPI")
elif domain_review_flagged:
    print("Account trust dropped")

Event dedup misconfigured or CAPI not sending properly halves your reported conversions - Meta thinks you're getting 50% of actual ROAS. Account trust can tank if your BM has too many old failed campaigns or domain review flagged a claim. Auction prices you as higher risk.

The thing about 10 ads in one adset is real: Meta weights winners internally. But turning off non-winners while the adset is still profitable can collapse everything. The top two ads rely on losers' audience data to improve targeting - kill the noise and you kill the signal.

One caveat: if all competitors in your niche launched at the same time, you're just saturated. Harder to pinpoint without seeing account metrics.

What's your EMQ sitting at? That'll tell you half the story.

cyberpunk

So just to clarify - when you say never touch the winning ad set, does that mean leaving individual ads inside it alone as well? Even the ones clearly underperforming?

conversionninja

Oh, the absolute gut punch of watching spend crater from a glorious $1k a day to a measly $20 - I swear Meta does this just to haunt us. That EMQ of 6.1 is screaming something, but it's like trying to read a menu in a pitch-black restaurant. Your ads are essentially wandering the digital gutter, ignored by everyone. Is your creative feeling stale or have you toggled anything in the audience settings recently? Because this level of drop feels less like a tweak and more like a full-blown curse.

backlinker

Correct. Let a winner win.

adcraft

Actually, that's not how CBO works in practice. Multiple ad sets can still be useful if you're testing different audiences or creatives - CBO just optimises across them based on performance. The idea that it's "pointless" is a misunderstanding of how the algorithm allocates budget. Done right, it gives you more data, not less.