bid caps gone rogue? full budget, double CPA

budgetburner

I've been running bid cap campaigns for years. The deal is simple: set a cap, and if the system can't find conversions at that price, it throttles spend. Good days spend full budget at or below cap. Bad days spend less, but you're still profitable. That's the whole point.

But around the 8th of May, something broke. Now it's either burning the entire budget at 2-3x my capped CPA - making campaigns unprofitable - or crawling so slowly it can't even hit a single conversion. No happy medium.

Someone in the replies said they're seeing the same: spending $600-$1k/day and averaging 30-50% above cap. Yeah, that's not a bid cap, that's a cost cap at best. And even cost caps are supposed to average out over a week, not spike instantly.

Maybe my understanding of bid caps is outdated? I always thought they were rigid - no spend if it can't hit the goal. But now it's spending like there's no tomorrow at triple the cost. Is this a glitch or did Meta quietly change the rules? Because these campaigns were working fine before.

socialbutterfly

Oh, tell me about it. I've been tearing my hair out with the same issue. My bid cap might as well be a suggestion - it's currently blowing through about 30-50% above what I set, and we're looking at $600 to $1k a day in spend.

Daily budget? Completely ignored.
Cap? Laughed at.
ROI? Crying in the corner.

It's like Meta just decided bid caps are optional now. No warning, no transparency, just "here's your surprise overspend - enjoy!"

liamcopy

Same here.
Cost cap? Broken.
Bid cap? Also broken.

Meta just does whatever it wants now.

analyticsjunkie

Oh you're definitely not alone. the thing that gets me is how bid caps used to have this one redeeming feature - sure, they'd sometimes throttle spend, but at least they'd protect your CPA from going completely off the rails. now Meta's like "here's your full budget, enjoy a 3x CPA, best I can do."

The theory was always that bid caps control auction entry, not the final cost. But when a campaign that used to respect the cap suddenly spends the whole budget while landing miles above historical performance, something's broken. have you checked if CPMs or conversion rates shifted around the same time? or is Meta just throwing you into more expensive auctions now because the algorithm's decided it knows better?

dailygrind

yeah, I've been seeing exactly this across several accounts. if those same bid cap campaigns were profitable before and now they're spending full budget way above target CPA, it's likely a delivery behaviour change rather than a bidding issue. I'd check attribution windows, conversion quality, and audience overlap since early May. Have you noticed any shifts there that might cause Meta to overvalue the traffic it's buying?

metricsmuse

Yeah, Meta's been absolutely cooked the past few weeks. Seen the same thing across a handful of accounts. The bid cap's supposed to be a hard ceiling, but right now it's behaving more like a "suggested" bid - which completely defeats the point of using it over cost cap or lowest cost.

What actually helped me was slashing the daily budget pretty hard and letting it recalibrate. Sometimes the system gets stuck in this weird spend pattern, and a sharp budget cut seems to snap it out of that loop. Also worth checking if any of your competitors ramped up spend recently - auction pressure can make Meta decide your bid cap isn't competitive enough and it just... ignores it, basically.

budgetburner

You running cost caps then? does your average CPA sit above the cap over a 7-day window? because that's exactly how cost caps are meant to behave - they're flexible, they'll lean towards that limit over a week. bid caps are the opposite. They're rigid. they won't spend a single pound if they can't hit the goal. or at least that's the theory. Lately I'm not so sure either one knows what it's doing

socialbutterfly

Oh, tell me about it. 😩 I've been pulling my hair out with cost caps lately too. The whole "7-day learning phase" thing is a joke when the spend just blows past the cap on day 3.

Some things I've noticed in our campaigns:

🔹 Cost caps? More like 'suggestions' at this point 🫠
🔹 Even with a super tight audience and conservative bids, Meta seems to ignore the cap entirely
🔹 No amount of pausing and re-launching fixes it - it's like the algorithm just does whatever it wants

Curious if anyone else has seen a difference between cost caps vs. bid caps? Or is it all the same broken mess right now?

dailygrind

if you're seeing lower quality buyers, almost no engagement signals, and customers behaving differently than before, take that seriously. it suggests Meta might still be finding conversions, but not the kind your business actually wants-especially if you're optimising for a broad purchase event rather than a high-value one.

I've run into this exact pattern at our FinTech startup. We saw a flood of conversions at a seemingly decent CPA, but retention was in the gutter and LTV tanked. The issue was that Meta's bid cap was hitting the volume target but optimising for any purchase, not the ones that actually stick.

Here's what I'd look at next:

Audit your highest-value customers from the last 6-12 months. Pull cohort data via your CRM or analytics tool (we use a mix of BigQuery and a lightweight BI layer). do they share any common traits? things like time-of-day conversion, device type, ad placement, or even which landing page they hit first. If you can isolate that signal, feed it back into Meta as a custom audience or a value optimisation target.
Rethink your conversion event. If you're tracking "Purchase" as the sole primary event, switch to something like "Subscribe" or "First Deposit" that aligns closer to monetisation. In Meta's event manager, set up a custom conversion with a higher value weight attached. You can even pass server-side value parameters via the Conversions API-that's been a game changer for us. screaming Frog won't help here, but Ahrefs for competitor landing page analysis can sometimes reveal how others are defining high-value actions.
Check your attribution window. If you're using a 7-day click, you're likely catching a lot of low-intent traffic that converts on impulse but churns fast. Shorten it to 1-day click + 1-day view and see if the quality shifts. Then re-evaluate your bid cap performance.
Use a holdout test. Turn off the campaign or shift to lowest cost for a week and compare the downstream revenue per customer. That's the only way to know if the bid cap is genuinely broken or if your audience is just degraded.

Meta's machine learning loves volume over value by default. you have to manually inject quality signals-otherwise, you're just optimising for cheap clicks that never pay off.