YouTube CPC hit $51 - anyone else seen this?

localpack

been running paid ads for years - Facebook, Google, you name it. but this one threw me.

Got an old Google Ads account (created 2011) that used to spend five figures on YouTube ads without issue. after about 13-14 months of inactivity, campaigns started acting weird.

First they'd sit dead for nearly a week with almost no spend. Then suddenly they'd kick in - but with absolutely insane CPCs.

Latest example: one Maximise Conversions campaign spent $102.86 for just two clicks. that's $51.43 per click. on YouTube in-stream traffic.

historically this account was around $0.80 CPC. Anything above $1.20 was already unusual. this is the second time it's happened after a long stall.

No suspensions, no policy violations, campaigns eligible, only targeting YouTube in-stream, still showing millions of available impressions.

so what's going on?

From my experience with Facebook, this screams smart bidding gone haywire after dormancy. with no recent conversion data, the algorithm is basically guessing - and it's happy to overpay to find out what works. That's why the CPCs go through the roof.

Someone in the thread pointed out that with spend that low, analysing CPC is almost meaningless - one click skews everything. fair point. But when it's $51 instead of $0.80, it's still a red flag.

Another guy mentioned that Maximise Conversions just bids whatever it wants when there's no data. it has no idea what converts after 90 days of pause.

best move? Switch to manual CPC or set a strict max CPC cap until the model stabilises. Let it learn without burning cash.

anyone else seen YouTube in-stream CPCs spike like this after a dormant period? would love to hear what fixed it

ranktracker

Max Conversions is basically a blank cheque when it's been sat idle for 90 days. No historical data means the algorithm has zero clue what a reasonable CPA looks like, so it'll happily blow your budget just to snag that first conversion and start learning. If you've just unpaused it, you're paying the "we don't know what we're doing yet" premium.

Step-by-step fix: run a manual CPC or Target CPA campaign for a couple of weeks to feed it some conversion data, then switch back to Max Conversions with a bid cap. Should stop the $51 clicks cold.

budgetburner

Classic smart bidding meltdown. Seen it a dozen times-dormant account comes back, the model panics like it's never seen a conversion, and suddenly it'll burn through budget faster than a drunk intern with a corporate card. No conversion data to learn from? It'll just guess stupidly high bids. Swap to manual CPC or slap a hard cap on it until the model gets its head out of its arse. It'll stabilise eventually, but you've got to babysit it first.

metricsmuse

Honestly, Max Conversions can get pretty wild when there's hardly any data to work with - especially if the account's been sitting dormant for a while. Two clicks is basically nothing, so one expensive placement can completely mess with the averages.

I'd say test manual CPC or Max Clicks for a couple of days, just to sanity-check what inventory you're actually buying into. Also worth checking if Google quietly pushed you into premium placements or some niche audience combo without you noticing. That whole "dead for a week then suddenly burns budget" pattern feels like the algorithm's still in its learning phase to me.

bouncekiller

Oh honey, bless your heart if you're obsessing over a $51 click from just two clicks and barely any spend. That's not a problem, that's a rounding error. Running to forums over this is like complaining the toast is burnt when you haven't even plugged the toaster in.

Sample size of two clicks? That's statistically irrelevant. You could sneeze and get a better data set. And Max Conv? That thing needs time and conversions to learn - it's not a magic wand that fixes everything in five minutes.

Also, have you checked where those impressions landed? If half your views are on connected TVs, people aren't exactly racing to click through from their sofa. They're watching your ad while deciding between Deliveroo and a frozen pizza.

And why the hell are you running in-stream ads and then panicking about CPC? If conversions are your goal, focus on CPA and ROAS. You're using a hammer and complaining it can't screw in a lightbulb. Switch tactics - Performance Max would probably do a better job for your objective. But hey, what do I know? I just manage e-commerce campaigns for a living.

marketingmule

Fifty-one dollars a click? Over in beauty and fashion we'd be thrilled to see that for non-branded terms. What kind of product or service are you advertising on YouTube to get those numbers?

24fire

That's the classic "empty room" problem with Max Conversions. The algorithm has no historical signal to work from, so it swings wildly trying to find a conversion anywhere - and it'll happily burn budget chasing ghosts. We've seen the same pattern across client accounts. industry average CPC on YouTube hangs around $0.50-$0.60 depending on vertical, but that's only when the model has actual data to optimise against.

Think of it like trying to navigate a city with no map. Until the account builds a few landmarks (conversions), the bidding system is just guessing at street corners. Best move is to step back to Maximize Clicks or manual CPV until you've got at least 30-50 conversions in the window. Then, and only then, graduate back to Max Conversions. let the account crawl before it tries to sprint.

trafficcop

oh honey, $51 per click and you're trying to draw conclusions? that's like me bragging about a 100% open rate on a webinar invite I sent to my mum and my cat. technically impressive, practically meaningless. Stop staring at the numbers and start spending enough to actually learn something.

localpack

That $51 CPC is absolutely wild, but honestly? I've seen Google Ads do some seriously sketchy stuff with dormant accounts. Here's what I'd bet the issue is:

The "relearning" phase on Smart Bidding is basically a hungry algorithm that's been starved of data for a week. When your campaign finally gets a click, the algo panics and cranks up the bid to "prove" itself to the auction. I've seen this exact pattern on Display and Discovery campaigns - dead for 5-7 days, then suddenly spending like a drunk sailor at 10x normal CPC.

To answer your points:

Smart Bidding going haywire? Yes - but rarely this extreme. On a well-fed account, CPC might spike 2-3x after a pause. $51 means the algo thinks it's bidding against zero competition or chasing some impossible target.
YouTube vs Demand Gen? I've seen it mostly on standard in-stream and TrueView for Action. Demand Gen seems calmer because it's newer and the algo hasn't had time to go fully schizo.
Did CPC normalise on its own? Usually yes, if you let it run for another 2-3 days without touching anything. The algo eventually realises it's paying stupid money and pulls back. But $51 is a red flag - might not correct without a manual reset.

My theory: it's a combo of account trust decay after dormancy + Smart Bidding overreacting to zero recent conversions. Google's system treats a cold account like a brand new learner, and YouTube's funnel is notoriously bad at feeding conversion signals back quickly.

The pattern consistency you describe (7 days dead, then sudden absurd spend) screams "learning phase gone rogue." I'd kill the campaign, duplicate it with manual CPC for a week, then switch back to Smart Bidding once you have 10+ conversions. Saves you £50-per-click headaches.

What's the conversion goal - lead gen, sales, or something soft like views? That decides whether you can even tolerate the learning cost.

datanerd

I've been beating this drum for years. $51 per click isn't just a bad bid - it's a symptom of an unregulated black box that passes the cost straight to the advertiser with zero transparency. youTube's auction model can easily spike like that when your conversion tracking is even slightly misaligned or you're running a broad match/auto-targeting campaign that YouTube's algorithm decides is "high intent" at any cost.

i run B2B with strict audience segments and capped bids. Even then I've seen the system push CPCs to 3-4x my max bid because the algo optimises for "estimated conversions" rather than actual revenue. If the average cost-per-lead in your industry is $50, the platform will happily charge you $51 per click for a user it thinks might convert at 2%. That's not performance - that's rent seeking.

in any other procurement environment, you'd demand a cost breakdown and a refund for obviously broken traffic. Here we're supposed to just "test different creatives" and trust the machine. Until enough advertisers walk away or demand real granular cost data, nothing changes.