MDrew Sanocki just posted the kind of note ad buyers usually keep to themselves.
COh, what's he seeing?
MHe says when frequency climbs on a cold audience, he reads it as Meta running out of new people at that price.
CThat's the quiet part out loud. Same ad hitting the same people means you're paying to annoy who you already have.
MBut is that Meta's fault, or just every ad platform after a while?
CIt's both. But advertisers are finally saying it. Some advertisers claim Meta's been pushing frequency to keep revenue up.
MHold on. Frequency going up could mean the creative is working.
COn a cold audience? That's not how it works. A cold audience hasn't bought from you. You want reach, not repetition.
MAnd Drew's advice is to check the new-customer cost against your Shopify file before you trust the pixel.
CBecause the pixel can count a customer as new even if they bought from you before.
MThat's a real gap. If Meta's own pixel overstates new customers, the whole campaign looks better than it is.
CAnd brands keep spending because the dashboard says it's working.
MI wouldn't go that far. Plenty of buyers already do this reconciliation.
CThen why did Drew post it like a warning?
MBecause pressure to hit scale makes people lazy. That's not a Meta conspiracy.
CIt's not a conspiracy. It's an incentive. Meta gets paid when you spend more, not when you find new customers.
MBut if advertisers cut spend, Meta has to fix it. That's the market working.
CMaybe the correction is coming. I just think it'll be slow.
MSmall businesses get hurt worst. They don't have Shopify data teams to double-check.
CExactly my point. A local shop sees "new customers up" and keeps burning cash.
MBut Drew's post is aimed at people who can check. The people who need it won't see it.
CThat's the real problem. The ones who get hurt never read the fine print.
MFine. But that's every ad platform ever, not just Meta.
CI'll give you that. I still think frequency climb is a warning signal, not a glitch.
MNoted. And listeners can decide if their dashboards are lying to them.
CSpeaking of things that sound better on a conference stage than in a data center, a speaker at a robotics event says a fresh PhD just spun off a startup to make server racks that use a tiny fraction of the energy.
MTiny fraction? That phrase is doing a lot of work without a single number.
CBut if it's real, it's huge. Data centers are the new factories, and they're eating power like crazy.
MEvery startup says their racks are more efficient until the utility bill shows up.
CCome on, a materials or thermal breakthrough could be real. We've seen stranger things.
MI'm not saying it's impossible. I'm saying "a tiny fraction" with no data is a pitch, not a result.
CThe woman just finished her PhD. It's early. But early is exactly when big leaps happen.
MAnd early is also when most of them die. We don't even know her name or the startup.
CFine, it's a conference anecdote. But it got a talk slot, so someone took it seriously.
MRobotics conferences give slots to plenty of things. Remember the Love Boat robot?
CAlright, low blow. But energy efficiency in server racks is a real problem.
MIt's a huge problem. I'm not dismissing it. I just want one number.
CIf she can cut data center energy use by even a third, every cloud provider would line up.
MNow you're moving the goalposts. The post said "tiny fraction."
CBecause that's how people talk before they publish. I'm giving it room.
MI'll give it room when there's a white paper. Until then, it's a promising rumor.
CFine. We'll keep an eye on it. But demand is real, so someone will crack it.
MOn that, we agree.
CAnd if it works, I reserve the right to say I told you so.