Media Buying is dead 😵

The big reveal was this. You can now plug Claude into your Google Ads account, and Claude will pause your underperforming campaigns. Saves you so much time. Media buyers are gone. The future is here.

I sat with that for a minute. Mostly the word underperforming.

Underperforming compared to what? On what time horizon? Against which blended target? In a learning phase or a scaling phase? Inside a placement shift the algorithm hasn't told anyone about yet? On an offer where the front end loses money on purpose because the back-end LTV pays for it?

"Underperforming" isn't a number. It's a judgment.

And every judgment in this work depends on context the platform doesn't have, the dashboard doesn't show, and the AI can't see, because that context lives in conversations you've had with the client, patterns you've watched across cycles, and parts of the business that don't show up in the ad account at all.

I closed LinkedIn and made another coffee.

Then I started writing this.

Because here's the thing. There's a version of media buying that absolutely is dead, or close to it. The version that's mostly button-pushing. Bid adjustments. Pause-and-scale on a CPA target. Following purely Google's recommendations. Setting up campaigns from a template. If that's what your media buyer does, the LinkedIn post is right about your media buyer.

It's just wrong about the job.

Most mornings, the work doesn't look anything like what that LinkedIn post described. Here are nine things about it.

1. CPA going up isn't always bad. Pausing the campaign impulsively -  almost always is.

A founder opens their account in the morning, sees CPAs spiking across campaigns that have been profitable for weeks, and the first instinct is to pause everything before it gets worse. It feels responsible. It's almost always wrong.

Spikes are usually a symptom, and the cause is rarely on your end. A silent placement shift, where Google decides to push your in-stream traffic to in-feed overnight. A bidding system stuck in relearning because someone changed a bid mid-cycle. A creative that fatigued and started attracting the wrong audience. A landing page issue that broke conversion tracking three days ago. The dashboard reads the symptom, not the cause.

If you pause, you don't fix the cause. You just lose the campaign learning the algorithm has been building for weeks, and you start over from zero on a problem that wasn't a campaign problem in the first place.

This is exactly the move the LinkedIn post is selling. Claude will pause your underperforming campaigns. On what basis? CPA reading? CPA is an indicator, not a diagnosis. When the dashboard goes red, the question is never what do I pause? It's what changed? Nine times out of ten, the change isn't where you think it is, and the answer doesn't live in the platform.

2. Your audience strategy is making you slower, not smarter.

I had a client a few years back running video creation software. Animation maker, that kind of thing. The audience was easy in theory. Content creators. Editors. Animation people. We layered in everything that made sense. Affinity for video. In-market for creative software. Custom segments based on YouTube editing tutorials.

The account stalled for months.

Out of pure frustration, I stripped all of it out one day. Every audience layer, gone. The campaign went broad. Just signal, no targeting. The account started converting almost immediately and scaled for nearly a year on broad alone.

When I finally pulled the audience insights report, the highest-converting segment was fast food lovers. I'm not making that up. People Google classifies as fast food enthusiasts, buying animation software. If I'd proposed that audience to the client at the start, I'd have been fired by lunch.

The lesson isn't that fast food lovers buy software. The lesson is I'd never have found out. Every audience layer narrows the pool of people Google's allowed to show your ads to. You're paying CPM premiums to compete in audiences your gut said made sense, while the actual buyers are sitting in audiences nobody would have proposed.

Your intuition about your customer used to be your edge. The algorithm now has more behavioral data on every user than you have on your entire customer base. Helping it less is helping it more.

3. The number that matters is the only one you're probably ignoring.

Demand Gen has a high CPA. So does cold YouTube. So does most upper-funnel work. Every quarter, someone tells me they want to cut Demand Gen because the CPA looks ugly compared to branded search.

Six weeks after they cut it, branded search collapses, and they can't figure out why.

If you kill the highest-funnel campaign because its attributed CPA is the highest, you're not optimizing. You're amputating. The number that matters is blended - you must know your blended goals.

4. Targeting is now the creative, and Google's been telling you that for years.

Five years ago, the question every paid team asked first was who are we targeting. The audience deck was where the strategy lived. The job of the creative was to land the message on a pre-selected group.

Google's been quietly walking the industry the other direction since at least 2021. Optimized targeting first, which let the algorithm expand beyond your audience selections if it found better converters. Then broad match becoming the default for search. Then lookalikes being demoted to "audience signals" in Pmax and Demand Gen, where they're hints, not constraints. Each one of those changes was Google saying the same thing in a slightly different way: stop trying to pick the audience, let us find them.

Most teams didn't update. They kept building the audience deck, kept layering interests, kept refining lookalikes, while the platform was telling them the lever had moved.

The lever moved to creative. The hook does the targeting. The headline does the targeting. The first three seconds of the video do the targeting. The algorithm reads who's leaning in and pulls more of them. Audience saturation isn't a real problem in 2026. Creative fatigue is.

5. Best practices are how you stay average.

Inceptly was the first agency to run and scale a VSL as a YouTube ad. At the time, "that won't work" was the consensus. Today, every serious DR brand has one in rotation.

Best practices describe what worked for someone else, in a different account, with a different audience, at a different time. They're a starting line, not a finish line. Every account I've ever scaled had at least one move in it that contradicted what a Google rep would have suggested.

This is where the AI argument gets things backwards. AI is, by construction, a best-practices machine. Useful for the bottom 50% of advertisers. Useless for the top 10%, because the top 10% are doing what the average hasn't caught up to yet.

Ask your media buyer one question this week. What did we learn last month that changes what we're doing this month? If the answer is "we found a winner and scaled it," you don't have a system. You have luck.

6. The job most media buyers don't realize they have isn't media buying.

The version of media buying that's dying isn't actually media buying. It's campaign management. Launch, monitor, pause, scale. Follow the recommendations. Trust the algorithm. Hit the targets. Fine work. Easy to automate. Probably will be.

The job that survives in 2026 is something else. It's a diagnosis. Reading the situation, not the dashboard. Knowing why CPA spiked when nothing on your end changed. Sitting with a frustrated client and figuring out the real problem, which is almost never the one they came in with. Holding an opinion the platform won't tell you to hold. Staying calm when everyone else wants to pause.

That's the part of the work AI can't do, because it depends on context the platform doesn't have. The patterns you've watched across cycles. The parts of the business that don't show up in the ad account at all.

Most mornings, I open the same accounts I opened a year ago, ask the same questions I asked a year ago, and notice things I didn't notice a year ago. That's the part nobody can plug into Google Ads for you. This will keep you alive, veeeeery much alive :)

Cheers!

💌 Like this newsletter? Let's continue the conversation

Get in touch with us by responding to this email or tagging us on LinkedIn or Instagram, and sharing your thoughts. Your feedback helps us keep our newsletter relevant and interesting.