Ad Buyers Are Being Replaced by Their Own Tools

A few days ago, a friend in the ad-buying business was venting to me.

He said the media buyers his team spent three years building up have less and less to do these days. It's not that budgets were cut — budgets are still climbing. It's that the machines have taken over the work they used to do.

The things he used to do every morning — check the data, adjust bids, segment audiences, test creatives — now, the moment he opens the dashboard, the system has already done it all for him.

He's a little rattled.

My first reaction: this isn't just happening on Google.

Seven Platforms, All Doing the Same Thing

Go look at the major ad platforms right now.

Google has Demand Gen and Performance Max. Meta has Advantage+. TikTok has GMVValue and Smart Campaigns. Snapchat has Dynamic Catalogs. Amazon has Performance Plus. Pinterest has Performance+. Microsoft has its own Performance Max too.

Seven platforms.

The names are all different, and each one pitches itself in its own way. But lay them side by side and look — underneath, they're all doing exactly the same thing.

What do I mean by "the same thing"?

It's this: the machine is making three decisions on your behalf — who sees your ad, how much you bid, and which creative runs.

These three decisions used to be judgment calls a media buyer made from experience. Now? Machine learning can run tens of thousands of trial-and-error cycles every second. Whose reaction time beats that?

Google: So You Don't Even Have to Pick Where Your Ads Run

Let's start with Google.

Google's Performance Max — the name alone tells you how ambitious it is. "Performance maxed out." You just hand over your creatives and your budget, and it finds the best-value slot in real time across search, display, video, shopping, and maps — every ad surface Google owns.

Demand Gen goes a step further. It identifies people before they've even searched, before they know they want to buy — based on search behavior and interest tags, it surfaces the ones who "might be ready to buy but haven't pulled the trigger yet," and hands them to you ahead of time.

Think about it: five years ago, how many media buyers staring at how many reports would it have taken to pull that off?

Meta: Moving Budget to Where It Wins

Meta's Advantage+ takes a different approach.

It doesn't manage ad placement (that's Google's core strength). What it manages is how your budget gets allocated. You launch ten creatives across five audiences, and the system watches — whichever creative performs well, budget piles onto it automatically; whichever audience converts at a higher rate, budget shifts over immediately.

It's not a weekly adjustment. It's adjusting every single minute.

That's what machine learning does. It doesn't get tired, and it doesn't get sentimental — it doesn't keep pouring money into a creative just because it performed well yesterday.

TikTok: Built to Sell

TikTok's GMVValue — the name gives away its goal right there: GMV (Gross Merchandise Value), meaning total sales.

Unlike other platforms that optimize vaguely for "clicks" or "engagement," it tells you straight: I'm here to help you move product. The machine learning predicts which users are most likely to place an order, then serves the ad to them.

For e-commerce brands, it's basically a sales associate who actually does the math.

TikTok also has Smart Campaigns, built for small shops. Don't know how to set bids? Don't understand audience targeting? No problem — tell it what you want (brand awareness? or conversions?), and it runs the whole thing for you.

The Rest Follow the Same Logic

Let me run through a few quickly.

Snapchat's Dynamic Catalogs — you upload your product catalog, and based on what a user has viewed or clicked, it automatically assembles a matching ad. One user sees one version; another user sees something different.

Amazon's Performance Plus — Amazon knows better than anyone who's in the mood to buy. It leans on that purchase-intent data to adjust your bids and creatives. When it comes to ROAS (Return on Ad Spend), nobody crunches the numbers better.

Pinterest's Performance+ — Pinterest is essentially a search engine. What people search for on it is what they want to buy or want to do. Its recommendation algorithm helps you find exactly that audience.

Microsoft's Performance Max — same name as Google's, covering Bing, LinkedIn, and the Microsoft Audience Network. B2B advertisers care about this one more.

So there you have it — seven platforms, and the core logic is identical.

Hand the decision-making over to the machine. Move the human off the operating console.

What This Means for You

Back to my friend.

I told him: the media buyer role isn't going to disappear, but its definition is changing.

The core skill of a media buyer used to be "tuning" — tune the bids, tune the audiences, tune the creatives. That work, the machine does faster, more reliably, and more cheaply than any human.

So where does the human value go?

It goes to the things machines can't do.

A machine can tell you "who to target," but it can't tell you "who your brand actually needs to be talking to." A machine can calculate which creative gets a higher click-through rate, but it can't craft a story that genuinely moves people. A machine can optimize the conversion funnel, but it can't decide whether you should build a completely different product this year.

The machines have taken over the operating console — but the strategy chair is still empty.

Whoever sits down in it stays in the game.

One Last Thing

My friend later said something that I think nails it.

He said: media buyers used to be the driver — foot on the gas, hands on the wheel. Now the platform has taken over the wheel and the gas pedal, and the buyer is sitting in the passenger seat.

But the passenger seat's job is to tell the car where to go.

The faster the car drives, the more the passenger matters.

May you, from the passenger seat, still see the road ahead clearly.