Alright, let's break it down. Drew Stein’s recent piece on curation in media buying, titled "How Curation is Reshaping Media Buying," carries some weight in the conversation around programmatic media. If you haven’t read it, let me catch you up: Stein says curation is no longer just a buzzword but is now a “strategic necessity”. And to his credit, he's not wrong in many respects. But, let's be clear here—the world has changed. Just two years ago, we were all on the curation bandwagon, but now, everybody’s got AI, and it’s reshaping the entire landscape, including the value of curation.

Let’s start with what Stein gets right: curation is undeniably crucial today. “Curation has shifted from buzzword to strategic necessity,” he says. And yes, he’s right. As Stein points out, fragmented data and the rise of privacy concerns have made curation more important than ever. If you’re still operating under the assumption that throwing a bunch of random data into your media-buying strategy is going to work, you’re behind the curve.

Stein points out that 66% of open exchange transactions now route through PMPs (Private Marketplace deals). That’s huge. PMPs have become the VIP room of digital media—a place where advertisers can avoid the chaos of open exchanges and have more control over their data and targeting.

So far, so good. But here’s the thing: the game has changed, and AI is the big player now. I was all in on curation a couple of years ago, just like Stein.

It was the shiny new tech, and everyone was all about how it could optimize buying strategies and deliver on privacy-first goals.

But now? AI is everywhere, and it’s integrated into every single step of the media buying process, including curation.

Just having curation as a tactic isn’t enough anymore—you need AI-driven optimization to really make the most of it. Curation, on its own, is no longer the dazzling innovation it once was.

Stein backs up his claim with a stat—“66% of open exchange transactions now route through PMPs.” Sure, that’s accurate and reflective of the trend towards more curated, private deals, and I don’t disagree. PMPs are the future. They provide more transparency, better targeting, and privacy compliance than the free-for-all of open exchanges. But what Stein doesn’t mention here, and what needs to be acknowledged, is how quickly the rest of the industry caught up.

I mean, sure, PMPs are growing, but AI-driven optimization is the rocket fuel here. If you’re not using AI alongside PMPs, you’re missing out. The ability to analyze real-time data, to optimize on the fly, and to drive even more precise targeting—that’s where the real magic is now. AI can’t be ignored anymore in this space. If your strategy doesn’t include AI, then you’re going to be stuck in 2019 while everyone else is already in 2025. So, while I’ll give Stein credit for pointing out the shift, AI is the next logical step in this evolution.

Stein claims, “Curation yields measurable improvements: lower CPA, improved ROAS, flexible goal alignment.” Now, I’m not here to tell you that curation doesn’t deliver. It absolutely can. But there’s an issue here: not all curated deals are created equal. Yes, the data shows improvements in CPA and ROAS for many, but it’s not guaranteed for everyone. Curation is not a one-size-fits-all solution, and Stein doesn’t address the fact that performance can vary wildly depending on factors like data quality, platform integration, and real-time optimization. It’s the “buy once, expect perfect results” mentality that’s been plaguing the industry for years. Curation doesn’t solve every problem—it can just help clean up the mess.

Stein also skips over the fact that AI and machine learning have radically altered the way we think about ROAS and CPA. With AI, optimization happens in real time, and decisions are based on patterns and predictive models, not just static data dumps. So while curation can improve efficiency, it’s AI that’s driving the major performance gains.

Now, let’s talk about the privacy thing. Curation is Stein’s answer to signal loss and privacy concerns. He says, “Curation helps address signal loss and privacy/regulatory challenges.” And, sure, there’s some truth to that. Curation allows advertisers to blend first-party, modeled, and contextual data, which helps navigate the privacy minefield left by the death of third-party cookies.

But let’s not pretend this is the full solution. Privacy laws are evolving faster than I can scroll through LinkedIn, and we all know that what worked yesterday might be obsolete tomorrow. Sure, curation is part of the solution, but AI and machine learning are the real heavy-hitters here. These technologies can help identify patterns and adjust in real time to comply with ever-evolving regulations, making them critical to future-proofing your strategy. So while Stein’s claim about curation is true, it’s only half the story—and AI is the other half.

Stein paints a glowing picture of curation, but there are some critical omissions here. What about the complexity of implementation? Curation isn’t as simple as clicking a button. It requires high-quality audience data, strong integration with SSPs, and real-time optimization—and that’s just to get started. In fact, it can be more complex to implement than the generic, open exchange strategy Stein is trying to replace.

Furthermore, Stein fails to address how AI has become the default optimization tool in the industry. AI isn’t just a nice-to-have anymore; it’s the core technology driving efficiency. If you’re not integrating AI-driven curation, you’re leaving money on the table.

And let’s not ignore: the trade-offs of curation. It requires investment in tech infrastructure, resource allocation, and staff training—all of which can create barriers to entry for smaller agencies or advertisers. So yes, curation works, but it’s not an easy button. You need to be ready to commit significant resources.

So, what’s the bottom line here? Drew Stein’s article on curation is solid, but it's a bit like talking about floppy disks in a world that’s already moved on to cloud storage. Curation is still important, but AI is the new frontier. Stein’s optimism about curation is not wrong—it’s just incomplete. Yes, it’s a critical tool for tackling fragmentation and privacy issues, but it’s no longer the shining star it once was. AI has arrived, and it’s already taking the wheel.

Curation is still a must-have, but if you’re not adding AI-driven optimization to your toolkit, you’re living in a world that’s already moved past you. The future of media buying isn’t just curation—it’s curation powered by AI.