From Scale to Curation. Creative Agencies Are Taking a More Intentional Approach to Programmatic.

Reshma Karnik is the Chief Media Officer at BarkleyOKRP, one of the largest independent creative idea companies in the United States. With extensive experience spanning digital media, ad tech, data, and analytics, she brings a deeply data-first perspective to how modern media organizations are built and how brands drive performance.  

Prior to MissionOne Media BarkleyOKRP, she held leadership roles at MediaLink and Dentsu, where she focused on accelerating growth through data, technology, and clean supply initiatives. Reshma was recently named to the 2025 Ad Age Media Buying Power List.

Curation Conversations sat down with Reshma to talk about how programmatic has shifted from scale to curation, the growing roles of first-party data and AI, and what it will actually take to turn those ideas into real, scalable outcomes

‍

You’ve been in programmatic since the early days, how would you describe how the ecosystem has evolved over time?

I started in programmatic back in 2012 when there was really just one display channel and open exchange was king. Scale was everything. Publishers were a dime a dozen, and most programmatic buyers were running campaigns through open exchange. PMPs and programmatic guaranteed deals were considered a luxury because they drove up CPMs, and clients were very cost conscious.

It was a very different environment—much more manual and much more focused on cost and scale. There wasn’t much understanding of curation or the importance of inventory quality. The idea that you had to run media in the right place at the right time was still pretty nascent.  

Over time, we started to see that evolve. We went from open exchange to contextual to PMPs to more intentional buying strategies, but the shift was gradual.  

‍

What does that evolution look like today, especially when it comes to curation?

Today, it’s almost completely flipped. What used to be something you’d “sprinkle in”—a PMP here, a curated buy there—is now table stakes. There’s really no such thing as “spray and pray” anymore. Programmatic and curation go hand-in-hand now. In my opinion, there shouldn’t be a programmatic campaign running today without a proper curation strategy behind it.

We’ve moved from manual, pre-packaged audiences and inventory to something much more dynamic and real-time, with a lot more intentionality and a much heavier reliance on data. First-party data has become a huge currency, and clients are much more focused on how their data is being used and how to control that.

At the same time, there’s been a big shift in power. It used to sit mostly on the demand side, with DSPs leading the conversation around identity. Now, publishers and SSPs are taking a much more active role—building data capabilities, partnering more closely with agencies, and helping create more curated, controlled supply environments.

If I had to sum it up, I’d say we’ve moved from open exchanges and ad networks toward curated marketplaces, with much more focus on quality, control, and transparency.

‍

Programmatic today spans so many channels. How has that fragmentation changed the role of curation?

The programmatic landscape today has evolved extensively, beyond the good old days of just display. Now you’re talking mobile, social, out of home, connected TV, retail media networks, streaming devices—it’s all connected.

You have audiences moving across platforms constantly. Someone might see something on a streaming device, browse online, read an article and then shop in-store—and all of that is part of one journey.

Take retail media as an example. You have companies like Walmart or Target with retail media networks, a massive ecommerce presence, and brick-and-mortar stores—all generating incredible amounts of data. That closed-loop feedback is hugely valuable.

As those channels have expanded, curation becomes front and center because you need to understand where your audiences are across all of these environments and how to reach them in a much more defined way.

‍

Let’s go back to something you said earlier about the evolution in programmatic. You mentioned that there’s been a shift in focus from the demand side to supply side. Do you see this as beneficial to clients?  

The whole ecosystem is important, of course, but the supply side does offer additional layers of transparency and control.  

On the demand side, it’s often a little bit of a black box. There’s an ad tax—everyone pays it when you’re piping through DSPs, and there are multiple layers involved. When you start working more directly with the supply side, you’re removing some of that clutter and those layers.

SSPs like PubMatic and Magnite are also evolving—they’re not just pipes anymore. They’re offering more direct buying options, and that gives you much more visibility into what you’re actually buying. You can really get a look under the hood.

The other big benefit is around data partnerships. You can now work directly with SSPs to curate inventory in a much more calculated way. WPP Media (formerly GroupM) is a great example—they’ve built their own curated marketplace through SSP partnerships, where they manage the inventory, the publisher relationships, and the types of deals they want to make.

That level of control is really powerful. You’re able to prioritize things like attention, quality, and outcomes in a way that’s harder to do if you’re only operating through the demand side.

When I was at Dentsu, we partnered directly with PubMatic through a clean supply initiative. That gave us early access to their capabilities and allowed us to shape how we used their data, their tech, and their inventory. Ultimately, it helped us create more control and better outcomes for our clients.  

We are currently in the process of putting more rigor around Supply Path Optimization strategies through demand and supply side partnerships at MissionOneMedia as well.

I wouldn’t say we should be replacing the demand side entirely, but leaning more into the supply side does create more transparency, more control, and better curation.  

‍

How are client expectations around programmatic performance and curation changing?

I think there’s definitely been a shift. Gone are the days where you could just say, “we’re running a full-funnel approach,” and everyone was okay with that. Clients are a lot more vigilant now, they really want to understand where their dollars are going and what they’re getting back from it. There’s a much bigger focus on ROAS, which means there’s a much bigger focus on data.

We’ve kicked the can down the road quite a bit on third party data and cookie deprecations, but net-net, there’s signal loss. Plus, there’s a lot of clutter in the market. First-party data is one of the best ways around these changes. I think clients are starting to understand that their data is the most reliable place to start.  

They also want to maintain some control over it rather than just handing it off. We’re seeing a shift from, “run my media for me,” to, “how are you using my data, and how is that driving outcomes?” And that’s a very different conversation.

The other piece is attention. Attention has become a real currency now. It’s not just about getting impressions—it’s about showing up in places where people are engaged.  

Overall, we’re seeing a move toward something much more deliberate—more focused on ROAS, more focused on control, more focused on quality—and that’s really where curation comes in.

‍

How do you actually figure out where attention is, not just impressions?

That’s where AI comes in. It helps you understand how people are engaging with content—what they’re reading, how much time they’re spending, where they’re clicking.  

Publishers are already doing this. If you think about something like The New York Times, they’re collecting first-party data by observing your behavior on the app or website, and then they curate content for you based on what you read. The next time you go to the site or open the app, you don’t have to spend time wading through stories you don’t care about. The ones you want show up first. I think that’s great. It’s all about an intentional move towards personalization.

That same idea applies to media. You’re looking at context, time spent, and engagement. All of that feeds into attention metrics. And that’s something clients are increasingly asking for. It wasn’t really a thing five or seven years ago, but now it’s becoming a core KPI. Companies like Adelaide are already integrating with DSPs like DV360 to enable custom bidding and making it easier for programmatic buyers to bring attention metrics into play.

‍

How do brands who don’t have as much first party data as a publisher like the New York Times find and measure attention?

It’s a good question. You do have to constantly balance scale and performance. First-party data is powerful, but it doesn’t always scale the way you want. This is where partnerships come in, especially partnerships with SSPs. They already have relationships with publishers, access to data, and AI systems that help identify high-attention environments.  

Combining that with the client’s first-party data allows you to build a more informed, curated approach.  

Clean rooms can also be used to bring in brand first-party data and matching that with Publishers and Supply partners first-party data; to analyze, model and activate only on high value audiences, can drive better personalization and ROAS.  

‍

We’ve been talking a lot about the technology, but where does that leave the creative?

The creative and human piece is very, very important. I mean, there are a dime a dozen AI tools out there—everyone’s talking about generative AI, and it’s at all of our fingertips. But there has to be a human component. It’s always a balance of art and science.

Creative has to appeal to me for me to click, take action, buy the product, or even react to it. So, in the curation process, it’s really important that we bring creative and media together from the start. Creative ideas can be so high in the sky that they’re not actually executable in media, and media can be so surgical that it doesn’t get anyone to engage. Again, it has to be a balance.

We at MissionOne Media  have built M1 Refinery, an agentic way to analyze and curate what elements of the creative resonate the best with an audience, which allows humans to tailor creatives for each audience.  

We are also building a product called Personified, which is focused on exactly that too. It’s about using first and third party data to really understand what audiences like, how they behave, and what they’re interested in, and then using those insights to inform messaging and drive personalization. The creative then conforms to the audience you’re trying to reach.

And that’s where the human piece comes back in—because you need teams to interpret that data and translate it into something that resonates.

‍

What do you see as the next evolution, and what should brands and advertisers be preparing for?

Putting it all together—AI, curation, first-party data—sounds great in theory, but it’s not as widely practiced as people think. There’s still a long way to go. We’ve been talking about first-party data for years, and I still don’t think it’s being leveraged to the extent that it should be.

I think the next evolution is a practical one. How do we actually use AI to scale personalization? How do we bring more contextual intelligence into the mix? And how do we use first-party data—both from the brand and from the agency—in a way that’s actionable?

The other piece is partnerships. You’re going to see a lot more collaboration with AI-driven platforms, with SSPs, and with publishers because no one can do this in isolation anymore.

Automation is also going to be key. The more we can remove manual processes from programmatic—whether it’s through custom bidding, activation and optimization—and create real-time actions, the more efficient everything becomes. It reduces the manual tasks and enables programmatic buyers to be more strategic with focus on precision. That’s why the shift to Agentic AI is important.

But it all kind of ladders up to the same thing: if you can use AI and first-party data to drive better personalization at scale, you’re going to drive better ROAS. That drives client satisfaction, which drives more investment—it becomes a bit of a flywheel. We’re not fully there yet, but we’re close.  

‍