Milena Traikovich stands at the forefront of the modern demand generation landscape, specializing in the intersection of deep-dive analytics and high-impact performance optimization. With a career rooted in refining how businesses identify and nurture high-quality leads, she has become a leading voice for marketers who feel trapped by the “black box” of automated advertising systems. Her approach moves beyond the surface-level metrics that often clutter dashboard screens, focusing instead on the strategic integrity of audience data. By bridging the gap between programmatic execution and first-party insights, she helps brands regain control over who sees their messages and why.
In this discussion, we explore the pitfalls of over-relying on platform-generated audience segments and the common tendency to prioritize delivery speed over targeting accuracy. We examine the inherent limitations of intent data—such as its short shelf life—and why a successful campaign must be anchored in marketer-defined data sources like CRM records and verified identities. Finally, we look at the tactical importance of the matchback process, a method that allows teams to audit their audience effectiveness and continuously refine their targeting strategy based on actual conversion behavior rather than platform assumptions.
Most marketers find themselves constantly adjusting bids and creative to improve performance, but you argue that we are often ignoring a more fundamental problem. How do we reconcile the immediate gratification of delivery metrics with the deeper question of whether we are reaching the right people?
The reality of our current environment is that programmatic platforms are exceptionally good at the mechanical act of delivering ads, but they are significantly less effective at deciding who should actually see them. It is incredibly easy to get caught up in the “dopamine hit” of delivery metrics because the feedback is instantaneous; you can see if your spend is pacing correctly, if ads are serving, and if conversions are coming in at a cost the team can live with. These numbers give you something to fiddle with while the campaign is still active, providing a sense of control. However, focusing solely on delivery is essentially backwards because you are letting the platform define the audience, deliver the media, and report the results all within its own closed loop. To truly drive value, we have to step back and realize that while these delivery numbers are necessary for managing a campaign day-to-day, they are a weak foundation for deciding if the audience itself was the correct choice.
When we look at the specific segments provided by programmatic platforms, such as an “SUV in-market” category, what are the hidden risks that marketers might be missing?
The primary risk is that intent data has a very specific shelf life that often expires long before the ad is served. On paper, targeting someone labeled as “in-market” for an SUV sounds like a precision strike, but the visibility into why that person was included in that segment is usually non-existent. For example, a person who was aggressively researching SUVs six months ago might appear to have high purchase intent, but they may have already bought a vehicle and moved on, or they might have just been comparison shopping for a friend. Both individuals still show up as “in-market” in the platform’s eyes, yet only one represents a genuine opportunity. Relying on these pre-packaged labels means you are trusting a system that might be working with stale data, and because the conversion report comes from the same system that created the audience, it becomes very difficult to independently verify if you are actually reaching a fresh, relevant group.
If the platform shouldn’t be the one defining the audience, where should the process actually begin to ensure media dollars aren’t being wasted?
The strongest campaigns—the ones that truly move the needle—begin with an audience that the marketer already understands and owns. This means building your target list from CRM records, first-party data, verified identities, or specific household attributes before a single dollar is spent on media. When you start with a marketer-defined audience, you shift the role of the programmatic platform from a decision-maker to an activation tool. You aren’t asking the platform to guess who belongs in your audience; you are giving it a specific list of people to find across the web. This creates a clear separation between the “who” and the “how,” ensuring that your strategy is rooted in data you have already vetted for quality and relevance.
You’ve mentioned that marketing rarely works in a straight line, especially in high-consideration categories. In this context, how does a “matchback” analysis help us judge the quality of our audience?
In categories with longer buying cycles, like the automotive industry, a buyer’s journey is rarely a simple “click and buy” event; they might see multiple ads, visit the website, speak with a salesperson, and wait several days before making a final decision. A matchback doesn’t try to claim that one specific impression caused a sale in a vacuum. Instead, it asks a more fundamental strategic question: were the people who ultimately converted actually part of the original audience we built? By comparing your conversion file against your initial audience file, you gain evidence of whether your campaign was pointed in the right direction. If you find that a significant portion of your converters came from outside your expected audience, it’s a signal that something is off—perhaps your original profile was too narrow, or one of your intent sources was stale, or maybe your strongest buyers share attributes you hadn’t even prioritized yet.
Once a campaign is over and the data is in, how can marketers ensure they aren’t just repeating the same mistakes in their next initiative?
The real value of this data-driven approach is that it turns every single campaign into a laboratory for the next one. Instead of starting every new flight with the same set of assumptions, you use the matchback results to build your next audience from actual conversion behavior. If the data shows that certain criteria didn’t lead to conversions, you remove those weak links. If you see that converters share a specific attribute that you previously ignored, you give that detail more weight in the next list. Without this constant refinement, it is too easy to keep reusing the same audience just because the last campaign “hit its numbers,” which often means you are missing out on an even better, more efficient group of buyers. By judging the audience against the people who actually acted, you ensure that your targeting improves in lockstep with your campaign performance.
What is your forecast for programmatic audience strategy?
I expect we will see a decisive shift away from “black box” platform segments as marketers demand more transparency and move toward 100% ownership of their audience definitions. In the coming period from 2026 to 2028, the industry will likely prioritize “identity-first” targeting, where the focus is on matching verified first-party data across environments rather than relying on vague interest categories. We will see the role of the programmatic trader evolve from someone who manages bids to someone who manages data integrity, ensuring that the feedback loop between conversion files and audience builds is shorter and more automated. Ultimately, the brands that win will be those that treat their audience list as a living, breathing asset that is constantly audited and refined, rather than a static setting they toggle on inside a dashboard.
