The digital marketing landscape is currently undergoing a massive paradigm shift, moving away from the superficial comfort of vanity metrics and toward the cold, hard reality of incremental growth. As a global leader in SEO, content marketing, and data analytics, Anastasia Braitsik has been at the forefront of this evolution, helping brands navigate the treacherous waters where high performance on paper doesn’t always translate to actual business expansion. Her approach challenges the status quo of traditional attribution, urging marketers to stop simply capturing existing demand and start generating new value through rigorous experimentation and AI-driven insights.
In this discussion, we explore the common pitfalls of relying on standard customer acquisition costs and the deceptive nature of branded search campaigns that often cannibalize organic traffic. We delve into the mechanics of incrementality testing, the importance of establishing firm decision rules before data hits the desk, and how modern AI tools are finally leveling the playing field for smaller teams. By shifting the focus from platform-reported ROAS to incremental CAC, the conversation provides a roadmap for marketers who want to ensure every dollar spent is truly moving the needle rather than just paying for conversions that would have happened anyway.
Many marketing teams feel a sense of accomplishment when they see a branded search campaign delivering a $15 cost per acquisition, but you’ve often cautioned that this efficiency can be an illusion. How should growth leaders differentiate between simply capturing existing demand and actually generating new customers?
The dopamine hit of seeing a $15 CAC on a dashboard is incredibly seductive, especially when compared to a paid social campaign that might be sitting at $60. However, growth leaders must look past the surface-level efficiency to understand the “why” behind those numbers. In many cases, those $15 conversions are coming from people who were already typing your brand name into a search engine, meaning they were already at your digital doorstep. If you are paying for those clicks, you might just be placing a toll booth on a road the customer was already traveling for free. The real challenge is determining how many of those customers would have found their way to a purchase even if the ad wasn’t there. When we look at that $60 social campaign, it might feel expensive and “inefficient,” but if it’s reaching people who had never heard of your brand, its actual value to the company’s long-term growth is significantly higher. We have to stop rewarding campaigns that just show up at the finish line and start investing in the ones that actually start the race.
You’ve mentioned that incrementality testing can lead to some surprising budget decisions, such as significantly reducing spend on high-performing channels. Could you walk us through the logic of why cutting a “successful” campaign, like a Google brand search test, might actually be the right move for a brand?
It sounds counterintuitive—almost like professional heresy—to suggest cutting a high-performing campaign, but the evidence from incrementality testing is often undeniable. In one specific instance involving a Google brand search test, we discovered that reducing spend by roughly 90% had almost no significant impact on actual customer acquisition. What was actually happening was a classic case of paid ads cannibalizing organic search results; as soon as the paid spend was dialed back, the organic listings naturally rose to capture that demand. This realization shifts the entire conversation from “how do we get more conversions” to “how do we stop paying for what we already own.” By pulling back that spend, the marketing team was able to reclaim a massive portion of their budget that was essentially being wasted on existing intent. Those funds can then be redirected toward non-brand search or other top-of-funnel activities where they can actually drive a 55% conversion lift, as seen in various lift studies. It’s about having the courage to trust the data even when it tells you to stop doing something that looks “good” on a standard report.
When it comes to non-brand search or reaching new audiences, the costs often look much higher on paper. How can marketers use the concept of incremental CAC to justify these “more expensive” investments to their finance teams?
The conversation with finance becomes much easier when you stop talking about platform-reported CAC and start talking about incremental CAC. If a platform says a customer cost $60, but an incrementality test shows that the ad was responsible for every single one of those conversions, then that $60 is a true reflection of growth. In contrast, if a $15 branded search conversion has an incremental lift of only 10%, the “real” cost of that additional customer is actually $150. When you present it this way, the “expensive” $60 social campaign suddenly looks like a bargain compared to the “cheap” $15 branded search. Using a conversion lift study to prove a 55% increase in total volume provides the empirical evidence needed to justify a higher spend. We aren’t just spending money for the sake of it; we are buying additional market share that wouldn’t exist otherwise. This shift in perspective transforms the marketing department from a cost center that “spends” into a growth engine that “invests” based on verified returns.
Establishing decision rules before a test begins is a critical part of your framework. Why is it so important to define what a “good” result looks like before the data starts pouring in, and how does that influence the way a team reacts to a $200 lifetime value?
Human nature is to see what we want to see in the data, which is why setting thresholds in advance is the only way to maintain objectivity. If we know a new customer has a lifetime value of $200, we need to agree on what we are willing to pay to acquire them long before the test results flicker onto our screens. For example, a team might decide to scale any campaign where the incremental CAC is under $120, hold steady if it’s between $120 and $180, and immediately cut spend if it exceeds that $180 mark. Without these rules, a team might see a $190 incremental CAC and try to justify it because they “feel” the campaign has brand awareness benefits that aren’t being captured. Having a pre-defined line in the sand prevents that kind of emotional backtracking and ensures that we are making evidence-based decisions. It turns the readout from a subjective debate into a tactical execution plan, allowing the team to move with much more speed and confidence.
AI is currently being used to streamline the more technical aspects of incrementality testing. How are you seeing growth teams utilize tools like LLMs or Meta’s GeoLift to handle the heavy lifting of experiment design and data analysis?
AI is effectively democratizing high-level data science, allowing smaller teams to run sophisticated experiments that were once reserved for companies with massive analytics departments. We are seeing teams use general-purpose LLMs like ChatGPT or Claude to analyze historical sales data and identify “twin” geographic markets that have similar sales patterns to serve as control and test groups. This used to take a human analyst days of crunching numbers in spreadsheets, but now it can be done in minutes with a high degree of accuracy. Beyond just the design, AI tools like Meta’s GeoLift or platforms like Haus are being used to monitor tests in real-time, flagging anomalies like a sudden regional promotion or a tracking error that could skew the results. If a conversion spike happens in a control market, the AI can alert the team immediately, saving a test that would otherwise be ruined. This allows the human marketers to focus on the strategy and the “why,” while the AI handles the repetitive, technical “how.”
Even with the best AI and testing frameworks, there are still risks involved in attributing conversions. What are the common environmental factors that can trick a measurement tool into crediting a campaign for a lift that it didn’t actually cause?
It is vital to remember that correlation is not causation, and even the most advanced AI can be fooled by the “noise” of the real world. A campaign might launch the same week that a competitor goes out of business, or during a seasonal peak that naturally drives interest in your category. If conversions jump by 20% during that period, the measurement tool might erroneously credit your new ad spend for the entire lift, when in reality, a pricing change or a viral social media moment was the true driver. External factors like regional holidays, shifts in consumer sentiment, or even a sudden change in the weather can all create “false positives” in your data. This is why a well-designed experiment with a clean control group is so essential; it provides the baseline of what would have happened anyway, allowing you to isolate the campaign’s impact from the background noise. You always have to verify the AI’s calculations against the context of the business and the broader market to ensure you aren’t making million-dollar decisions based on a statistical fluke.
What is your forecast for the evolution of growth marketing metrics over the next few years?
The era of “set it and forget it” attribution is coming to a definitive end, and we are moving toward a future where incrementality is baked into the very core of every marketing platform. I predict that the standard dashboard metrics we use today—like basic ROAS or total CAC—will eventually be relegated to secondary status, replaced by real-time incremental lift scores powered by edge computing and privacy-safe data modeling. We will see a shift where every marketing dollar is tagged with a “probability of influence,” allowing brands to dynamically shift budgets across channels based on which one is generating the most actual growth at any given second. Smaller brands will have access to the same “measurement-first” toolkits as the giants, meaning competition will be won not by who has the biggest budget, but by who has the most disciplined testing culture. The ultimate goal is to reach a point where we are no longer guessing what works, but rather operating with a level of financial precision that makes every marketing investment as predictable as a high-yield savings account.
