In the rapidly evolving landscape of paid social, the era of manual audience tweaking is effectively over. Anastasia Braitsik, a global leader in data analytics and content marketing, joins us to discuss why the heavy lifting has shifted from the media buyer’s dashboard to the creative studio. As Meta’s machine learning becomes more adept at identifying high-intent users, the differentiator for brands is no longer just who they target, but the visual and emotional hooks used to capture them. This conversation explores the nuances of creative diversity, the limitations of the platform’s latest metrics, and the strategic shift required to combat creative fatigue in a market where the algorithm, not the human, finds the customer. We will delve into the mechanics of algorithmic learning, the difference between volume and variety, and the practical workflows that turn a single production shoot into a powerhouse of data-yielding assets.
As machine learning takes over the heavy lifting for audience optimization, how have you seen the role of traditional targeting methods like interest stacking and lookalike audiences evolve?
The shift has been dramatic, moving us away from the days when we spent hours building dozens of ad sets for different micro-segments. Today, those manual levers like interest stacking and lookalikes matter far less because Meta’s AI has become incredibly efficient at finding the right people to achieve specific campaign objectives. We’ve moved toward broader targeting, which essentially hands the optimization work over to the machine, allowing it to test and learn in real-time. This means our role as advertisers has transitioned from being technical button-pushers to being creative strategists who provide the fuel for the AI. When you strip away the complex targeting layers, you realize that the creative is now the primary way we communicate with the algorithm to tell it who our ideal customer is.
Meta recently introduced a “Creative Diversity” metric in Ads Manager; based on your testing, does this score truly reflect the performance health of a campaign?
It is a fascinating addition, currently labeled “in development,” and it gives us the first real glimpse into how Meta evaluates our visual variety. The metric categorizes diversity as Low, Medium, or High, and it is clearly designed to help advertisers identify when their assets are becoming too visually similar. However, in my recent audits of various client accounts, I’ve seen cases where campaigns I felt were highly diverse—using static images, carousels, and partnership ads—were still rated as “Low.” This suggests that the thresholds are either very strict or that Meta is looking for deeper structural differences that haven’t been fully disclosed yet. Despite these early discrepancies, the metric is a loud signal from the platform that we need to stop being repetitive and start being more adventurous with our visual assets.
Many advertisers mistake “creative volume” for “creative diversity.” How do you define the difference from an algorithmic perspective?
This is perhaps the most common trap: an advertiser uploads 20 ads and assumes they’ve checked the box for diversity. But if those 20 ads feature the same spokesperson, the same product shot, and the same opening line, Meta’s algorithm views them as mere edits of a single creative rather than distinct options. True diversity means giving the system fundamentally different ways to communicate the same core message so it can learn which specific combinations resonate with different people at different times. It’s about changing meaningful elements like the emotional appeal, the customer pain point, or the visual style—moving from lifestyle photography to a meme-style creative, for instance. When you provide meaningful variations, you give the AI a much richer set of signals to work with, which ultimately leads to better optimization.
Video hooks are often the first point of failure for an ad. What strategy do you recommend for developing hooks that actually stop the scroll?
The first few seconds are absolutely everything; they determine whether a user keeps moving or stops to engage with your brand. I always recommend producing several versions of the same video that each begin with a completely different hook to appeal to various motivations or stages of brand familiarity. For a skincare brand, you might test a “POV” hook like finding a non-greasy moisturizer alongside a “three mistakes” educational hook or a “wish I knew this sooner” testimonial hook. The body of the video can remain identical, but changing those first three seconds allows you to speak to the person who is skeptical, the person who is looking for a solution, and the person who just wants to see a transformation. It’s a sensory game where the right combination of text and movement catches the eye before the brain even fully processes the message.
Beyond just toggling between images and videos, what are the “deeper layers” of diversification that high-performing accounts are using right now?
The strongest accounts I see are diversifying across multiple dimensions simultaneously to provide Meta with the broadest possible learning environment. We aren’t just talking about format; we are talking about testing founder videos against user-generated content (UGC), or customer testimonials against unboxing videos. You should be looking at partnership ads with creators, lifestyle photography versus cold product shots, and even educational content or reaction videos. Each of these formats provides a different “flavor” of social proof and emotional connection, helping the system identify patterns in consumer behavior. By rotating through social proof, reviews, and problem/solution angles, you ensure that your brand remains fresh and relevant to a wide variety of audience segments.
Creative fatigue is a notorious performance killer. How does a diverse creative strategy act as a shield against rising costs and frequency?
Creative fatigue occurs when the system has a limited pool of similar assets, leading it to serve the same ad to the same people repeatedly until the frequency spikes and performance craters. When you introduce fundamentally different creatives, you give Meta the flexibility to pivot before that fatigue sets in. As you diversify, the AI can identify which hooks attract specific demographics and which messaging styles convert colder versus warmer audiences. This creates a deeper pool of assets that can rotate naturally as performance shifts, so you aren’t constantly scrambling when your “one winning ad” finally stops working. It makes your campaigns more resilient and efficient because the system always has a “Plan B” or “Plan C” ready to deploy to the right user.
For a team that currently only produces a few ads a month, how can they scale their output without completely reinventing their production process?
The key is to rework the workflow so that one single production effort yields a massive volume of usable assets. Instead of thinking about one shoot as one ad, look at how you can take a winning concept and recreate it in 10 new ways by simply swapping out elements. A single creator can record multiple hooks, different captions, and several different calls-to-action (CTAs) in one sitting, providing you with a library of components to mix and match. You can also experiment with different aspect ratios, thumbnails, and creator-first versus product-first edits to see what sticks. By planning these variations ahead of time, you give yourself a much longer runway for testing without needing to schedule a brand-new shoot every time you need a fresh asset.
What is your forecast for Meta advertising?
I believe we are heading toward a future where the “advertiser” becomes more of a “creative director of AI.” As Meta’s systems continue to automate bidding, placement, and targeting, the competitive advantage will shift entirely to those who can produce high-velocity, high-quality creative that feeds the machine unique data points. We will likely see even more sophisticated AI-driven reporting that tells us not just which ad won, but why it won—analyzing specific colors, words, and emotional triggers. Brands that fail to build a robust creative engine will find themselves priced out by rising costs, while those who master creative diversity will see their efficiency and scale reach levels that were previously impossible.
