How Dynamic Creative Optimization (DCO) Should Work
We all know it when we see it: You search online for a potential family vacation to Disney World, only to start seeing online ads with images of Mickey Mouse with copy that reads, “Book your dream vacation today!”
That tailored ad is an example of dynamic creative optimization (DCO) in action. Digital marketers have been reliant on DCO tactics for years, using real-time data—including user behavior, preferences, and contextual signals—to automatically tailor ads to whoever’s on the other side of the screen. This reliance has taken away a marketer’s agency.
But the rules of audience targeting are changing, both literally and figuratively—and marketers are wondering if DCO is equipped to deliver results they seek after all.
The status quo and limitations of DCO
With DCO, marketers can upload pre-approved headlines, images, and CTAs into ad platforms run by Google, Facebook, and Amazon. From there, the algorithms mix and match those elements to serve individual users with a personalized ad that is, in theory, optimized for peak performance.
It’s an audience targeting tactic that has served marketers well over the years. In fact, a Digiday survey found that 99% of marketers say that DCO is a significant factor in their marketing efforts.
But in my opinion, new regulations and technological advances have dramatically undercut the value of traditional DCO.
Here’s why:
It’s still manual
On paper, DCO is every performance marketer’s dream. All you need to do is upload your creative assets to Meta, Amazon, or Google and watch as the algorithm automatically generates ads that connect with your target audience.
But the reality of traditional DCO leaves much to be desired. One of the biggest reasons for its lack of complete adoption is that, before automation and optimization can begin, you must first do a lot of heavy lifting:
- Creating endless ad variants from scratch
- Manually resizing assets for every channel
- Uploading assets across every platform
- Chasing internal stakeholders for approvals
- Rebuilding creative mid-flight to account for any insights
That cycle repeats with every new campaign or test you want to run.
The kicker is that when it comes to creative testing, most teams aren’t limited by ideas or data, but rather, the number of variations that can be created in a period. Most designers can only create 100 or 150 variations a month.
With limited bandwidth, teams often opt for safer plays that are quicker to execute—like applying different background gradients to ads that are easy to create and swap—instead of lifestyle-focused and deeply personalized creative, which she says are essential for increasing engagement and avoiding ad fatigue.
User tracking and data is limited
Data is the lifeblood of DCO. It’s what makes engaging ad personalization possible. But to get meaningful outputs, you need a higher volume of meaningful inputs—like your ad creative, for example. That’s where most DCO technology fails to deliver.
With increasing privacy restrictions that limit cross-device tracking and user-level data, platforms only have a few chapters of the story. Marketers can see who clicked, but not why users engaged, which creative aspects resonated, or how different variants performed.
Faced with these constraints, DCO tools often optimize based on the data they can see, such as click-through rate (CTR) and impressions. But these insights fail to tell the whole story. This puts marketers at a strategic disadvantage: Your team doesn’t know what creative worked, what fell short, or what to test next.
Marketers aren’t in control
Between algorithms and walled gardens, there’s a certain amount of control marketers don’t have access to with programmatic advertising.
Ad platforms withhold data and insights, limit personalizations, and make governance difficult. And today’s DCO platforms take even more control by limiting much of the creative flexibility and brand oversight you need.
On the creative side, most platforms only allow for limited personalizations. Marketers are often confined to a few predefined parameters you can personalize, like headline or CTA, instead of being able to fully customize everything.
Then there’s the brand side. I was chatting with a friend, Luke Heinecke, recently. He’s the founder & CEO of Linear and an LP at Pixis.
He pointed out that DCO is really a relinquishing of control: The marketer is essentially handing the wheel to the algorithm. He said, “That’s fine until it ignores your strategy or mismatches headlines with the wrong visuals or priorities.”
That’s where internal friction starts. As performance teams scale up to create as many creatives as possible, brand teams push back to maintain guardrails, guidelines, and consistency.
The marketing team wants every iteration, while the brand team wants guardrails and control. Unfortunately, you can rarely have both.
The Pixis Philosophy: How DCO should work
We’re not claiming DCO is broken, but its limitations—manual workflows, incomplete data, and creative constraints—are holding us back from realizing that dream that looks so enticing on paper.
That’s the belief that drives everything we build at Pixis: that marketers are more inhibited by current tools than helped.
Our goal is to eliminate the manual work at every step—from getting creative recommendations to launching assets that change in real time based on who’s seeing them.
But automating the busywork doesn’t mean eliminating you from the equation. It’s more about freeing you up to focus on higher-value work.
I feel strongly that marketers should focus on strategy and analysis. You shouldn’t spend hours reviewing campaigns and ad sets you’ve already built. Your time is better spent driving new ideas and A/B tests.
Quick aside: that’s why we’ve built an LLM specifically for performance marketers for analysis, insights and analysis. We believe no marketer should ever have to export a csv. All the non-strategy, non-creative, non-empathetic work woul ideally fall to an AI agent like this:
How performance marketing teams get more from DCO with Pixis
That’s our high-level philosophy on DCO, but let’s get down to the nuts and bolts: What does this look like in practice, and how does it help you unlock the technology’s potential?
First, to be clear, we don’t offer a complete DCO solution, and don’t claim to. Not yet, at least!
But because so much of the limitation of true DCO is a creative bottleneck, we put our focus on solving that problem.
Because some people would prefer to watch than read, here’s a demo of Adroom, our generative AI product for ad creative. Feel free to skip if you’d just rather read on to what we consider best practices.
Make every element of your ads dynamic
Does shuffling headlines or product images technically count as DCO? Yes, but to realize the full value of true DCO, marketers should think way bigger than simple swaps in fixed templates (and it does with Adroom).
You need the ability to dynamically adjust every creative layer—product imagery, backgrounds, headlines, CTAs—based on who you’re targeting, where, and the context around them.
When you do that, you open doors to better ad engagement.
For example, some brands are personalizing ads to niche audiences, interests, events, and trending topics. Instead of delivering a generic ad to 25-to-50-year-old men in Boston, you can build tailored versions for microsegments within that audience, such as 25-to-50-year-old men in Boston who are browsing for DIY projects on weekday nights.
Brands like Swiggy are putting this into practice right now. (For my friends elsewhere in the world, Swiggy is like Doordash. They’ve got a market cap of over $12B USD, and process 20 million orders a day.
When they set out to improve the quality of app installs from their Facebook campaigns, we helped them scale creative production and performance. In just six weeks, they identified more than 100,000 distinct personas using Targeting AI, launched 60,000 custom creatives tailored to them, and continuously optimized their efforts toward a single goal: reducing the cost per first transaction.
The result: a 270% jump in installs, and a 41% drop in both cost per install and cost per first transaction.
Scale without sacrificing your brand
Brand consistency is always a challenge, whether you're producing one ad or thousands. With DCO, the challenge of maintaining ad standards is even steeper, especially for organizations in highly regulated markets, like pharmaceuticals and finance.
In other words, more volume means more risk, whether that means brand inconsistencies, not looping in the right stakeholder before something goes live or failing to consider the creative requirements of each channel, like running a text-heavy ad on a visual-first platform like Instagram.
With Pixis, structure matters just as much, if not more, than speed. Not only is every creative variation automatically brand-aware, reflecting fonts, colors, logos, and tone of voice, but built-in approval workflows also ensure that every stakeholder can leave feedback before the ads go live.
Turn data into creative strategy
DCO isn’t only about creating the right ad for the right person and moment. The real opportunity lies in what you learn through that optimization process, and how you turn those insights into better creative. For example:
- What messaging drives specific audiences to act?
- Which images perform best on which channels?
- When are certain CTAs more likely to convert?
This is another area where many marketers hit a wall with DCO in the past: The platforms shuffle assets and create new variants, but you don’t know what worked, what didn’t, or where to focus your efforts.
DCO is no longer a “nice to have”
It used to be that most brands were just testing out DCO.
Now they’re actively looking for solutions like Pixis that help them scale creative production intelligently while maintaining brand consistency across thousands of different dynamic personalizations.
But while AI and automation are helping with personalization, scale, and consistency, one pain point is still impossible to ignore: workflows.
According to the same Digiday survey, 71% of marketers reported that their creative and ad-serving process workflows are only somewhat well-integrated and efficient, while another 28% called theirs somewhat poorly integrated and efficient.
That disconnect matters. A lot. Because as DCO scales and becomes more central to performance marketing strategies, the level of complexity will grow. More variants, more platforms, more stakeholders.
More everything—and without the proper structure, technology, and philosophy, things can spiral quickly.
That’s where Pixis comes in. With AI that optimizes creative based on contextual signals, built-in guardrails that keep your brand consistent, and the flexibility to test and scale what works, you get the structure, control, and agility DCO has always needed.
Want to learn more about how Pixis can help you dynamically scale your creative? Schedule a demo today.
