Correction, 2026-10-01: an earlier version said DCO assembly happens in under 100 milliseconds, that DCO sits inside every major platform's automated ad products in 2026, and that Meta's Advantage+ creative adjusts contrast and converts aspect ratios per placement. We could not source those. Meta's own Advantage+ creative page lists enhancements such as image and video expansion, image animation, background generation for catalog ads and text variations, so the page now uses that list. The closing line also suggested batch generation; 8frame has no batch tool.
Dynamic creative optimization, or DCO, is an ad technology that assembles a personalized ad in real time by combining modular creative elements, headlines, images, CTAs, backgrounds, and product shots, choosing the combination predicted to perform best for each individual viewer and context.
The key idea is that you don't build a finished ad. You build a library of interchangeable parts and let a machine-learning engine construct the ad, per impression, from those parts. Instead of one static banner that every viewer sees, DCO produces many possible variants and serves the one it predicts will work for the person looking at it. It has been a core capability of programmatic advertising for years, and the big ad platforms now offer automated versions of it, such as Meta's Advantage+ creative.
How dynamic creative optimization works
DCO has three moving pieces: a library, a decision engine, and a feedback loop.
The element library holds your modular assets, sorted into slots: backgrounds, hero images, headlines, CTAs, prices, product shots. For the system to work, these have to be genuinely combinable. Any headline needs to read cleanly on any background under any CTA. If the elements aren't designed as modules, DCO can't recombine them without producing broken-looking ads.
The decision engine does the assembly. When an impression becomes available, the engine reads the available viewer and context signals, queries the library, selects the combination it predicts will maximize the desired action, and assembles a complete ad. This happens during ad serving, so nobody sees the engine choosing. They just see an ad that happens to be assembled for them.
The feedback loop is what makes it "optimization" rather than just "dynamic." Performance data from served combinations flows back into the engine, which learns which combinations work for which audiences and shifts future assembly toward the winners. Over time it converges on the strong combinations and stops serving the weak ones.
The historical bottleneck was library depth. A shoot produces only so many usable elements, so most DCO libraries stayed shallow and the personalization stayed shallow with them. AI generation loosens that constraint, which is why DCO is getting renewed attention.
When you use dynamic creative optimization
Retargeting with product feeds. The classic DCO use case. An ecommerce catalog feeds the library so the engine can assemble an ad showing the exact product a viewer browsed, with pricing and availability pulled live. Every viewer sees a different assembled ad.
Large prospecting audiences with segment variation. When your audience spans many segments and the winning message differs by segment, DCO can find per-segment combinations at the element level that you'd never manually guess. It optimizes within the message space you defined.
Multi-market or multi-placement campaigns. DCO handles versioning across languages, markets, or placements by swapping elements rather than rebuilding whole creatives. Meta's Advantage+ creative also lists image and video expansion among its enhancements.
DCO is the wrong tool when you're still searching for the concept itself. It optimizes within a system of parts. It can't tell you the whole angle is wrong.
Examples
A furniture retailer's retargeting campaign (illustrative). The DCO library holds product shots pulled from the live catalog, three background room settings, four headlines, and three CTAs. A shopper who viewed a specific sofa gets an assembled ad pairing that exact sofa with a living-room background, a benefit-led headline, and a "Shop now" CTA, all built when the impression loads. A shopper who viewed a desk gets a different assembly entirely, from the same library. The retailer produced roughly a dozen elements; the engine serves hundreds of combinations.
A DTC brand on Meta Advantage+ creative (illustrative). The brand uploads 40 AI-generated backgrounds, 12 product hero shots, and a set of headlines and CTAs. Meta's system combines the elements, can apply enhancements such as image expansion, image animation and text variations if the brand leaves them on, and optimizes toward early engagement signals. The brand never assembles a single finished ad by hand. It supplies the parts and the objective and reads the results to see which pairings won.
FAQ
Can 8frame run DCO? No. 8frame generates the elements; the assembly and optimization happen in the ad platform or a DCO vendor.
What does a library of AI backgrounds cost on 8frame? Nano Banana 2 is 11 credits an image at 1K and returns up to four images a run, so 40 backgrounds are 440 credits (about $4.40 at pack rate).
Related concepts
- Dynamic creative optimization with AI is the hands-on workflow: how to generate a deep, modular element library with Nano Banana Pro and Kling v3 and load it into Advantage+ or a programmatic DCO platform.
- Creative volume is the input DCO depends on. Shallow libraries produce shallow personalization; the multiplication that makes DCO worthwhile only happens when the element library is deep.
Sources
- Meta, Advantage+ creative (lists image expansion, video expansion, image animation, background generation for catalog ads, text generation and variations; page served in German): https://www.facebook.com/business/ads/meta-advantage-plus/creative, checked 2026-10-01
- Nano Banana 2 price and outputs per run; credit value at pack rate: 8frame tool registry and billing catalog, read 2026-10-01
DCO is only as good as the library feeding it. Generate modular backgrounds and hero shots from one product reference on 8frame, then read the full AI DCO workflow.
Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.