Ad formats are not static. They evolve as platforms evolve, as user behavior shifts, and as technology enables new experiences. The brands that adopt new formats early capture a window of attention that late adopters never get. The first brands to use Instagram Stories ads got 50% lower CPMs than they pay today because the format was underpriced while adoption was low. The same pattern repeated with Reels, with TikTok Spark Ads, and with every new format that platforms introduce.
The evolution follows a predictable cycle. Phase 1: the platform introduces a new format. Early adopters get cheap reach because supply of ad inventory in the new format exceeds demand. Phase 2: adoption increases, CPMs normalize, and the format becomes standard. Phase 3: the platform introduces an even newer format, and the cycle repeats. The brands that participate in Phase 1 consistently get better economics than the brands that wait for Phase 2.
The current Phase 1 formats worth attention are: interactive carousel ads on Meta (where each card can have a different CTA), shoppable video ads on TikTok (where products are tagged directly in the video), conversation ads on LinkedIn (where the ad includes branching responses), and performance max creative sets on Google (where the algorithm mixes and matches creative elements). Each of these formats is underpriced relative to standard formats, and each rewards early experimentation.
The production challenge with new formats is that they require different creative approaches. An interactive carousel needs a narrative arc across 5-10 cards. A shoppable video needs product placements that feel organic. A conversation ad needs branching copy trees. A performance max set needs modular creative elements that work in many combinations. These are different creative products with different production requirements, and they are difficult to produce manually at scale.
This is where AI generation becomes particularly valuable for format evolution. A generation engine that understands the conventions of each format can produce format-specific creative without the founder learning each format's production requirements. You select the format, the engine applies the format's conventions, and you get creative that works within that format's structure. The format-specific knowledge is encoded in the system, not in your head.
I have tracked format adoption across the Downshift portfolio for two years. The pattern is consistent: we adopt new formats within 2-3 weeks of launch, run test campaigns with AI-generated creative, and either scale or abandon based on performance data. The early adoption window, typically 3-6 months, consistently delivers 30-50% better CPMs than the same format after mainstream adoption. Over a year, those savings compound into a significant budget advantage.
The strategic question is: how do you know which new formats to adopt? Not all new formats succeed. Some are abandoned by the platform. Some never gain advertiser adoption. Some are effective for specific verticals but not others. The answer is cheap testing. Generate creative for the new format, run a small test campaign ($50-$100), measure performance. If it works, scale. If it does not, move on. The cost of testing is trivial compared to the cost of missing a format that works.
There is a longer-term trend worth noting: ad formats are becoming more interactive and less static. The trajectory is from passive (user sees the ad) to active (user interacts with the ad). Static images are passive. Carousels are partially active. Shoppable videos are fully active. AR try-on ads are immersive. Each step along this trajectory increases engagement and, generally, increases conversion rates. The brands that can produce interactive creative at scale will have a structural advantage over brands that can only produce static creative.
The format evolution also creates a complexity burden. If you need to produce creative for 4 platforms, each with 3-5 active formats, you need 12-20 different creative products per campaign. Manual production of 12-20 different creative products is a full-time job. AI generation of 12-20 different creative products is a Tuesday morning. The complexity burden of format evolution is manageable with AI and unmanageable without it.
The future is format-fluid creative: creative assets that adapt their structure to whatever format they are placed in, without human intervention. A single creative concept, expressed as a static image on Instagram, a carousel on LinkedIn, a shoppable video on TikTok, and a responsive display ad on Google. Same message, same brand, four completely different creative products. That future is closer than most founders think.
Mani tracks format evolution and updates generation capabilities as new formats emerge. When a platform launches a new format, we add support within weeks. Your Brand DNA stays the same. The format-specific generation rules adapt. You benefit from new format economics without learning new production skills. That is the advantage of platform as a service: the platform evolves even when you are focused on building your product.
The production infrastructure needs to evolve with formats. A generation engine that only produces static images will miss the shift toward interactive and video formats. The engine needs to be format-aware and format-expandable, adding support for new formats as they emerge without requiring the user to learn new production skills.