Should I create an original ad creative or replicate a proven competitor format?

Replicate the proven competitor format when you're entering a new category or testing on a limited budget — you're buying a validated structure, not assets. Reserve original concepts for when a winning format fatigues or your brand has a genuinely differentiated angle. At AI production costs (~$75 and under an hour per ad), the real answer is often: ship both and let ROAS decide.

Default to replicating the proven format when you're entering a new category, launching a new product, or working with a limited test budget — a competitor's winning ad is a validated hypothesis about what converts in your market, and rebuilding its structure costs a fraction of proving an original concept from scratch. In one documented production, a winning competitor UGC ad found in the morning was rebuilt as a brand-native version the same day for approximately $75 in under an hour — against $300–$600 per creator and at least two weeks of turnaround through creator marketplaces. The reference ad's product had sold out in 2 hours, which is performance proof no untested original concept carries.

Replicate the structure, never the assets. Deconstruct the hook type, beat sequence, shot timing, camera movement, pacing, and CTA — then rebuild every frame with your own brand, product, characters, and copy. invideo is an agentic video creation tool with all the current generation models available, so you can upload the reference ad to the invideo agent and it breaks it down shot by shot: in one production it detected all 9 cuts in a reference ad and extracted one frame per scene; in another it decomposed an 11-second winner into 7 shots with timestamps, camera movement, and frame composition, explained why the ad converted, identified the hook, and generated 7–8 hook variations adapted to the new brand. Load your brand context before the rebuild — without it, raw generation can't distinguish the elements that drove conversion (which must stay) from the brand-specific elements (which must be swapped). This is also where the legal and strategic line sits: borrowing hook logic, format, and pacing is standard competitive practice; reproducing visuals or copy verbatim is not — every asset in the rebuilt ad should be yours.

Go original when the replication case weakens. Three signals: your current winning format is fatiguing in-platform, you have a differentiated brand angle no competitor format expresses, or the category's dominant pattern has gone stale — polished fake-testimonial UGC no longer converts, while value-first native content does, so copying a tired format buys you a validated loser. Before abandoning a proven format entirely, exhaust its variants first: re-upload your own high performer and generate structural variants across audiences, seasons, and products — real brands run a variant system for A/B testing rather than betting on one finished ad.

At AI production costs, stop treating it as either/or. You can't predict which ad will win without shipping it, so produce the replicated format and an original concept in parallel and let platform data decide. One documented run produced three complete ad formats — a product showcase film, a faceless UGC, and a full-character UGC — for $225 total (890 credits); in a Meta Ads Manager comparison, a person-led UGC ad ran 6.2x ROAS against 0.8x for a product-only cut, the kind of gap you only discover by shipping both. Once either path produces a winner, scale it instead of restarting: documented workflows generate 8–10 variations of a single winning ad in one day, at roughly $30 per product swap and $70–$145 per market localization.

Watch some of these to see what works for you:

It watches the ad, breaks it down shot by shot, figures out what made it work, and rebuilds the whole thing. Storyboard, shot list, generations, final edit.

— invideo's creative team