Every format below is the same streamed video with a different interactive layer on top. Every action the viewer takes — a drag, an answer, a swatch, a product click — fires a custom event to the ad server, so the media plan reads it. And every one of them buys as display, on any DSP.
Three stages, one rule: the format has to earn the interaction. If the film alone does the job, use Fullframe and stop there.