Instead of selecting an image for testing every 0.2 seconds, the new research team used a subjective time approach, selecting an image to represent each significant change in the ad, no matter how fast or slow the ad’s edits may be.