A “fake door” or “painted door” experiment uses the analogy that we build a fake door to see how many people open it, which is a good proxy for the efficacy of that door.
In TOCE, they discuss adding a coupon field to a checkout process that always returned “invalid coupon code” to see how that negatively affected conversion. When they ran that through experimentation, they had an understanding of how much money coupon codes would need to make to offset the conversion loss.