A bid is not a purchase. It is a sentence in an argument about what something is worth, and the reply comes from a stranger who has been watching the same hands hold the same bowl.
That gap — the few seconds after a bid lands, while everyone else decides — is the whole mechanism. It is also the part that is easiest to ruin.
The two ways to ruin it
Let the clock run out mid-argument. Two people are ten seconds from agreeing on a price and the timer ends it at an arbitrary moment. The seller is short-changed and the under-bidder goes away feeling cheated by a stopwatch. So on most items a bid landing late pushes the clock back out, and the auction ends when the bidding stops rather than when a number reaches zero. A seller who wants the opposite can choose a sudden-death format instead, where the clock is short and never moves — and the item says so before you bid.
Show people a price that is not real. A bid you place has to be a bid everybody sees, in the order it happened, with the same number on every screen in the room. That sounds obvious and it is the hardest part of the build — a hundred phones on a hundred connections, and the person on hotel wifi must not be bidding against a price from four seconds ago.
What we will not do
We will not let a seller bid on their own work, through any account. The whole value of a room full of bidders is that the price is somebody's real opinion.
We will not hide who is winning. You can see that you have been outbid, immediately, which is the only thing that makes the next ten seconds mean anything.
And when it is over
The hammer falls and the money does not go anywhere yet — we hold it until the thing arrives. The seller finds out they have sold it. You find out what you are paying and when it is coming.
Then they pick up the next piece.
Related: why a show has an ending.