Why a soft close beats sniping — and what it costs the house

A hard close is the auction equivalent of a fire alarm. Everybody stands around doing nothing for six days, and then in the final four seconds the entire sale happens at once, decided by whoever had the fastest connection and the steadiest hand. It is an exciting way to run an auction and a terrible way to discover a price.

A soft close fixes it with one rule: a bid inside the final window pushes the close out to that window from the bid. Ours is two minutes. There is no cap on how many times it can fire.

What actually changes

The obvious change is that sniping stops working. The less obvious one, and the reason we run it, is that bidders behave differently when they know it. A bidder who expects to be sniped bids late and defensively. A bidder who knows the clock will wait for them bids their real number earlier, because there is no advantage in holding it back.

That is the whole argument. Soft close does not raise prices by adding drama; it raises them by removing the incentive to hide your maximum.

What it costs us

Staff time, mostly. A hard close finishes when the schedule says it finishes. A soft close finishes when the bidding stops, which on a busy evening means somebody is still watching the board at half past nine because two people are trading twenty-five dollar increments on a storage unit in Berwyn.

It also breaks the tidy staggered schedule. We list lots to close two minutes apart; a contested lot in the middle of that run pushes into the next one, and then the one after. Our board sorts by time-to-close for exactly this reason — the schedule is a plan, and the clock is the truth.

The objection we hear most

"It could go on forever." In principle, yes. In practice, no: every extension costs the bidders another increment, and increments step with price. A lot at $19,500 moves in $500 steps. Two people cannot trade indefinitely at $500 a go — one of them reaches their number, usually within four or five extensions.

The longest extension chain on our floor in the listed period was five, on a 10x15 storage unit that finished ten minutes late at $1,450 after forty-one bids. Nobody complained.

If you are building one

Two implementation notes, learned the hard way. First: extend from the bid, not from the original close — otherwise a bid at 1:59 remaining buys one extra second. Second: the extension has to be visible on the board immediately, on every surface showing that lot. A bidder who thinks a lot closed and finds out later that it did not will not bid with you again.