The rule that beats the better offer

An auction was running. Twenty bids, two days left, sitting at forty-one dollars. A buyer messaged with an offer: fifty dollars, including shipping, if we did the deal directly.

Good buyer. Nine years on the platform, perfect feedback. Nine dollars over the current price, right now, no waiting.

My agent brought it to me with an analysis. The auction was already at forty-one with two days to run and real bidding momentum, so the expected final price beat fifty. Recommendation: decline, on the numbers.

The analysis was right. The recommendation was right. And the whole exchange was a mistake, because the numbers were never the question.

Why the answer cannot be economic

A sale arranged off-platform has no seller protection. Payment reverses and there is nothing to appeal to. The listing has no record of the deal. And the platform reads those messages, so proposing it risks the account that the entire operation depends on.

So the answer is no when the offer is nine dollars better. It is also no when it is fifty dollars better. There is no number.

The reply I sent was five words: never sell outside the platform.

The difference between a recommendation and a rule

The agent had treated this as a judgment call and done a good job of the judgment. That is exactly the failure.

A judgment call gets re-made every time. It arrives at whatever answer the numbers support, so the day someone offers double, the analysis flips and the recommendation flips with it. And it costs me attention every single time, which is its own tax.

A rule gets applied. It costs nothing, it does not drift, and it produces the same answer on a bad day as on a good one.

What went in the charter:

Any buyer proposing a price off-platform, a direct payment, a “cancel and I’ll buy it direct,” or a sale after an auction ends unsold: decline or ignore. Never counter, never negotiate the number, and never bring the offer to me as though it were a live question. The answer is already given.

That bolded clause is the part that changed behaviour. Without it the agent still surfaces the offer, framed neutrally, which puts the decision back on my desk. A rule that still requires me to say no has not saved me anything.

How to tell which one you have

Ask what would change your mind. If a number exists that flips the answer, it is a judgment call and the agent should get a framework. If nothing flips it — because the downside is categorical rather than quantitative — it is a rule, and the agent should never be running the analysis in the first place.

Most things are judgment calls. But the handful that are not tend to be exactly the ones where an agent’s analysis is most persuasive and most beside the point.

The precision it needs

I widened this agent’s remit later, which meant the rule needed rereading. It is not “this platform is the only place anything may ever be sold” — finding new channels is literally its job. It is narrower: a sale that originated on the platform gets completed on the platform.

An overbroad rule gets worked around, because at some point it obviously contradicts the job. Better to write the narrow version and let a different rule handle the rest.