Scarcity: Why the Last Cookie Tastes Better

The experiment is almost too neat. In 1975 Stephen Worchel, Jerry Lee and Akanbi Adewole gave people a chocolate chip cookie from a glass jar and asked them to rate it. Some jars held ten cookies. Some held two. The cookies were identical. The cookies from the nearly empty jar were rated as more desirable, and people said they would pay more for them.

Then the researchers added the condition that makes the study worth remembering. For a third group, the jar started with ten cookies, and partway through the experimenter swapped it for a jar of two. Newly scarce cookies were rated higher than cookies that had been scarce all along. And when the reason for the swap was that other raters had wanted them, the ratings went higher still than when the swap was explained as a mistake by the supplier.

So scarcity does not simply raise value. Change in availability raises it more than the level of availability, and scarcity caused by competition raises it more than scarcity caused by accident.

Two mechanisms are running. The first is inferential and fairly reasonable: rare things are often rare because they are good, so rarity is a usable proxy for quality when you have nothing better. The trouble is that the proxy fires whether or not the rarity has anything to do with quality, and it is trivially easy to manufacture. The second mechanism is Jack Brehm's psychological reactance. When a freedom is threatened — including the freedom to have something later — people experience a motivational surge to reassert it, and the object of the lost freedom becomes more attractive in the process. Reactance explains why the competition condition amplified the effect: another person wanting the cookie is a more credible threat to your access than a supply error.

Once you know the shape, the commercial infrastructure built on it becomes conspicuous. The countdown timer on a checkout page, which resets when you reload. The stock counter reading only three left in your size. The limited drop, whose limit is a decision rather than a constraint. The seat map that tells you eleven other people are looking at this flight. The exploding offer that expires at midnight for no reason connected to anything in the world. All of these are the cookie-jar swap, executed as software.

They work because they are answering the wrong question for you. Scarcity is information about how much of a thing exists. It is not information about whether you want it. Those two questions feel similar under time pressure and are completely unrelated, and every deadline in a sales process exists to keep you from noticing the difference.

The observable tell is that manufactured scarcity is almost always paired with urgency and almost never with an explanation. A genuine constraint has a cause that someone can state: the venue holds four hundred people, the print run was two thousand, the harvest was poor. Manufactured scarcity is asserted, not explained, and if you ask why the deadline is midnight you will get a reason about the offer rather than a reason about the world.

The countermeasure follows directly from Worchel's result. Scarcity acts on your desire, not on the object. So the question to answer is not whether you can still get it, but what you would pay for it if there were a warehouse full — and whether you wanted it at all before you learned that you might not be able to have it.