Scarcity

Evidence

The theoretical basis is commodity theory: a good gains value the harder it is to obtain. The most cited test is by Worchel, Lee and Adewole (1975). The 134 students in the analysis — the sample consisted entirely of women — rated cookies from a jar containing two or ten cookies. The ordering is revealing.

The F tests whether groups differ more than chance would lead you to expect; all three values sit well above the conventional threshold. So scarcity works not only as a state but as a change and as a social cue.

Two findings from the same paper are almost never quoted with it. First, the study also tested whether it helps when many others still want the cookie. It did not, and on estimated price the finding reversed: there the cookie was dearer in the condition with few other takers. Second, the second experiment showed the effects only among participants who had not seen through the purpose of the study. Among the rest the pattern ran the other way, and counting them in left nothing of the effect. The authors write themselves that support for commodity theory may not be obtainable without carefully screening out suspicious participants.

Taste, incidentally, was untouched: asked how good the cookies tasted, participants showed no effects whatever. Scarcity shifts the judgment of worth, not the experience.

Lynn (1991) pooled 41 studies with 51 tests of commodity theory. The result is reliable and small at once. Across all studies the association is clearly different from zero, but the mean effect size is only r = 0.12 — and 0.17 once five studies with clearly reversed findings are removed as outliers. The r measures how closely two quantities go together; 0 means not at all, and values around 0.1 count as small. Individual results range from r = −0.54 to r = 0.43 and come out heterogeneous. Lynn converts the figure into a more tangible form: instead of landing at 50–50, judgment lands at 44 against 56 when something is scarce rather than plentiful.

Two things belong with that. The cookie study is one of the 41 works in this meta-analysis, so it is not an independent confirmation but part of it. And Lynn notes himself that hardly any of the studies used the scarcity cues that actually occur in marketing; he calls for research with more realism. The explanation by way of a need for uniqueness, incidentally, held in only 4 of 11 tests.

Aggarwal and colleagues (2011) separate two forms that practice constantly conflates: quantity scarcity (“only 2 left”) and time scarcity (“today only”). In two experiments with students, quantity scarcity beat time scarcity, both times by about one point on a seven-point purchase-intention scale. The route is the same in both cases: perceived competition with other buyers. A limited number of units produces it more strongly than a deadline, because with a deadline you do not have to get there before anyone else. Shi, Li and Chumnumpan (2020) map the literature in a review of 66 papers published between 1970 and 2017. It is not a meta-analysis and gives no pooled effect size; it sorts the field by four explanations: the need for uniqueness, for conformity, for avoiding later regret, and for freedom of choice.

Two of its findings matter for shops. Scarcity claims work only where customers expect scarcity in the first place; where the claim does not fit the situation, what arises is suspicion about the motive. And reasons given in terms of other people’s demand — “in popular demand”, “over X sold” — count as especially liable to be read as deception, because customers do not credit sellers with being able to quantify demand reliably.

Limits

What is measured is usually valuation, not purchase. The three checked papers bear that out: Worchel records value judgments about cookies, Lynn pools mostly such judgments, and Aggarwal measures purchase intention after a fictitious advertisement shown in a lecture hall. None of them observes a purchase decision with the participant’s own money.

One of the three papers explicitly excludes online shops. Aggarwal and colleagues write that online buying lies outside the scope of their study, because deals are easy to compare there and that is likely to change how scarcity cues work. Anyone citing this study for a shop is citing something it does not claim.

Scarcity has to be plausible. A justification that does not fit the situation — a software download with “only 3 licences left” — does not read as scarce but as dubious. Reactance is the likely mechanism: people who notice the pressure push back against it. That link is not verified.

Repetition devalues it. A shop where everything is always nearly sold out teaches customers that the display means nothing. The effect is then not merely gone; the signal has been spent for the honest case.

Manufactured scarcity is deception. Countdown timers that start over on the next visit, and stock figures with no relation to the warehouse, are not a design decision. Assessing any specific case legally belongs in a professional review.

The evidence is reliable in direction, not in size. Lynn’s meta-analysis supports the effect on average but contains five contrary studies, and the second cookie experiment is shaky. Transfer to shops remains the open question.