Why “free” changes what people take — Ariely’s argument, and the chocolate-counter experiments underneath it.
What the evidence says
- Ariely argues that “free” is not simply a very low price but a distinct emotional category, producing a qualitatively different response from even a trivial cost[1]
- On that account people will choose a free item over an objectively better paid one, even where the price difference between the options is identical in both cases[1]
- Applied to food: people take and eat things they would not buy at even a trivial price — the offer, not the appetite, is what initiates the eating[1]
- The effect is measured, not only argued. In real purchases at a chocolate counter, excluding those who took nothing, the share choosing a Hershey’s Kiss rose from 27% (priced 1¢ against a 15¢ Lindt truffle) to 69% (priced 0¢ against a 14¢ truffle), while Lindt’s fell from 73% to 31% (both ps < 0.001)[2]
- The strict version is stranger than “people like free things”: demand shifts when a price falls to zero even though the price difference between the options is held constant, which no cost-benefit account predicts. A third condition discounting the truffle far more steeply (0¢ against 10¢) was statistically indistinguishable from the first free condition (t(142) = -1.0, p = 0.31)[2]
- It is not an artefact of free things being easier to take. A further experiment equalised transaction costs, so every option including the free one cost the same attention and time (232 customers): Hershey’s share still rose from 21% to 71%, Lindt’s fell from 79% to 29% (t(92) = 5.6, p < 0.0001)[2]
- These are choice experiments, not eating experiments. They measure which item people take between two priced goods, one of which becomes free — not appetite, not consumption, not how much is eaten[2]
- The figures above are from Shampanier’s 2007 MIT doctoral thesis; the experiments were published the same year in Marketing Science, the version of record[2]
- A free sample also puts food in a person’s path unplanned, so its effect on eating may run through proximity and visibility as well as through price: an inference joining Ariely’s account to the proximity studies, not a claim either makes[1]