Why Does $9.99 Feel Cheaper Than $10?

What changed in the left-digit experiments?

The left-digit effect describes price judgments that give disproportionate weight to a change in the leftmost digit. In their author-written NYU summary, Thomas and Morwitz describe pen-price judgments that differed when the dollar digit changed in $2.99 versus $3.00. They contrast this with one-cent changes that leave the dollar digit unchanged. Authors’ NYU account

This is not the same as literally reading $2.99 as $2.00 or failing to see cents. Seeing a digit and weighting it in a judgment are different things. The study measured how high prices seemed, not how many pens those students purchased.

The bottle comparison also changes the number of digits before the decimal point: $9.99 has one and $10.00 has two. That adds a difference absent from $2.99 and $3.00 when displayed consistently. A familiar example can illustrate a question without reproducing all the controls of a study.

Anderson and Simester’s publisher abstract reports increased demand for $9 price endings in three retail field studies, with stronger effects for new items and some evidence of weaker effects alongside sale cues. This result is tied to those settings, rather than a universal sales rule or proof of one cognitive mechanism. It does not establish that a one-cent change in our bottle example increases purchases or profit. Publisher abstract

A nine ending can also resemble a deal signal. It does not prove a markdown occurred. To know that, you would need the actual earlier price and comparable offers, rather than the shape of the new price tag alone.

Impressions, recall and purchases differ

A meta-analysis found a small average advantage for just-below prices on purchase-related measures, with substantial variation. Those measures combine different outcomes, and alternative analyses and publication-bias adjustments weaken the result. It is not a percentage sales forecast. Troll and colleagues

An experiment with real-money lottery decisions did not find reliable price-ending effects on the tested choices. That does not erase a price-judgment result; it asks a different question in a different task. Fenneman and colleagues

Two preregistered student experiments published in 2026 found price-recall differences without reliable direct effects of price endings on purchase intentions or quality judgments. These results concern those experimental tasks, rather than every retail setting. Remembering a number, intending to buy and actually placing an order are distinct outcomes. Escher and colleagues

Example assumptions & distinctions

The switch holds the fictional 500 mL bottle, type size and terms constant. Prices are before tax with no other charges. It illustrates a comparison, not a perception experiment; reactions are not measured or recorded.

Research on price judgments, recall, intentions and actual purchases examines different outcomes. Neither a nine ending nor a stronger impression establishes a universal sales increase, a real markdown or a useful purchase. Checking arithmetic need not remove the feeling of a bargain; a round price does not settle usefulness either.