Why prices that end in .99 change what people are willing to pay

Where the .99 habit came from

Charm pricing's exact origin is disputed. Wikipedia's summary traces the practice to the late 19th century, though it notes the historical evidence suggests 9-ending pricing only became widespread in the 1920s, not earlier, as some popular accounts of the tactic's history claim.

What has stayed remarkably stable is how common it is. The 1997 Marketing Bulletin estimate of roughly 60% of advertised prices ending in 9 sits within a percentage point of Capital One Shopping's more recent figure of 60.7%, which also found another 28.6% of prices end in 5 — meaning fewer than one in ten prices today end in a round number. The share ending in 9 has barely moved in that time.

The brain reads the first digit, not the last

Manoj Thomas and Vicki Morwitz named this pattern the left-digit effect in a 2005 Journal of Consumer Research paper. Across five experiments, they found people encode multidigit prices by anchoring disproportionately on the leftmost digit, so a price like $2.99 is processed as roughly "2 point something" rather than as one cent short of $3 — the digit before the decimal point does most of the psychological work, no matter how small the true gap it hides.

The effect is not constant. Thomas and Morwitz found it is strongest when the leftmost digits of the two prices being compared actually differ, as with $2.99 versus $3.00. It is much weaker when the leading digit is the same on both sides, as with $3.95 versus $3.99. They also found the bias gets stronger the closer two prices sit to each other, since shoppers are more likely to run an active side-by-side comparison at that distance — precisely the moment a one-cent gap can be read as a far bigger one.

How much a single cent can move a choice

Ipsos researchers tested the mechanism directly using Price Sensitivity Meter methodology across nine separate studies, covering 625 price points and 4,865 respondents, on categories including fast food and beverages priced between $0.49 and $10.00. They found a price change from 99 cents to the next whole dollar produced a significant jump in how expensive people perceived the item to be, and that a shift from $2.49 to $2.50 produced a similar, though smaller, jump in perceived expensiveness. Ipsos's applied conclusion for brands already seen as pricey was that an $X.99 or $X.49 ending is preferable to a round number, since round endings disproportionately signal higher cost in their data.

A 2009 Journal of Consumer Research experiment by Kenneth Manning and David Sprott, described in a press summary on ScienceDaily, showed how far that perception can swing an actual choice. Offered a $2.00 pen against a $3.99 pen, 44% of participants picked the pricier one. Offered the same trade-off reframed as $1.99 against $4.00 — a one-cent shift on each price — only 18% did. Moving which item carried the .99 ending, not the underlying gap between them, more than doubled the pricier item's share of choices.

What happened when retailers tested it in the field

Eric Anderson and Duncan Simester tested the effect outside the lab, running a series of field experiments and publishing the results in Quantitative Marketing and Economics in 2003. A $9 price ending increased demand in all three experiments they ran, confirming that the bias translated into real purchase behavior rather than staying a quirk of how people answer survey questions.

One detail in that result matters more than the headline. The demand increase was stronger for new items than for items the retailer had sold in previous years, which the researchers attribute to customers having less independent information about what a new item is worth — they lean harder on the price-ending cue when they have nothing else to go on.

That pattern helps explain a puzzle in plain sight: if 9-endings reliably lift demand, why don't retailers use them on every single item? The Kellogg summary of the research points to the answer — the ending's value depends on how much the buyer already knows.

Moving which item carried the .99 ending, not the underlying gap between them, more than doubled the pricier item's share of choices.

Why some prices round up on purpose

Stiving's 2000 Management Science paper, "Price-Endings When Prices Signal Quality," starts from the opposite premise: firms that use a high price to signal quality are more likely to set that price at a round number instead of a 9-ending. The model treats the price ending itself as a cue, separate from the number's raw value, that shoppers read alongside the price.

Wikipedia's summary of the broader literature backs up the same pattern from the retail side: restaurants and high-end retailers often price in even numbers specifically to increase brand image, treating a 9-ending as a signal that would undercut the quality claim the price is meant to make. The Manning and Sprott pen study found something consistent with this — the left-digit effect's pull on choice disappeared when people were buying gifts for close friends or making small-dollar purchases, contexts where squeezing out the cheapest option mattered less than in the earlier pen and catalog examples.

Matching the ending to what the price has to prove

Put these findings together and a simple test emerges for a price that has not yet been set. Charm pricing does its most reliable work on value-positioned goods, where the buyer has little independent basis for judging worth — new SKUs, unfamiliar categories, first purchases. That is exactly the condition under which Anderson and Simester found the largest demand lift, and it lines up with Manning and Sprott's finding that reframing which price carries the .99 ending can more than double the pricier item's share of choices.

Round pricing does more work where the number itself is part of the pitch for trust or quality — premium retail, professional services, anything sold on the strength of the price rather than in spite of it. How big any charm-pricing lift actually is varies a great deal by study and category: the research Capital One Shopping compiled cites estimates ranging from roughly 24% to 60%, a spread wide enough that no single percentage should be treated as a guarantee for a specific product. The right test is against the item's actual condition and what the price is being asked to prove, not a rule of thumb.