Why do prices ending in 99 work?
A British supermarket feared the ending would cost it its quality reputation. The research points to how people read numbers, and to retailers using the tool less than shoppers respond to it.
Answer to the question
Why do prices ending in 99 work?
Prices ending in 99 work because shoppers weigh the leftmost digit most heavily, so 2.99 reads as nearer two than three. Thomas and Morwitz (2005) found the effect only when the leftmost digits differ, and field experiments found that nine endings raised demand. The size of the effect varies by item and context.
Key takeaways
- Thomas and Morwitz (2005) ran five experiments showing a nine-ending price is perceived as smaller than a price one cent higher only when the leftmost digits differ.
- Lacetera, Pope and Sydnor (2012) analysed more than 22 million wholesale used-car transactions and found sale prices drop at 10,000-mile odometer thresholds, so the bias is not limited to price tags.
- Anderson and Simester (2003) ran three field experiments in which a nine ending raised demand each time, more for new items and less alongside sale cues.
- Strulov-Shlain (2023), using scanner data on 3,500 products from 25 US chains, estimated that shoppers treated a one-cent rise from a 99 ending as more than a 20-cent rise.
- Shotton (2023) reports a small street test of 79p against 80p chocolate in which taste ratings of 7.1 and 7.6 did not differ significantly.
A major British supermarket once held a firm view that prices ending in 99 looked tacky and would cost the chain its reputation for quality. It would not pay for a study to find out. Richard Shotton, in The Illusion of Choice, describes how he and a collaborator ran a cheap one instead. Shoppers on a busy London street tasted chocolate presented as a new brand launching in the UK, priced at either seventy-nine or eighty pence, and rated the taste out of ten. The gap in ratings was not statistically significant.
Shotton notes that its direction did not support the supermarket's fear. That small test touches one half of the question. The other half is why prices ending in 99 work at all, and for that there is a mechanism, a set of field results and a puzzle about how retailers use them.
Shoppers read numbers from the left, and cars show it
The best-supported explanation is about how a number is read. Manoj Thomas and Vicki G. Morwitz ran five experiments, published in the Journal of Consumer Research in 2005, on when a price ending in 99 is perceived as smaller than the price one cent above it. They report that the effect appears only when the leftmost digits of the two prices differ, and that it depends on how close the compared prices are, both numerically and psychologically. It also appears with other multidigit numbers, so it is not confined to money.
The rule is easy to apply. A price of 2.99 sits below 3.00 in leftmost digit, so it is read as noticeably smaller. A move from 3.59 to 3.60 leaves the first digit alone, and the rule gives no reason to expect a gap. A price ending in 99 is simply the highest price that still keeps the lower first digit. Why the First Price You See Decides What You Pay covers a related pull of the first number seen.
The cleanest test sits outside the shop. Across more than 22 million wholesale used-car transactions, sale prices drop discontinuously at 10,000-mile odometer thresholds, with smaller drops at 1,000-mile thresholds. Nicola Lacetera, Devin Pope and Justin Sydnor read this as left-digit bias in how buyers process odometer values, and they report that the pattern appears to come from final customers rather than professional agents, and that it shows up in a market with large stakes and easily observed information.
No seller chose those digits to flatter a price. A mileage figure is not a price ending, and no seller can tidy it, yet the first digit still carries more weight than its place in the number warrants. That fits the claim that the mechanism belongs to number reading, with prices ending in 99 as one place where it shows.
Demand moves in the field, and the fear is separate
Reading a number differently would not matter to a retailer unless demand followed prices ending in 99. In three field experiments reported by Eric T. Anderson and Duncan Simester in 2003, a retailer varied price endings and a price ending in 9 raised demand every time. The effect was stronger for new items than for items the retailer had sold in earlier years, and it was weaker when the retailer also used sale cues. The authors suggest that nine endings may work best when customers have limited information. The experiments establish a direction for prices ending in 99, and the size of the gain varied with the item and with the cues around it.
The left-digit account explains why prices ending in 99 look smaller. It says nothing about what a shopper infers about the shop, which is the supermarket's concern. Shotton's test addressed that concern with taste ratings, not purchases. The average ratings were 7.1 and 7.6, and he does not give the number of shoppers, so it counts as one small observation. The pair he chose, 79p and 80p, does cross a leftmost-digit boundary in the pence, which makes it a fair setting for the rule above. Precision is a different property from a nine ending, and The Price That Isn't Round covers that side.
On our reading, a direct test of the supermarket's worry would ask shoppers to judge the shop itself, not one product's taste, across price endings. Shotton's design does not do that, and the studies above do not either. They measure how big a price looks and how much is bought, and both of those point the same way for prices ending in 99.
| Study | Setting | What it reported |
|---|---|---|
| Thomas and Morwitz (2005) | Five experiments | A nine-ending price is read as smaller than a price one cent higher only when the leftmost digits differ |
| Lacetera, Pope and Sydnor (2012) | More than 22 million wholesale used-car sales | Sale prices drop at 10,000-mile odometer thresholds |
| Anderson and Simester (2003) | Three field experiments | A nine ending raised demand each time, more for new items and less alongside sale cues |
| Strulov-Shlain (2023) | Scanner data on 3,500 products from 25 US chains | A one-cent rise from a 99 ending treated as more than a 20-cent rise |
Retailers use it less than the data suggest
Scanner data records what shoppers bought, not what they said, and the sample in Avner Strulov-Shlain's 2023 paper in the Review of Economic Studies was 3,500 products sold by 25 US chains. From it he estimated that consumers respond to a one-cent increase from a price ending in 99 as if it were an increase of more than 20 cents.
Retailers, he found, set prices as though the bias were much smaller than the demand they face reflects. He describes their pricing as closer to rules of thumb than to optimisation.
Shops that use prices ending in 99 are therefore using a tool whose pull on shoppers they have understated. His estimate is that retailers forgo between 1 and 4 per cent of potential gross profits as a result.
Common questions
Why do prices ending in 99 work?
Prices ending in 99 appear to work because shoppers weigh the leftmost digit most heavily, so 2.99 reads as nearer two than three. Thomas and Morwitz (2005) found the effect only when the leftmost digits differ. Field experiments by Anderson and Simester (2003) found that nine endings raised demand, though the size varied by item and context.
Is it only a trick used in shops?
No. Lacetera, Pope and Sydnor analysed more than 22 million wholesale used-car transactions and found sale prices drop at 10,000-mile odometer thresholds, which they attribute to left-digit bias. Thomas and Morwitz also report the effect with other multidigit numbers. That points to a general feature of reading numbers, with prices ending in 99 as one use of it.
Do prices ending in 99 make a product look cheap in quality?
The evidence here is thin. Shotton reports a small street test of chocolate priced at 79p or 80p in which taste ratings did not differ significantly, but it measured ratings rather than purchases and gives no sample size. No claim can be made about every category from that single result.
Sources
Shotton, R. (2023). The Illusion of Choice. Harriman House. Seed book; vault hub Books/The Illusion of Choice.
Thomas, M., & Morwitz, V. (2005). “Penny wise and pound foolish: The left-digit effect in price cognition.” Journal of Consumer Research, 32(1), 54-64. doi:10.1086/429600 https://doi.org/10.1086/429600
Lacetera, N., Pope, D. G., & Sydnor, J. R. (2012). “Heuristic thinking and limited attention in the car market.” American Economic Review, 102(5), 2206-2236. doi:10.1257/aer.102.5.2206 https://doi.org/10.1257/aer.102.5.2206
Anderson, E. T., & Simester, D. I. (2003). “Effects of $9 price endings on retail sales: Evidence from field experiments.” Quantitative Marketing and Economics, 1(1), 93-110. doi:10.1023/A:1023581927405 https://doi.org/10.1023/A:1023581927405
Strulov-Shlain, A. (2023). “More than a penny's worth: Left-digit bias and firm pricing.” The Review of Economic Studies, 90(5), 2612-2645. doi:10.1093/restud/rdac082 https://doi.org/10.1093/restud/rdac082