Charm pricing
A Socratic walk-through of charm pricing — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a price ending in 99 sell better than the round number just above it?
A penny is a penny. Between 4.99 and 5.00 there is a difference no shopper would stoop to pick up off the pavement, and every shopper alive knows the trick is being played on them. Yet retailers keep paying that penny, on every unit, for decades.
So either they are all wrong together, or something is happening that surviving the trick's obviousness does not prevent. Which is it — and if the effect is real, what exactly is the shopper responding to?
Reasoning it through
REASONING #Start with how a price is taken in. Nobody parses a price tag the way they would check a bill. The eye moves left to right, and a rough sense of magnitude forms from the leading digit before the rest of the number has properly arrived. Four-point-nine-nine is filed as "four-something", and the filing has happened before the nines land.
If that is what is going on, the account makes a sharp prediction: the effect should hinge on the leftmost digit changing, not on the ending being a 9. Going from 3.99 to 4.00 crosses a boundary; going from 3.44 to 3.45 does not, though the penny is identical. Broadly, that is what studies of the left-digit effect find. It is a clean, mechanical story, and it makes charm pricing a species of misreading.
But a rival story fits the same facts. Endings in 9 are a convention. Decades of use have made them a signal — this is a value price, this shop competes on price, this item has been marked down. On that account the shopper is not miscounting at all; they are reading a code, correctly, and adjusting their estimate of whether the thing is a bargain.
How would we tell these apart? The two accounts diverge on context. Truncation should work anywhere, since it is a property of how digits are read. Signalling should weaken wherever the message is redundant or wrong.
Field experiments give the sharpest evidence available. Anderson and Simester, working with mail-order catalogues, printed genuinely different prices for the same item in otherwise identical catalogues, and found demand higher at $39 than at $34 — more sold at the higher price, which no ordinary demand curve permits. That is charm pricing doing real work rather than a laboratory artefact. Tellingly, they also found the 9-ending's advantage shrank when the item was already flagged as being on sale. If the ending's job is to whisper "this is a discount", it has nothing left to say once a sign is shouting it.
Now run the other direction. If a 9-ending signals "cheap", it should sometimes be a liability. It is: research on how numbers feel — Wadhwa and Zhang's work is the usual reference — finds round prices do better for purchases made on feeling rather than calculation, and you will notice that perfume, jewellery and fine dining rarely price in 99s. There, the convention is working against the seller, which is exactly what a signal can do and a misreading cannot.
The analogy
THE ANALOGY #A price tag is read like a headline, not like a contract. The leading words fix what the story is about, and everything after is small print you technically saw and did not use.
a headline is written by someone trying to summarise honestly, whereas the shortening here is the reader's own and the seller merely arranges the digits to exploit it — and a headline carries no house style, while an ending in 9 has itself become a message about what kind of shop you are standing in.
Clarifying the model
THE MODEL #The refinements matter more than the headline claim here, because the headline claim is routinely overstated.
The effect is real, replicated in the field, and modest. It is not the difference between selling and not selling; it is a shift in the odds, whose size varies a great deal by category, by shop, and by what else is on the label. Popular write-ups quote dramatic percentage lifts from single studies as though they were constants. They are not, and the honest summary is that magnitudes are inconsistent across contexts.
The two accounts are also not rivals in the way they first appear, and the evidence does not cleanly pick one. Truncation cannot easily explain why the ending stops helping once a sale sign appears; signalling cannot easily explain effects that turn on the leading digit changing in settings with no discount convention to read. The likeliest reading is that both operate, with their relative weight set by how much the shopper is calculating and how strong the local convention is.
Which yields the practical corollary, and it is the one most often missed: the recipe does not generalise. Where the buyer is judging value, a 9-ending helps a little. Where the buyer is judging quality or buying on feeling, a round number often does better, and the same digits that read as a bargain read as cheap. Charm pricing survives not because the effect is large but because a penny is a trivial price to pay for a small edge repeated across millions of units.
A picture of it
THE PICTURE #How to readEach point is a buying situation, not a product. Read across for how the buyer is judging — on feeling at the left, on price at the right — and up for whether the odd ending or the round number does better. Top right is charm pricing's home ground, where the catalogue experiments were run. Bottom left is where it backfires, and the perfume counter is priced accordingly. The two interesting points are in between: an item already advertised as reduced drifts down because the ending's message is redundant, and a house asking price sits low on the right because precision there signals a considered figure rather than a bargain.
What became clearer
WHAT CLEARED #Charm pricing is not one effect but two overlapping ones — a genuine quirk in how a number is read from the left, and a learned convention about what a 9 means — and the field has not cleanly separated them. What is well established is smaller and more interesting than the folklore: the penny buys a modest, context-dependent edge, it can be worth having at scale, and in the parts of the market where buying is done on feeling it is worth avoiding.
Where to go next
ONWARD #- Why a precise number ("£412,750") anchors a negotiation more firmly than a round one.
- How the same left-to-right reading shows up in mileage, salaries and exam marks, where nobody is selling anything.
Key terms
TERMS #| Term | What it means |
|---|---|
| Charm pricing | setting a price just below a round figure, most often ending in 9, in the expectation of higher demand. |
| Left-digit effect | the tendency for a number's perceived magnitude to be fixed disproportionately by its leading digit. |
| Price ending | the final digits of a price, treated here as a signal in its own right rather than as part of the amount. |
| Field experiment | a trial run in a real market with real buyers, as opposed to a laboratory judgement task. |
Every term the collection defines is gathered in the glossary.