The Math Nobody Does: Why 'Buy 3, Pay 2' Convinces Even With a Small Discount

The offer is clear: buy three, pay two. The arithmetic should be too. For someone who was already going to buy three units, it's a one-third discount. For someone going to buy two, it's an invitation to spend the same money and walk out with more goods—which isn't a discount at all on what they intended to spend. For someone going to buy one, it's a proposal to triple their expense in exchange for a lower per-unit price on a product that might sit in the cupboard for months.

So the number announced is rarely the number that matters. The actual discount depends entirely on information the label doesn't have: how many units that person would have bought if the promotion didn't exist. And yet the offer works—even for people planning to buy just one. The interesting question isn't whether people can do the division. It's why the division isn't what they're doing.

Deal pleasure counts separately

One proposed explanation for this gap starts by separating two things that standard economic theory treats as one. On one side is the benefit of having the product: what it's worth to whoever uses it, minus what you gave up to get it. That's consumer surplus, and for a purely rational agent the story ends there. On the other side is the pleasure—or irritation—of having made a good or bad deal, assessed not against its use value but against the price the person expected to pay. This expected price is what's called a reference price. When the actual price falls below it, there's additional satisfaction, a bargain; when it's above, there's the feeling of being robbed.

The point of this distinction is that the second part exists even when the product is identical. A classic study with MBA students who drank beer regularly illustrated this: researchers asked how much they'd pay for a bottle of their favorite brand brought by a friend to the beach. In one version, the beer came from a resort hotel; in the other, from a small, run-down grocery store. The beer is the same, the location of consumption is the same, the person doesn't even enter the establishment. Still, the median responses, adjusted for inflation, were US$7.25 for the resort beer and US$4.10 for the grocery store beer.

The proposed explanation is expectation: you expect a resort to charge more, partly because its costs obviously are higher. Paying seven dollars at the hotel is annoying but predictable; paying the same at the corner store is an outrage. What changes from one case to another isn't the product or the benefit—it's the comparison with a price the person carries in their head before reaching the counter.

An uncomfortable consequence follows from this, and the formulation itself spells it out: because this component can be positive or negative, it both prevents purchases that would be worthwhile and induces purchases that are waste. The case that sparked the idea is domestic and recognizable: a researcher's friend went to buy a comforter for a full-size bed. Regular prices were US$300 for king, US$250 for queen, and US$200 for full. That week, all sizes were US$150. She bought the king—too large for the bed, hanging off on all sides, but irresistible as a deal.

Why the extra unit weighs more than a penny off

If the mechanism is comparison against a reference price, then how the discount is announced ceases to be a detail. A one-third discount can be presented as a lower per-unit price or as one free unit. Financially, for someone buying three, it's the same thing. Psychologically, there are reasons to suspect it isn't.

The basis for this suspicion is how gains and losses are evaluated. The value function that behavioral economics inherited from prospect theory has three properties: people experience life in terms of changes, they feel diminishing sensitivity to both gains and losses, and losses hurt roughly twice as much as equivalent gains please. Diminishing sensitivity means the second unit of a gain is worth less than the first. The same gain split into three dilute portions yields less satisfaction than that same gain presented all at once.

The application to 'buy 3, pay 2' is a plausible reading, not a measured result: the free unit is a segregated gain, presented all at once, while a few cents off each item are three tiny gains each below the threshold of barely being felt. It's worth noting that available research supports the general mechanism of the value function solidly, but doesn't include experiments directly comparing bulk discounts against percentage discounts in this format. The explanation is consistent with what we know; it's not demonstrated for this specific case.

There's a second effect, more mundane and probably more powerful: the quantity offer scrambles the comparison. Price per kilo against price per package, three units against the competitor's value pack—the arithmetic exists, but it takes time and attention in the aisle. And a general warning from behavioral economics itself applies here: modeling the consumer as someone who does all the necessary calculations, with enough self-control to execute the optimal plan, is exactly the kind of assumption psychologists find implausible. The person facing the shelf isn't solving an optimization problem. They're deciding fast, using shortcuts.

The other side of the shelf: a menu, not a courtesy

Until now, the story has been about the buyer. But the format is no accident: it solves a concrete problem for the seller. A seller with any market power would like to charge a high price from those willing to pay it and a low price from those who only buy with a discount. The obstacle is that they don't know who is who—and can't ask.

The quantity discount gets around this obstacle without requiring anyone to reveal anything. It's a menu: buy one unit, pay full price; take three, pay less per unit. The buyer sorts themselves by choosing. It's the same logic that appears formalized in contract design literature, where the seller's central problem is inducing the high-valuation buyer to reveal their type—and where the binding constraint is exactly that one. Charge different prices to different people without announcing different prices: a quantity discount does this with a single label.

Literature on price discrimination also records the limits. The gains are small unless there are substantial differences in price elasticity between groups and a relevant share of the product sells at the higher price point. And arbitrage corrodes the scheme: when the product is easily resold or transferred between buyers, whoever accesses the low price can pass it on to someone who'd pay more. In a supermarket, neighbor-to-neighbor arbitrage is limited, but home storage serves a similar function over time—someone who bought three doesn't come back next week.

Why promotion instead of simply lowering the list price

The more strategic question remains: why a temporary promotion instead of simply lowering the list price? There's a direct argument for this in the classic analysis of rivalry between competitors, and it's not about the consumer. It holds that rivalry works through moves that provoke retaliation, and that price competition is the most unstable form of all: price cuts are quickly and easily matched by rivals, and once matched, they reduce revenue for all companies unless the sector's price elasticity of demand is high enough. Advertising battles, by contrast, might expand demand or strengthen differentiation, benefiting everyone.

A promotion in quantity format, with defined timeframe and product, sits in a middle ground: it changes the effective price without changing the list price, and it's easier to reverse than a permanent cut. This is a reading built from the argument about price-war instability, not a finding specific to this promotion type—the marketing literature that would directly measure the effect isn't among the available evidence here. What can be said with certainty is that the distinction between reversible and permanent moves is central to analyzing how rivals react, and the relevant question for whoever acts is always the same: how likely, how fast, how effective, and how severe will retaliation be.

There's also the time dimension. Under standard elasticity of demand frameworks, we'd expect a price change's effect not to be exhausted in the period it occurs: the immediate effect would be smaller than the total adjustment, and demand would be relatively more elastic in the long run than the short run. It's also argued that this adjustment is faster for food, bought weekly, than for durables, bought infrequently. If this reasoning holds for a supermarket promotion, part of its results wouldn't be new customer conversion but temporal displacement of purchases that would happen anyway—which is implication of the general mechanism, not measurement.

The side effect: destroying the ruler that makes the offer attractive

The mechanism has an embedded trap for whoever uses it. If pleasure comes from the gap between the price charged and the reference price, and if the reference price is learned from experience, constant promotion re-educates the buyer downward. The retail response to this problem has often been to manipulate the ruler: announce a largely fictitious 'suggested price' that works as a misleading reference. One proposed explanation for this pattern is that products sold this way share two characteristics: they're bought infrequently and their quality is hard to judge. Infrequent purchase helps because the consumer doesn't notice the promotion is permanent.

The opposite is also documented, and instructive. American retailers tried to wean customers off constant promotion with 'everyday low price' strategies, and these experiments generally failed. One reduced coupon use by 30% in spring 2007 compared to the previous year; sales plummeted and the chain quickly promised to return to coupon abundance. Another abolished fictitious suggested prices, ended traditional promotions, and rounded prices ending in ninety-nine cents; it claimed the effective final price was the same. Sales and the stock price plummeted when the changes took effect in 2012, the executive responsible was fired the following year, coupons returned, and by 2014 sales hadn't recovered.

The offered reading for the failure: it might even be true that consumers weren't paying more under the new regime, but they were losing a lot of transaction utility—including the small pleasure of paying US$9.99 instead of US$10. It's worth noting that big discount retailers operate with permanent low prices and didn't eliminate this component; they did the opposite, convincing the customer that the entire shopping experience is a bargain hunt.

Three questions that restore the real discount

None of this means looking for good deals is irrational. Saving on one purchase makes another possible. The risk is specific and bounded: buying something you won't use just because the deal is too good to pass up. Somewhere in the garage or attic of many people sits their personal version of that king comforter on a full-size bed.

• How many units would I buy if this promotion didn't exist? The real discount is on that number, not on three.

• What's the per-unit price, and what am I comparing it against—the announced list price or what this product usually costs?

• If I take three, do I skip buying for the next few weeks, or will I end up buying again before I finish what I took?

The useful inversion is there. The question the package answers is how much do I save. The question that matters is different: how much would I spend without it. And that's exactly the second calculation—the one that depends on knowing what you'd do before seeing the label—that the discount's format was designed to make difficult.