012 min
How free changes a choice
Here is an illustrative case, using roles instead of people. A product manager at a subscription note-taking app sells two plans. The Basic plan costs $1 a month and the Pro plan costs $9 a month. Pro is clearly better, and most people who pay choose it.
The product manager considers a change: make Basic free and cut Pro to $8. The gap between the plans stays at $8. Both plans become $1 cheaper in the same way.
Standard price theory says this change should barely move anyone from Pro to Basic. A buyer who weighs benefit against price sees the same difference between the two plans as before. The only people who should react are those who were not buying at all, and some of them may now buy.
The zero-price effect predicts something else. When one option becomes free, some people who preferred the better option switch to the free one anyway. They give up the plan they valued more.
The studies in this entry measured that pattern with chocolate, not with software. The app prices above are an invented example of what the researchers looked for, not a result.
This is why the effect has its own name. Most price changes move demand in proportion to their size. A change to zero moves it by much more than its size predicts. The rest of this entry covers how researchers separated that extra movement from an ordinary rise in demand, what they found, why it may happen, and where the finding stops applying.
022 min
Where the free effect comes from
Kristina Shampanier, Nina Mazar and Dan Ariely published "Zero as a Special Price: The True Value of Free Products" in the journal Marketing Science in 2007 (volume 26, issue 6, pages 742 to 757). When they wrote it, they worked at MIT, the University of Toronto and Duke University. An earlier version appeared as a Federal Reserve Bank of Boston working paper titled "How Small is Zero Price?"
The authors started from an observation that they described as intuition and anecdotal evidence: many people wait in long lines for a free ice-cream cone or a free coffee that they could buy on another day for two to three dollars.
That observation alone proves little. Demand for any product is high when its price is very low, so long lines at a price of zero do not show that zero is special. The authors needed a design that separated an ordinary rise in demand from a special reaction to zero.
They also placed their work among older findings that zero behaves differently from small positive numbers in other areas of psychology, such as rewards and probabilities. Their contribution was to test the idea for prices, using real purchases where they could.
A note on the term. The paper's title and abstract use "zero price" and "free" for the same thing. Later writers call the result the zero-price effect or the free effect. This entry uses all three names for one finding.
032 min
How the free effect was tested
The researchers built every experiment on one structure. Each person chose between three options: buy a low-value chocolate, buy a high-value chocolate, or buy nothing.
There were two conditions. In the cost condition, both chocolates had positive prices, for example 1¢ for a Hershey's Kiss and 15¢ for a Lindt truffle. In the free condition, both prices were reduced by the same amount, so that the cheaper good became free. The Kiss was then 0¢ and the truffle 14¢.
Standard theory makes a clear prediction for this change. Buyers pick the option with the largest benefit minus price. When both prices fall by the same amount, that difference stays the same for each chocolate. According to the authors' model, there should be no switching from one product to another. The only expected change is that some people who bought nothing now buy one.
The design works because it rules out the usual explanation. Demand for a product goes up when its price falls, and that is not surprising. But demand for the truffle should not go down when both prices fall together. If it does, free must be adding something beyond the price cut.
The authors also described a second model in which a price of zero adds extra value to the free product. The two models give different predictions for the same data, so the experiments can tell them apart.
| Experiment | Setting | Products | Cost condition | Free condition |
|---|---|---|---|---|
| 1 | Hypothetical survey, 60 participants | Hershey's Kiss and Ferrero Rocher | 1¢ and 26¢ | 0¢ and 25¢ |
| 2 | Real sales at a booth in the MIT student center, 398 people | Hershey's Kiss and Lindt truffle | 1¢ and 15¢ | 0¢ and 14¢, and 0¢ and 10¢ |
| 3 | Real sales at an MIT cafeteria, 232 customers | Hershey's Kiss and Lindt truffle | 1¢ and 14¢ | 0¢ and 13¢ |
Experiment 1 asked 60 participants to make a hypothetical choice among a Hershey's Kiss, a Ferrero Rocher chocolate and buying nothing. Experiments 2 and 3 used real chocolates and real money.
043 min
What happened when the cheaper chocolate became free
Experiment 1: a survey
In the survey, a 1¢ price change between positive prices did nothing. Moving prices from 1¢ and 26¢ to 2¢ and 27¢ left demand almost unchanged. Moving them from 1¢ and 26¢ to 0¢ and 25¢ raised demand for the Kiss and lowered demand for the Ferrero Rocher. Only when one of the prices became zero did the shift take place in Experiment 1.
Experiment 2: a booth with real sales
Experiment 2 used real chocolates. A booth in the MIT student center held two boxes of chocolates and a sign that read one chocolate per person. Price signs lay flat on the table, so only people standing close could read them. The researchers alternated the price signs about every 45 minutes. A person who read the prices and took nothing was counted as choosing nothing.
Among people who took a chocolate, the Kiss's share rose from 27% in the 1¢ and 15¢ condition to 69% in the 0¢ and 14¢ condition and to 64% in the 0¢ and 10¢ condition.
The third condition shows how strong the effect was. In it, the truffle's price fell by 5¢ instead of 1¢, a much larger cut than the one that had already shifted demand. Demand for the truffle did not rise when its price fell by 5¢, and the small difference went in the opposite direction. The authors found that a price of zero was more powerful than a price cut five times larger that stayed above zero.
Experiment 3: removing the hassle of paying
Experiment 2 had a weakness. Taking a free chocolate, or taking nothing, needs no payment. Buying needs change, a wallet and a transaction, so effort alone might explain the result.
Experiment 3 removed that difference. At an MIT cafeteria, the price of a chocolate was added to the customer's bill for their meal, so every option needed the same payment steps. In total, 232 customers took part in Experiment 3. Among those who took a chocolate, the Kiss's share rose from 21% in the 1¢ and 14¢ condition to 71% in the 0¢ and 13¢ condition.
Experiment 4: no money at all
Experiment 4 tested whether the effect needs money. Thirty-four trick-or-treaters were each given three Hershey's pieces and offered a small or a large Snickers bar. In the free condition, the child could take the small bar for nothing or trade one Hershey's piece for the large bar. In the cost condition, the small bar cost one piece and the large bar cost two. The authors reported that the zero-price effect stayed strong even when the trade involved similar products and a similar currency.
How to read these numbers
All the shares above are among people who took a chocolate. They exclude people who took nothing. The experiments used cheap candy and mostly student participants.
052 min
Why people react to free as they do
The authors tested three explanations and judged one most likely.
Social norms
Social norms are unwritten rules about how to behave. One idea is that a priced item makes people think like buyers, while a free item makes them think like guests receiving a gift, and gifts feel more valuable. The authors set this aside as the main cause, because earlier work shows that merely mentioning a price brings in market rules, and every condition here showed a price.
Mapping difficulty
Mapping difficulty means that benefits and costs are hard to compare when they are in different units. A chocolate is not money, so people may weigh the two sides badly. A free item has no cost to compare, so it becomes easy to judge. This was the reason for Experiment 4, which traded chocolate for chocolate. The effect stayed, so mapping difficulty cannot be the whole story.
Affect
Affect is a quick feeling of liking or disliking. Research on the affect heuristic shows that people often use that first feeling as a cue when they decide. The authors argued that free options produce a stronger positive feeling than their value explains.
They tested this in two steps. In Experiment 5, 243 participants rated how attractive one chocolate offer was by circling a smiley face. The offers were a free Kiss, a Kiss for 1¢, a truffle for 13¢ and a truffle for 14¢. The free Kiss was rated higher than all three others, and the other three did not differ from each other. The authors wrote that the free good elicits more positive affect than standard cost-benefit analysis predicts.
In Experiment 6, 200 students made a hypothetical choice. Half first answered two questions that forced them to compare quality and price. The zero-price effect appeared in the group that chose directly. The zero-price effect did not appear in the groups that compared quality and price before choosing. When people stop to compare, the feeling matters less.
The authors concluded that affect emerges as the most likely account for the effect. They also warned that several forces may act together, and they listed other possible effects of zero, such as inferences about quality and lower barriers to trying a product.
062 min
What later studies and real cases add
Amazon and France
The paper reports one real-world case. When Amazon introduced free shipping in some European countries, the price in France was mistakenly reduced not to zero but to one French franc, a negligible positive price of about 10¢. The number of orders increased dramatically in the countries with free shipping, and not much changed in France.
The authors called this anecdotal evidence. The paper gives no order counts, so treat it as a reported case that fits the theory, not as a measurement.
Follow-up experiments
Michal Gal and Daniel Rubinfeld reviewed the follow-up studies in a 2014 paper on free goods and competition law. They describe two that matter for business decisions.
- Buy one, get one free. Spiegel, Benzion and Shavit (2011) compared marketing forms with the same final price. The experiment showed that consumers usually preferred getting one product for free over getting a 50% discount on each of two products.
- Free breakfast. Nicolau and Sellers (2012) studied hotel choice. When the low value hotel offered a free breakfast, the demand for the low value hotel increased, beyond the market value of the breakfast.
What happens to value after the offer ends
A different question is what a free product is worth to the buyer afterwards. Palmeira and Srivastava (2013) studied free gifts that come with a required purchase. Their research demonstrates that willingness to pay for the product after the promotion is retracted is higher when it was offered for free than when it was offered at a low, discounted price.
This conflicts with earlier work. Palmeira and Srivastava wrote that a free offer does not devalue the product at all and, at a minimum, devalues the product less than if it were offered for a low, discounted price. Gal and Rubinfeld report the opposite pattern from other work. A study they cite found that firms that offered their products for free and tried to charge a low price lost a significant proportion of their customers. The two results may differ because the products and offers differ, and the question is not settled.
The word you choose
Koo and Suk (2020) tested how a free offer is written. Results of ten experiments, including a field study, demonstrate that a free promotion is evaluated more favorably when presented as "$0" than "free." They explain this with framing effect logic: "$0" makes people focus on the cost they avoid, and avoiding a cost weighs more than gaining an item.
072 min
How to use the free effect in pricing decisions
The studies tested chocolate, gift certificates and television choices, not software or services. The steps below are practical suggestions based on the findings. They are not results from the studies. Each step depends on the one before it.
Write down the choice and the gap.
The effect appeared when one option became free while the price gap between options stayed the same. List the options a buyer sees, their prices, and the gap. If your change also changes the gap, you cannot attribute a shift to free.
Test a free arm against a low-price arm.
This step only works after step 1, because the two arms must keep the same gap. Run an A/B test with one arm at $0 and one at a small positive price. A product manager or founder should compare which option people choose and revenue per visitor, not only sign-ups.
Cost the free option.
Every free user has a serving cost. An engineer or finance lead estimates it. If the free option wins on choice but loses on cost, the test result alone does not justify the change.
Measure what happens after.
The research is mixed on how people value a product once the free period ends. Track conversion to a paid plan, and track cancellations, for the people who started free.
Choose the wording.
A designer writes the label. Given the Koo and Suk result, test "$0" against "Free" on the same offer.
The authors of the 2007 paper made one suggestion of their own: when considering promotions at a low price, companies should experiment with further discounts to zero.
082 min
Where the free effect has limits
Small stakes. The authors wrote that one limit of their experiments is that they are restricted to relatively cheap products and relatively unimportant decisions. They ran one hypothetical survey on flat-panel televisions to test larger sums. In it, demand for the smaller, cheaper television was 40% at $299, 40% at $199, 43% at $99 and 83% when it was free. The authors suspected that for a car, a price of $100 might be perceived as close enough to zero to work the same way, but they did not test it.
Few kinds of buyers. Gal and Rubinfeld note that the studies they reviewed were performed on relatively inexpensive goods and many were performed on students.
Products that go together. According to Gal and Rubinfeld, Spiegel and colleagues found that the free effect disappeared when two products were perfect complements. Buyers treated those two products as one package.
Careful comparison. In Experiment 6, the effect weakened when people first compared quality and price. A buyer who is shown a comparison table, or who is spending a company's money after review, may react less.
Choice is not profit. The experiments measured which chocolate people chose, not whether the seller earned more. A free option can win the choice and still lose money. Step 3 above is the check.
Contested after the offer. As the previous section shows, studies disagree on how a free offer affects the later value of a product.
092 min
The free effect vs nearby concepts
Three concepts are often confused with the free effect because they also change what a price seems to mean. The deciding question is which part of the situation changed.
| Concept | What changed | Question that separates it |
|---|---|---|
| Not in the library yetFree (zero-price effect) | ||
| MarketingPrice anchoring | ||
| MarketingDecoy effect | ||
| PsychologyLoss aversion |
Loss aversion is related to one explanation of free. Koo and Suk argue that "$0" works better than "free" because it focuses people on the cost they avoid. That is a proposal about wording, not a replacement for the affect account above.
"Free" has other meanings that this entry does not cover. Free software means software with freedom to modify and share it, and it may still cost money. Freemium is a business model that offers a free tier alongside paid plans. A free tier is one place the zero-price effect can appear, but the model itself includes costs and conversion that the effect does not address.
?8 questions
Questions people ask
What is the zero-price effect?
Who discovered the free effect?
What is an example of the zero-price effect?
Is a price of 1 cent the same as free?
Why do people choose a free item over a better one that costs little more?
Does free shipping work because of the zero-price effect?
Does the zero-price effect work for expensive products?
How can a team test the free effect on its own pricing?
§5 sources
Sources
Shampanier, K., Mazar, N. and Ariely, D. (2007). Zero as a Special Price: The True Value of Free Products. Marketing Science, 26(6), 742-757. (open copy: )
Shampan'er, K. and Ariely, D. (2006). How Small is Zero Price? The True Value of Free Products. Federal Reserve Bank of Boston Working Paper 06-16.
Palmeira, M. and Srivastava, J. (2013). Free offer is not equal to cheap product: A selective accessibility account on the valuation of free offers. Journal of Consumer Research, 40(4), 644-656. (record: )
Koo, J. and Suk, K. (2020). Is $0 Better than Free? Consumer Response to "$0" versus "Free" Framing of a Free Promotion. Journal of Retailing, 96(3), 383-396.
Show all 5 sourcesShow fewer sources
Gal, M. and Rubinfeld, D. (2014). The Hidden Costs of Free Goods: Implications for Antitrust Enforcement.
