Equal-Value Checkout Promotions Can Get Different Responses By Said Maybe, Said Maybe. Sources checked 2026-11-11. An analysis of three promotion experiments separates the size of an offer from the way shoppers see it. Compare equal-value offers in dollars, percentages and context. ## Quick answer In a randomized hypothetical hotel study, participants favored the uncertain offer more often when the alternatives were a 1% chance at a free $480 stay or $4.80 off than when they were a 10% chance at a free $48 stay or the same $4.80 off. A second study found that a chance at free was selected 4.2 percentage points more often than a chance at half price at equal expected value. These were stated choices. For an abandoned checkout test, compare paid orders and margin separately from promotion selection. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) · Gaertig and Simmons, Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.celiagaertig.com/uploads/1/1/5/6/115680147/gaertig_simmons_2026_uncertain_promotions.pdf) ## Key findings and sources ### Equal $4.80 savings came with different promotion preferences Gaertig and Simmons’s hypothetical hotel Study 1 compared a 1% offer on $480 with a 10% offer on $48. Both guarantees saved $4.80 and the matched chance-at-free offers had that same expected saving. More participants selected the uncertain offer in the $480, 1% condition than in the $48, 10% condition, the authors report. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### A 35.1-point choice difference for the 1% versus 10% offers In Gaertig and Simmons’s randomized hypothetical hotel Study 1 with 852 US MTurk participants, the 1% condition raised choice of the uncertain offer by 35.1 percentage points compared with the 10% condition. This is a stated choice between equal-value offers within each condition. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### Chance at free increased selection by 4.2 points In Gaertig and Simmons’s Study 2 with 2,338 US MTurk participants, a 10% chance at a free hotel stay led to 4.2 percentage points more selection of the uncertain offer than a 20% chance at a half-price stay. Both uncertain offers had an expected saving of 10%, equal to the sure discount. Responses were hypothetical. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### The same yearly membership saving shifted preference for a separate offer Gaertig and Simmons’s Study 3 held the personal-training offer fixed. The mean preference for a chance at free was 2.92 out of 6 beside a $48-per-year membership discount and 2.60 beside the equivalent $4-per-month discount. Said Maybe calculates a 0.32-point gap from the reported means. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### The lower-price effect appeared in both hotel studies Gaertig and Simmons report that the $48 hotel condition increased choice of the uncertain promotion by 7.2 percentage points in Study 1 and 5.6 points in Study 2 compared with the $480 condition. The two randomized scenario studies tested different offer designs and measured stated choices. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### Equal average promotion savings can have very different payout ranges On a $50 order, a 10% guaranteed discount and a 10% chance of a free order each have $5 in expected savings. The guaranteed saving is always $5. The chance offer pays $0 or $50 under these worked inputs. Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) ### The framing effect was roughly twice as large in the low-price condition In Gaertig and Simmons Study 4, percentage rather than dollar framing of an equivalent sure discount increased preference for the uncertain offer by 0.699 points at $11 and 0.347 points at $311. The outcome used a six-point preference scale. Gaertig and Simmons, Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.celiagaertig.com/uploads/1/1/5/6/115680147/gaertig_simmons_2026_uncertain_promotions.pdf) Randomized imagined purchase scenarios with 2,371 study participants. Compare the reported framing contrasts within the low- and high-price conditions. ### The framing interaction was 0.352 scale points The price-by-framing interaction in Gaertig and Simmons Study 4 was 0.352 scale points, with a normal-approximation interval of 0.056 to 0.648 based on the reported standard error. The outcome was stated preference. Gaertig and Simmons, Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.celiagaertig.com/uploads/1/1/5/6/115680147/gaertig_simmons_2026_uncertain_promotions.pdf) Study 4 on equal expected offers in imagined shopping scenarios. Use 0.352 plus or minus 1.96 times the reported 0.151 standard error. ## Compare the results Figure description: How offer design changed stated choice. Exact values are also in the study data. How offer design changed stated choice. ## Two $4.80 offers can look different When a shopper stops at checkout, an offer only helps if they understand what it changes. Gaertig and Simmons isolated one part of that decision by giving participants two hotel promotions with identical average savings. One guaranteed a discount. The other offered a chance at the full stay for free. In their first study, a 1% discount on $480 and a 10% discount on $48 both saved $4.80. A 1% chance at a free $480 stay and a 10% chance at a free $48 stay also had the same expected saving of $4.80. The researchers reported that participants favored the uncertain option more often in the $480, 1% condition than in the $48, 10% condition, even though the two guaranteed savings were equal. Across the first study, the difference in selecting an uncertain offer was 35.1 percentage points for the 1% versus 10% comparison and 7.2 points for the $48 versus $480 comparison. Those two coefficients describe the contrasts the experiment tested. They do not provide a formula for predicting what happens if a store doubles its coupon. The authors did not find a statistically significant interaction between promotion percentage and price in Study 1. ## Hold average savings fixed and change the payoff In Study 2, the guaranteed option was always 10% off. The uncertain alternatives were a 10% chance at the whole stay for free or a 20% chance at half price. Both uncertain designs average to a 10% reduction before anyone chooses. Participants selected the uncertain offer 4.2 percentage points more often when the possible outcome was free than when it was half price. The observed preference difference is informative precisely because the offers have the same expected value. The selection statistic does not tell a merchant what the realized expense would be when different numbers of shoppers accept an offer. The team separately changed how a neighboring promotion appeared. In Study 3, the personal-training sessions always cost $50, with a choice of $5 off or a 10% chance of getting them free. People saw the membership discount next to that choice as either $48 per year or $4 per month, two expressions of the same yearly saving. The mean preference for the uncertain add-on deal was 2.92 versus 2.60 on a six-point scale. Both means were below the midpoint, so the uncertain option gained ground while the average response still leaned toward the guaranteed saving. ## Calculate the price the customer can actually pay The most direct test for a DTC brand is often between two clear ways of describing a guaranteed offer. If a percentage applies only to certain items, calculate the dollar saving from the qualifying subtotal and confirm what happens to shipping, tax and any existing offer. A shopper with a $96 cart might only have $48 of eligible goods. The defensible description of a 10% offer in that case is a $4.80 reduction on the eligible items. Show the payable total only when the cart calculation can confirm it. If availability changes with a variant, destination or time, the recovery message has to use the current eligibility rather than the previous cart snapshot. Run the presentation test on the same eligible offer and compare completed paid orders, refunds, and contribution after discount. Record offer selection separately. A message can draw more selections without generating more retained revenue. If the experiment also changes the promotion itself, record that as a separate comparison. This study makes the distinction worth measuring. The amounts and probabilities in its hotel and gym scenarios do not set a target for checkout recovery. ## Equal average savings can create different customer outcomes A guaranteed 10% saving on a $50 order pays $5 every time. A 10% chance of receiving the order free has the same average payout across repeated draws. One person can receive $50 or zero under the second offer, so the checkout should make the odds and the confirmed final price visible. The merchant sees a different distribution too. With $32 of variable cost, the guaranteed offer leaves $13 of contribution on every qualifying order. The chance offer can leave $18 when no discount applies and a $32 loss when the item is free. Under the stated odds and stable order value, the average contribution is still $13. Actual offer performance depends on demand, order mix and how the promotion is implemented. Before considering a variable offer, verify that the customer understands the lowest and highest possible price and that the confirmed outcome stays with the cart. If the outcome cannot be verified during a recovery conversation, use the checkout’s recorded result before quoting a price. ## The same dollar saving changes how people compare the offers In Gaertig and Simmons’ Study 4, participants compared a 10% chance of getting a product free with an equivalent guaranteed discount. The researchers kept the value constant and varied whether the guaranteed saving was shown in dollars or as a percentage. The effect of percentage framing on preference for the uncertain offer was 0.699 points for the $11 product and 0.347 points for the $311 product on a six-point scale. The difference between the framing effects was 0.352 scale points. Applying a normal approximation to the reported interaction standard error gives an interval of 0.056 to 0.648 scale points. These numbers describe stated preference in an imagined purchase. The result suggests a testable presentation question, with the size of the measured change depending on price and the comparison a shopper was shown. ## Test whether the same offer is understood across product price bands Use an offer that the shopper can verify at checkout. For a guaranteed saving, show the exact value for the eligible variant and the resulting price. Ask a reviewer to find the same amount in the cart, the recovery message and the payment step. If the effect you care about is whether customers understood the saving, test that directly before looking for a paid-order change. A low-ticket product may have a clear absolute saving even when the percentage is visually prominent. A high-ticket product can make the shopper focus on the total cost and uncertainty instead. Compare presentation within the same product and offer terms, keep outcome and margin measurements separate, and check the refund or contact rate when introducing any new complexity. ## Study data Table: Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? Measure | Value | Context Study 1 sample | 852 participants | Final analyzed US MTurk sample. Study 1 lower hotel price | $48.00 | One-night price shown to the low-price group. Study 1 higher hotel price | $480.00 | Ten-night price shown to the high-price group. Study 1 smaller promotion percentage | 1% | Sure discount percent and chance of a free hotel stay in the smaller-promotion condition. Study 1 larger promotion percentage | 10% | Sure discount percent and chance of a free hotel stay in the larger-promotion condition. Study 1 lower percentage difference | 35.1 percentage points | Published OLS coefficient 0.351 scaled to percentage points for the choice of the uncertain promotion, 1% versus 10%. Study 1 lower price difference | 7.2 percentage points | Published OLS coefficient 0.072 scaled to percentage points for the choice of the uncertain promotion, $48 versus $480. Study 2 sample | 2,338 participants | Final analyzed US MTurk sample. Study 2 guaranteed discount | 10% | Sure discount percent in all assigned conditions. Study 2 chance of a free stay | 10% | Probability of receiving 100% off in the free-stay condition. Study 2 chance of a half-price stay | 20% | Probability of receiving 50% off in the half-price condition. Study 2 lower price difference | 5.6 percentage points | Published OLS coefficient 0.056 scaled to percentage points, $48 versus $480, averaged across uncertain offer type. Study 2 free-outcome difference | 4.2 percentage points | Published OLS coefficient 0.042 scaled to percentage points, chance at free versus chance at half-price, averaged across price. Study 3 sample | 2,221 participants | Final analyzed US Prolific sample. Study 3 yearly membership saving shown | 48 USD per year | Saving on the neighboring membership offer in the annual-price group. Study 3 monthly membership saving shown | 4 USD per month | Saving on the neighboring membership offer in the monthly-price group. Study 3 add-on price | $50.00 | Personal-training add-on price shown in both groups. Study 3 guaranteed add-on saving | $5.00 | Certain dollar saving on the personal-training add-on in both groups. Study 3 annual frame mean | 2.92 preference points out of 6 | Mean on the add-on promotion preference scale when membership savings were shown per year. Study 3 monthly frame mean | 2.6 preference points out of 6 | Mean on the same add-on promotion preference scale when membership savings were shown per month. Table: Gaertig and Simmons, Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? Measure | Value | Context Study 4 framing contrast at the lower price | 0.699 scale points | Published contrast or standard error Study 4 framing contrast at the higher price | 0.347 scale points | Published contrast or standard error Study 4 interaction standard error | 0.151 scale points | Published contrast or standard error ## What changes when you compare the figures Table: Calculations from the published values above. Comparison | Result | Calculation Guaranteed saving at 1% on $480 | $4.80 | = 480 * 1 / 100 Guaranteed saving at 10% on $48 | $4.80 | = 48 * 10 / 100 Expected saving at a 10% chance of free on $48 | $4.80 | = 48 * 100 / 100 * 10 / 100 Expected saving at a 20% chance of half price on $48 | $4.80 | = 48 * 50 / 100 * 20 / 100 Expected saving at a 10% chance of free on $480 | $48.00 | = 480 * 100 / 100 * 10 / 100 Expected saving at a 20% chance of half price on $480 | $48.00 | = 480 * 50 / 100 * 20 / 100 Yearly membership discount at $4 monthly | 48 USD per year | = 4 * 12 Preference gap in gym study from displayed means | 0.32 points on a six-point scale | = 2.92 - 2.60 Worked case, 10% guaranteed on a $50 order | $5.00 | = 50 * 0.10 Worked case, expected payout from a 10% chance of a free $50 order | $5.00 | = 0.10 * 50 Worked case, operator contribution on a free $50 order with $32 cost | $-32.00 | = 0 - 32 Low-price framing effect divided by high-price framing effect | 2.014 times | = 0.699 / 0.347 Lower approximate interaction interval bound | 0.056 preference scale points | = 0.352 - 1.96 * 0.151 Upper approximate interaction interval bound | 0.648 preference scale points | = 0.352 + 1.96 * 0.151 ## Methodology and sources Said Maybe transcribed Gaertig and Simmons’s Studies 1–3. We restate the four hotel choice coefficients in percentage points and recompute the expected saving as probability times discount share times price. We subtract rounded reported preference means for Study 3 and convert monthly membership savings to a yearly amount using 12 months. The experiments used hypothetical decisions, so no result is an observed cart recovery rate. [Download the source data (CSV)](decision-table.csv) [Download the calculations (CSV)](calculations.csv) [Download the citation notes (CSV)](citation-notes.csv) [Download the evidence behind the new findings (CSV)](finding-evidence.csv) ## Related reading Online discount statistics from a 314-trader pricing sweep (https://saidmaybe.com/guides/online-discount-statistics/) Shopping cart coupon statistics from Alibaba’s randomized promotion study (https://saidmaybe.com/guides/shopping-cart-coupon-statistics/) Countdown timer statistics and how to verify an ecommerce offer (https://saidmaybe.com/guides/ecommerce-countdown-timer-and-scarcity-statistics/) Check the real price before you promise two Shopify discounts (https://saidmaybe.com/resources/shopify-discount-combination-cases/) When to offer a discount in an abandoned checkout conversation (https://saidmaybe.com/guides/offer-a-discount-in-checkout-conversation/) Browse checkout recovery research (https://saidmaybe.com/research/) ## Sources and definitions Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.researchgate.net/publication/393132524_Why_and_When_Are_Uncertain_Price_Promotions_More_Effective_Than_Equivalent_Sure_Discounts) Gaertig and Simmons, Why (and When) Are Uncertain Price Promotions More Effective Than Equivalent Sure Discounts? (https://www.celiagaertig.com/uploads/1/1/5/6/115680147/gaertig_simmons_2026_uncertain_promotions.pdf) The source data includes the population, period, denominator and limitations for every figure. Said Maybe · saidmaybe.com