Asking customers to reconsider a subscription choice increased selection of an annual plan by more than eight percentage points in a field experiment, but a similar prompt reduced original purchases in a separate retail test.
The Journal of Consumer Research study examined a feature the authors call a confirmation nudge: a prompt asking someone to confirm an initial choice or switch to an alternative.
Kellen Mrkva, Shannon Duncan, Marissa Sharif and Stanley Zuo combined two business experiments with preregistered laboratory tests. Their findings show that an extra decision can influence what people buy, while also creating another opportunity to leave.
A second question changes the decision
In the subscription experiment, app users were randomly assigned to a standard process or a prompt offering a choice between confirming their initial selection and switching to an annual plan. The increase of more than eight percentage points describes that test, not an expected gain for every subscription company.
An annual plan can offer a lower effective monthly price, but it also requires a longer commitment or a larger payment upfront. Whether it saves a particular customer money depends on the stated price, renewal terms and how long the customer actually uses the service.
The researchers found stronger effects when the alternative was desirable and participants were more likely to benefit from it. In other experiments, participants viewed the prompt as more manipulative than comparison conditions.
The paper appeared online in 2025 and in the journal’s October 2026 issue. Wharton’s account of the research describes the subscription test and a separate jewelry-retailer experiment.
More add-ons, fewer completed purchases
At the jewelry retailer, the share choosing an added service plan increased from about 19% to nearly 32% after the confirmation prompt. More shoppers also abandoned the original purchase, and revenue per visitor was $0.64 lower in that test, according to Wharton.
Higher attachment of an add-on is not the same as higher total revenue or profit. A seller needs to count customers who leave, the cost of providing the extra service, refunds and longer-term effects on trust.
A company’s own trial could randomly assign comparable visitors to a normal checkout and a clearly worded prompt. The business can then compare completed orders, revenue per visitor and complaints, while keeping other conditions stable. A small sample or seasonal promotion can make the results harder to interpret.
Choice still has to be clear
For customers, an offer should make the payment amount, billing period and renewal terms easy to understand before they commit. The ability to choose the original option should remain clear and available.
Our earlier reporting on delivery and payment friction at checkout examined a separate survey. Together, the studies illustrate why improving one stage of a purchase does not guarantee that more people finish the entire process.
