Display advertising can change shoppers’ expectations about a product’s price and quality even when the ad does not state either attribute, according to a 2026 working paper from researchers studying a simulated online store.
The researchers found that those expectations affected whether participants searched for and purchased an advertised product. Visibility explained most of the advertising effect in their model, but changes in beliefs accounted for a meaningful share as well.
Jean-Pierre Dubé, Ilya Morozov, Franklin She and Anna Tuchman describe the work in a National Bureau of Economic Research working paper issued in August. A working paper is preliminary research and may be revised.
What the experiment tested
Participants shopped in a simulated online store where the researchers varied exposure to display ads, their content and the advertised product’s price. They also asked participants about their expectations and observed choices under the experiment’s incentives.
The design let the researchers study two ways an ad could affect demand. A prominent placement can make a product easier to notice. The ad may also cause someone to infer something about the product before viewing its details.
For example, a shopper might infer that a product is more expensive or higher quality from how it is presented, even if the advertisement does not explicitly say so. The precise inference depends on the ad and the shopping context; it should not be assumed that every ad creates the same impression.
Advertising is commonly measured by clicks, impressions and sales. Our guide to online advertising explains the formats businesses use to place paid messages in digital spaces. The experiment adds another question: what did the customer believe before clicking?
Why clicks alone may not explain the result
A prominent ad can increase product views simply because it is easier to see. If the same placement also changes expectations, the group of people who click may approach the product with different assumptions about its value.
For a retailer testing a new placement, that means a higher click rate should be read alongside product-page behavior, completed orders and returns where those measures are available. A click does not by itself establish whether the customer understood the price or whether the ad generated a profitable sale.
Businesses can compare matched versions of a campaign, varying one feature at a time, and ask customers what they expect before showing the full product information. Any claims used in the test still need to be accurate and consistent with applicable advertising rules.
The researchers also use a theoretical model to examine the information an advertiser might choose to disclose. Their model suggests there can be incentives to withhold some product details. That is a prediction within a model, not evidence that withholding information improves every campaign or benefits customers.
The experiment took place in a simulated store rather than across all retail platforms. Its estimates and mechanisms may differ for other products, audiences and placements. The paper’s contribution is to show that an ad can affect both what people notice and what they expect to find.
