Platform Design and Market Power

Last registered on August 08, 2025

Pre-Trial

Trial Information

General Information

Title
Platform Design and Market Power
RCT ID
AEARCTR-0016524
Initial registration date
August 06, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
August 08, 2025, 7:17 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Region

Primary Investigator

Affiliation
University of Innsbruck

Other Primary Investigator(s)

PI Affiliation
University of Innsbruck
PI Affiliation
University of Duesseldorf
PI Affiliation
University of Innsbruck
PI Affiliation
University of Munich

Additional Trial Information

Status
In development
Start date
2025-08-07
End date
2026-08-06
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
We examine how the design of an online retail platform - the complexity of price comparisons or the existence of a buy box - affects the market power of sellers competing on this platform. To this end, we conduct an online experiment in which we vary platform design and seller prices for a homogeneous product. To capture consumer behavior, we estimate a structural model that separates consumer attention and choice stages. We then derive equilibrium prices by simulating seller behavior in a pricing game where consumers' behavior is given by the structural parameters.
External Link(s)

Registration Citation

Citation
Duell, Adrian et al. 2025. "Platform Design and Market Power." AEA RCT Registry. August 08. https://doi.org/10.1257/rct.16524-1.0
Experimental Details

Interventions

Intervention(s)
The following description refers to the baseline treatment (which will be labeled BB-complex). We invite subjects on Prolific to an online experiment which consists of two parts. The first part is a survey in which we elicit demographic information (age, gender, education), online labor supply, online shopping experience, as well as measures of cognitive ability, trust, risk and time preferences. After the survey, we describe the second part, which consists of an experimental shopping task.

In the shopping task, subjects have to purchase a hypothetical book (HB). They will not actually purchase a book, but their earnings in the shopping task increase in the price savings that they realize. There are several sellers that offer the book. If a subject purchases HB from at the total price p, then her payoff in the shopping task is 15 – p USD.

In the baseline treatment, the total price consists of four partial prices that we call base price, shipping fee, delivery time costs, and seller rating costs. The total price is the sum of these partial prices. Each seller offer defines values for these partial prices.

Upon starting the shopping task, participants first see the overview webpage. This webpage does not only display HB, but also other books (by design, it is not possible to purchase any other book). In the baseline treatment, there is a “buy box”: Participants see the price vector of the top-ranked seller and can directly purchase the book from it by clicking on a button. Alternatively, they can click on a button called “All buying options”, in which case they enter the price comparison webpage. This page shows the offers of all 12 sellers. Both the overview and the price comparison webpage are designed like a generic online retail platform so that subjects have an intuitive understanding how they work. On the price comparison webpage, sellers are ordered according to the sum of the last three entries in the price vector (shipping fee, delivery time costs, seller rating costs), from lowest to highest value.

Subjects can purchase from the top-ranked seller on the overview webpage through the buy box or they can purchase from any seller on the price comparison webpage by clicking on a “Buy” button next to the corresponding seller’s price vector. Upon clicking on such a button, subjects need to confirm their choice. After confirming a choice, the shopping task is completed. If they do not trade with any seller, their payoff in the shopping task is zero.
Intervention Start Date
2025-08-07
Intervention End Date
2026-08-06

Primary Outcomes

Primary Outcomes (end points)
(1) Choices and overpayment on the platform

(2) Time spent on the different webpages

(3) Survey measures of demographics, cognitive ability, risk and time preferences, trust, online labor supply, and online shopping experience
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We implement six main treatments that we call BB-complex, BB-medium, BB-simple, NBB-complex, NBB-medium, and NBB-simple. BB-complex is the baseline treatment described above. In all main treatments, the sellers’ ranking and their total prices (and hence subjects’ choice sets) are identical, only the presentation of prices and the design of the platform changes between treatments.

BB-medium is identical to BB-complex except that the total price in this treatment consists of only two partial prices (instead of four), base price and shipping fee. For each seller, the base price in BB-medium is identical to the base price in BB-complex, and the shipping fee in BB-medium equals the sum of shipping fee, delivery time, and rating costs from BB-complex. On both the overview and the price comparison webpage, only the base price and the shipping fee are displayed.

BB-simple is identical to BB-complex except that there is only one price component, the base price. For each seller, the base price in BB-simple equals the sum of the four partial prices in BB-complex. On both the overview and the price comparison webpage, only the base price is presented. Finally, NBB-complex (NBB-medium, NBB-simple) are identical to BB-complex (BB-medium, BB-simple) except that there is no buy box: On the overview page, participants do not see the price vector of the top-ranked seller and also cannot purchase the book right away. Instead, they have to access the price comparison webpage.

Parallel to the six main treatments, we run additional treatments in which we vary the sellers’ total prices (otherwise these treatments are identical to the main treatments).
Experimental Design Details
Not available
Randomization Method
Randomization by a computer
Randomization Unit
We randomize at the individual level.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
At least 1800 individual observations
Sample size: planned number of observations
Our goal is to recruit at least 1800 subjects in total. The task will be kept on Prolific until 1800 subjects completed the study.
Sample size (or number of clusters) by treatment arms
At least 1800 individual observations
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Board for Ethical Questions in Science of the University of Innsbruck
IRB Approval Date
2025-04-09
IRB Approval Number
Certificate 35/2025