Back to History Current Version

Integrating Predictive Models into Two-Sided Recommendations: A Matching-Theoretic Approach

Last registered on January 23, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
Integrating Predictive Models into Two-Sided Recommendations: A Matching-Theoretic Approach
RCT ID
AEARCTR-0015446
Initial registration date
January 12, 2026

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
January 23, 2026, 7:31 AM EST

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

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Suguru Otani
Affiliation
Graduate School of Economics, The University of Tokyo, Market Design Center
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Shunya Noda
PI Affiliation
Graduate School of Economics, The University of Tokyo
Contact Investigator
PI Name
Kazuki Sekiya
PI Affiliation
Graduate School of Economics, The University of Tokyo
Contact Investigator
PI Name
Yuki Komatsu
PI Affiliation
Department of Economics, The University of Tokyo
Contact Investigator
PI Name
Sachio Ohkawa
PI Affiliation
MiDATA Co., Ltd.
Contact Investigator

Additional Trial Information

Status
In development
Start date
2026-01-13
End date
2026-03-31
Keywords
Labor, Welfare
Additional Keywords
two-sided platform, matching theory, recommender system, machine learning, deferred acceptance, linear programming, algorithmic market design, efficiency and equity
JEL code(s)
C78, D47, C48
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Two-sided platforms require two-sided recommendations to ensure both sides benefit from mutual interest. While conventional one-sided recommenders focus on individual preferences, they ignore how over-recommending attractive users can exacerbate inequality. We propose a data-driven framework for efficient and equitable two-sided recommendations by integrating machine learning-based predictive models (e.g., like probabilities and login rates) with insights from matching theory. Through a simulation study using real-world data from CoupLink, a Japanese online dating service, we demonstrate three main findings: (i) when integrating predictive models using deferred acceptance, employing the probability of successful matches as users' preferences leads to better outcomes, (ii) exposure-constrained deferred acceptance (ECDA), which incorporates an improved capacity design, further enhances equity, and (iii) adopting linear programming-based approaches improves both efficiency and equity beyond what ECDA achieves.
External Link(s)

Registration Citation

Citation
Komatsu, Yuki et al. 2026. "Integrating Predictive Models into Two-Sided Recommendations: A Matching-Theoretic Approach." AEA RCT Registry. January 23. https://doi.org/10.1257/rct.15446-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The intervention involves implementing a new recommendation algorithm, the Exposure-Constrained Deferred Acceptance (ECDA) algorithm, into a Japanese online dating platform. This algorithm is designed to optimize matching efficiency while balancing equity in match allocations. The intervention is introduced in treatment markets, where the platform’s recommendation system is modified to use ECDA, while the control markets continue using the existing recommendation algorithm.
Intervention Start Date
2026-01-13
Intervention End Date
2026-01-26

Primary Outcomes

Primary Outcomes (end points)
Total Number of Dates
Average Likes
Unique Dating Proposers
Unique Dating Receivers
Gini Index for Proposers
Gini Index for Receivers
Likes Sent to Receivers
Post Engagement Actions
Total Number of Effective Users
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
This study is a two-arm cluster-randomized controlled trial with geographic areas as clusters.
Prior to rollout, one of two areas (Kanto vs. Kansai–Chubu) was randomly assigned to receive the ECDA recommendation algorithm, while the other continued using the existing algorithm.
The primary estimand is the intention-to-treat (ITT) effect of ECDA on pre-specified marketplace outcomes during the intervention window.
The experiment runs for two weeks and we track post-treatment activity to obtain post-engagement data. The primary outcome measures include:
- Total Number of Dates (efficiency metric)
- Unique Dating Users (equity metric)
- Proposer and Receiver Gini Index (inequality metric)
- User Engagement Activity (retention measure)
- Total Number of Effective Users (post engagement effectiveness)

The estimation identifies the impact of the ECDA algorithm by comparing pre- and post-intervention differences between treatment and control groups, with market-specific fixed effects controlling for baseline differences.
Experimental Design Details
Not available
Randomization Method
Simple random assignment (1:1) at the geographic area (cluster) level. Unit of randomization is the geographic market/area. Prior to the intervention start date, we randomly assigned treatment status across the two areas (Kanto vs. Kansai–Chubu) using a computer-generated random number. The randomization was implemented by a member of the research team using a computer-generated random number prior to the intervention. The unit of randomization is the geographic market (area). One area (1 cluster) was assigned to treatment (ECDA deployed) and the other area (1 cluster) to control (status quo algorithm). No within-area randomization is used due to interference/spillovers among users within the same market.
Randomization Unit
Areas (e.g., Kanto and Kansai-Chubu)
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
2 geographical areas
Sample size: planned number of observations
4,438 males directly affected as proposers and 11,685 females indirectly affected as receivers for treatment 2,342 males directly affected as proposers and 9,122 females indirectly affected as receivers for control The numbers are based on the active user data from 26 December 2025 to 8 January 2026.
Sample size (or number of clusters) by treatment arms
1 area for treatment
1 area for control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
The Office for Life Science Research Ethics and Safety at the University of Tokyo
IRB Approval Date
2025-11-07
IRB Approval Number
E25ALS0255
Analysis Plan