Consumer Demand and Market Competition with Time-Intensive Goods

How to cite

List, John; Goodman, Joseph; Frahan, Lancelot Henry de; Holz, Justin; McKay, Evan; McMenamin, Niall; Mogstad, Magne; Sadoff, Sally; Sider, Hal. 2026. "Consumer Demand and Market Competition with Time-Intensive Goods." AEA RCT Registry. AEARCTR-0016397. https://doi.org/10.1257/rct.16397-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

John List* , Joseph Goodman , Lancelot Henry de Frahan , Justin Holz , Evan McKay , Niall McMenamin , Magne Mogstad , Sally Sadoff , Hal Sider

* Primary investigator

Registry ID
AEARCTR-0016397
DOI
https://doi.org/10.1257/rct.16397-1.0
First registered
Last updated
Status
Completed
Trial dates
2023-05-01 – 2023-07-01
Countries
United States of America ()
JEL codes
C93, D12, L40, L41
Registry record
http://www.socialscienceregistry.org/trials/16397

Abstract

We leverage Becker’s time allocation theory to examine consumer demand and market competition for time-intensive goods. The Beckerian model predicts higher diversion ratios for goods with substantial time shares and those with high time costs relative to monetary prices. Applying this model to data from two field experiments, we analyze demand for Facebook and Instagram, focusing on substitution patterns across online activities and offline time use. Our findings indicate that users exhibit low elasticity to ad load, the primary user cost, and that time shares and time costs significantly influence diversion ratios. We explore the implications for user costs and benefits on these platforms and assess the potential impact of a Federal Trade Commission-proposed de-merger of Facebook and Instagram.

Intervention

We conducted two concurrent experiments. In one, participants received $4 per hour for reducing their average daily Instagram usage below a personalized baseline. In the other, participants received the same incentive for reducing Facebook usage below their respective baseline.

Experimental design

We recruited participants with ads on Facebook and Instagram, third-party publishers, and online panels maintained by firms for internet-based market research. Interested individuals were directed to a screening survey to confirm eligibility. Criteria included being 18 or older, residing in the U.S., primarily using Facebook or Instagram on an Android device (with sole access to that device), and having an average daily usage of at least 15 minutes on Facebook or 10 minutes on Instagram over the prior 28 days, based on Meta’s internal data. Eligible participants were instructed to download a RealityMine app from the Google Play Store, developed for passive device monitoring, and received $5 upon installation. They were guided to enable activity tracking, which recorded time spent on all apps on their Android device. Participants who successfully installed the app, enabled tracking, and could be matched to Meta’s data via email and Facebook or Instagram IDs were compensated an additional $10, provided their actual usage met the daily criteria.

We conducted two concurrent experiments: one for Facebook and one for Instagram and participants were first assigned to either the Facebook or Instagram experiment. Within each experiment, we then randomly allocated participants to either the treatment or control group. Participants in the treatment groups of both the Facebook and Instagram experiments received $4 per hour (prorated for partial hours) for reducing engagement below their compensation baseline, which we define as the average daily minutes of use of the week with the highest average daily use from the four weeks before the treatment period, rounded up to the nearest quarter hour. Participants in the control groups received compensation for continued data provision via their device monitor but faced no financial incentives to reduce engagement.

We collected data on daily minutes spent on phone applications using the RealityMine Android app, recorded participants using over 35,000 distinct apps during the experiment.

Design details

Randomization method
Blocked randomization by baseline usage of Facebook, Instagram and Snapchat. We followed a re-randomization approach to ensure balance across demographics, baseline usage of other apps and app categories, baseline Facebook friend count, and baseline Instagram follower and following counts.
Randomization unit
Individual (user) level.
Clusters
In total, the Facebook experiment included 3,500 participants, and the Instagram experiment included 2,768 participants.
Observations
In total, the Facebook experiment included 3,500 participants, and the Instagram experiment included 2,768 participants.
Arms
Facebook experiment: 1750 to treatment and 1750 to control. Instagram experiment: 1384 to treatment and 1384 to control.
Primary outcomes
Daily minutes spent on personal social networking (PSN), non-PSN and offline. Also, time spent on individual apps, like Facebook, Instagram, Snapchat, TikTok, YouTube, etc.
Sponsors
Name: Meta Location: Menlo Park, CA Url: https://www.meta.com
Partners
Name: Compass-Lexecon Partner_type: private_company Url: https://www.compasslexecon.com

Cited by

No citation data available for this registration yet.

Keywords