File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Estimating Effects of Long-Term Treatments

TitleEstimating Effects of Long-Term Treatments
Authors
KeywordsA/B tests
digital platforms
long-term treatments
surrogates
Issue Date12-Jul-2023
PublisherAssociation for Computing Machinery, Inc
Abstract

Randomized controlled trials (RCTs), also known as A/B tests, have become the gold standard for evaluating the effectiveness of product changes on digital platforms. Accurately estimating the effects of long-term treatments still remains a challenge. Product updates such as new user interfaces or recommendation algorithms are intended to persist in the system for an extended period. However, A/B testing is typically conducted for short durations, often less than two weeks, to facilitate rapid product iterations. Conducting lengthy experiments to capture the long-term impact of product changes becomes impractical due to potential negative impacts on user experiences, high opportunity costs associated with user traffic, and delays in decision-making processes.


Persistent Identifierhttp://hdl.handle.net/10722/338870
ISBN

 

DC FieldValueLanguage
dc.contributor.authorHuang, S-
dc.contributor.authorWang, C-
dc.contributor.authorYuan, Y-
dc.contributor.authorZhao, J-
dc.contributor.authorZhang, J-
dc.date.accessioned2024-03-11T10:32:09Z-
dc.date.available2024-03-11T10:32:09Z-
dc.date.issued2023-07-12-
dc.identifier.isbn9798400701047-
dc.identifier.urihttp://hdl.handle.net/10722/338870-
dc.description.abstract<p>Randomized controlled trials (RCTs), also known as A/B tests, have become the gold standard for evaluating the effectiveness of product changes on digital platforms. Accurately estimating the effects of long-term treatments still remains a challenge. Product updates such as new user interfaces or recommendation algorithms are intended to persist in the system for an extended period. However, A/B testing is typically conducted for short durations, often less than two weeks, to facilitate rapid product iterations. Conducting lengthy experiments to capture the long-term impact of product changes becomes impractical due to potential negative impacts on user experiences, high opportunity costs associated with user traffic, and delays in decision-making processes.</p>-
dc.languageeng-
dc.publisherAssociation for Computing Machinery, Inc-
dc.relation.ispartof24th ACM Conference on Economics and Computation (09/07/2023-12/07/2023, London)-
dc.subjectA/B tests-
dc.subjectdigital platforms-
dc.subjectlong-term treatments-
dc.subjectsurrogates-
dc.titleEstimating Effects of Long-Term Treatments-
dc.typeConference_Paper-
dc.identifier.doi10.1145/3580507.3597701-
dc.identifier.scopuseid_2-s2.0-85168162798-
dc.identifier.spage907-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats