Writing More Compelling Creative Appeals: A Deep Learning-Based Approach

Jiyeon Hong () and Paul R. Hoban ()
Additional contact information
Jiyeon Hong: University of Wisconsin-Madison, Madison, Wisconsin 53706
Paul R. Hoban: Amazon.com, Inc., Seattle, Washington 98109

Marketing Science, 2022, vol. 41, issue 5, 941-965

Abstract: We present a deep learning algorithm to provide personalized feedback on creative appeals, written content intended to persuade readers to undertake some action. Such appeals are widespread in marketing, including advertising copy, RFP responses, call center scripts, product descriptions, and many others. Although marketing research has produced several tools to help managers glean insights from online word-of-mouth, less attention has been paid to creating tools to assist the innumerable marketers responsible for crafting effective marketing messages. Our approach leverages the hierarchical structure of written works, associating words with sentences and sentences with documents, and the linguistic relationships developed therein. We score each sentence in an appeal by its expected contribution to success accounting for its substance and persuasive impact. The sentences with the lowest scores make the appeal less compelling and are the most effective points to focus a revision. The approach has proved effective in a randomized control trial, with subjects rating essays revised with the aid of algorithmic feedback as being 4.5% more likely to achieve their objectives. In addition to providing automated feedback to authors, we leverage the model’s output to derive substantive insights into what makes an appeal compelling.

Keywords: deep learning; natural language processing; recurrent neural networks; creative appeals (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (4)

Downloads: (external link)
http://dx.doi.org/10.1287/mksc.2022.1351 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:inm:ormksc:v:41:y:2022:i:5:p:941-965

Access Statistics for this article

More articles in Marketing Science from INFORMS Contact information at EDIRC.
Bibliographic data for series maintained by Chris Asher ().