Business Insights: Negative online reviews may be valuable, new tool for auditors

Online reviews expressing customer dissatisfaction with products like shoes and clothes may be a promising new tool for auditors, offering important signals to better assess a company’s sales returns, according to two Rutgers Business School researchers.

For the study, accounting and information systems Ph.D. student Muhammad Talha Afzal worked under the supervision of Professor Hussein Issa to gather 19 million customer reviews from Amazon and employed artificial intelligence to identify reviews indicating customer intentions to make returns. The data information was then integrated into standard auditing procedures.

The study is one of the first to demonstrate how Generative AI (GenAI) based textual analytics can be employed to identify and use return-related reviews as part of audit analytics.

“The results reveal that return-related reviews serve as a significant external indicator of sales returns,” said Afzal, whose interest was motivated by news in early 2026 about Amazon’s $309 million settlement to shoppers over returns. “This highlights the paper’s contribution because it demonstrates the ability of unstructured customer-generated textual data to provide audit-relevant insights,” 

Because returns can be an important item in financial statements, particularly for manufacturer’s where product defects are common, the ability to predict misstatements in sales return figures is critical for both auditors, company officials, and investors. 

The study focused on clothing, shoe and jewelry brands sold on Amazon because those products are generally more likely to be returned because of a customer’s dissatisfaction with size, color or quality. Those issues are exacerbated in online shopping when what a shopper sees can look so different than what they receive in the mail.

Afzal prompted ChatGPT to identify reviews referencing returns and refunds that reflected the intentions of dissatisfied consumers. The researchers analyzed the data of a variety of apparel-related industries from 2000 to 2018 to examine the relationship between return related reviews and actual sales returns.

As part of the research, Afzal and Issa developed a proxy for sales returns using the receivables turnover rate and then validated the proxy through a series of analyses to prove an association between return-related reviews and sales returns. 

They found that return-related reviews were more associated with sales returns for the prior quarter rather than the coming quarter. To corroborate this, Afzal performed a secondary text classification on a sample of 452,655 reviews using detailed prompt-engineered instructions for ChatGPT. The process categorized the reviews into “future return” and “past return” or “no return.” The analysis showed that consumers are more likely to initiate product returns before leaving reviews. Those insights help to identify the timeframe for incorporating customer reviews into auditing procedures. 

The findings of the study offer a new source of evidence that can be combined with records from a company’s management enabling auditors to more accurately assess the likelihood of misstated sales return figures in the planning stages of the audit, ultimately enhancing the quality of the audit.

“This study underscores the value of incorporating alternative data sources, or exogenous evidence, such as customer reviews into the audit process,” said Issa, who teaches how emerging technologies can be used in auditing.

Earlier studies – some done by Rutgers Business School professors and Ph.D. students –have demonstrated the growing relevance of non-financial data, including social media and reviews for financial reporting and auditing purposes. Less research has focused specifically on how product reviews relate to sales returns.

The research by Issa and Afzal is an example of the innovative, industry-changing work produced by Rutgers Business School’s Continuous Auditing and Reporting Lab. As the flagship business school of New Jersey, Rutgers Business School fosters innovation among its faculty researchers who influence industry with their thought leadership and help to provide a strong return on investment for students.

The research paper is currently under review at the International Journal of Accounting Information Systems.

-Susan Todd