Judith Chevalier

Who they are

Chevalier is the William S. Beinecke Professor of Finance and Economics at the Yale School of Management and Faculty Director of its Program on Social Enterprise, Innovation, and Impact. Her faculty profile lists a PhD from the Massachusetts Institute of Technology in 1993 and a BA from Yale University in 1989.

Her listed research areas are wider than reviews. They include finance and industrial organisation, including work where the two overlap, the effects of new technologies on firms, individuals and policy, retail economics covering both e-commerce and physical stores, career choice, incentives and gig work, and studies of public health during COVID-19. Consumer product reviews appear in her profile as an interest within retail economics. This page covers only the part of her work that bears on reviews and sales, and says nothing about the rest beyond naming it.

Key papers

The first paper is Chevalier and Mayzlin (2006), published in the Journal of Marketing Research, volume 43, issue 3, pages 345 to 354. It compares consumer reviews and relative sales of books at Amazon.com and BarnesandNoble.com. Its abstract reports three findings. Reviews were overwhelmingly positive at both sites, although Amazon.com had more of them and longer ones. An improvement in a book's reviews raised its relative sales at that site. And one-star reviews had a larger impact than five-star reviews. The abstract concludes that new forms of customer communication on the internet have an important impact on customer behaviour.

The second is Mayzlin, Dover and Chevalier (2014), "Promotional Reviews: An Empirical Investigation of Online Review Manipulation", in the American Economic Review, volume 104, issue 8, pages 2421 to 2455. This site's note on that paper records its design, which compares the same hotels on Expedia, where reviews could come only from customers who had booked through the site, and on TripAdvisor, where anyone could post. The author order differs between the two papers. Chevalier is first on the 2006 paper and third on the 2014 one, and the two are sometimes confused for that reason.

Key ideas

Three ideas from the 2006 paper matter for this publication.

The first is that reviews are a measurable influence on sales and not only a record of opinion. The comparison of two retailers lets the authors look at how relative sales differ with relative reviews, and the abstract states that improvements in a book's reviews raised its relative sales at that site.

The second is the asymmetry. One-star reviews had a larger impact than five-star reviews. That result is the nearest thing in sales data to the popular saying that one bad review outweighs several good ones, and it points in the direction that psychologists describe as negativity bias. The abstract reports no exchange rate between the two ends of the scale, and the saying's figure of five good reviews to one bad one does not come from it.

The third is the setting. The reviews were overwhelmingly positive, which means that the average book at both retailers carried praise and that a low rating was the less common event. This site draws no conclusion from that beyond noting it, since the abstract does not say how the authors interpret it.

The article on this site that uses the 2006 paper, "Why Does One Bad Review Hurt So Much", is careful about its reach. The sample is books at two retailers, and the finding does not show how large the effect is for other products. Later consumer research, including Rohini Ahluwalia's 2002 paper and the 1991 paper by Paul Herr and colleagues, suggests that familiarity with the brand changes how much a single piece of word of mouth moves a judgement. Neither of those papers is by Chevalier, and neither was written about reviews posted online.

---