The Ebb and Flow of Online Word of Mouth

Abstract

Motivated by the importance of having reliable online rating systems, we examine the ebb and flow of online ratings with a focus on whether aberrations in online ratings cast a long shadow in subsequent ratings. Our study consists of two complementary pieces: first, we reveal the dynamics following rating aberrations using an econometric analysis of Yelp ratings for 8,943 restaurants in the state of Washington, over a period of 26 months; then, we follow up with two randomized online experiments of joke ratings to corroborate the field evidence and establish the causal nature of our findings. Overall, we find that aberrations in online ratings do not appear to cast a long shadow on future ratings: in the observational study, we find correction percentages of 102.8 % and 103.2% for positive and negative aberrations respectively; in the experiments, the corresponding numbers are 63% and 74% respectively. By and large, these findings suggest that online rating systems are fairly resilient to aberrations. Furthermore, we also find that the correction strength for positive aberrations is on par with those for negative ones, and larger aberrations are followed by stronger corrections. Our findings contribute to the nascent literature on the reliability of online ratings and hold important implications for consumers and online rating platforms.