The Problem(s) with Online Reviews
Online reviews through Google, Yelp, Amazon are ubiquitous and a common step in most people's process for buying anything more consequential than toothpaste or bottled water. The presence of a 4.5+ star rating is reassuring and often enough to convince most buyers to take the plunge. However, this doesn't always lead to great results. Here we look at the problems with online reviews.
There are 4 major issues with online reviews:
- Fake Reviews
- Business first models from major aggregators
- Social proof, herding and self-selection
- The eye of the reviewer — does it match yours?
Fake Reviews: Studies show that about 20% of reviews online are fake. Although some of these are fairly obvious to a reader, many of these are sophisticated enough to fool an anti-fake algorithm and contribute to the inflated star rating which is often the only thing the majority of consumer look at. Even if a consumer reads some of the reviews, fake reviews help drown out honest critical reviews. With fake reviews easy to create either by motivated businesses or outsourced to ‘astroturfing' review farms who employ freelancers from low-cost countries to submit hundreds of enthusiastic fake reviews.
Business first models from major aggregators: Many aggregators make much of their money from businesses rather than customers — allowing businesses prominent paid placements on searches out of line with their organic relevance, or even tools to manage or massage their reviews, for a fee of course. Additionally, while they often try to stamp out fake reviews, algorithms are easy to game or bypass — leading to a sharp spike in fake reviews from 5% in 2013 to over 20% today. While they do provide a valuable service, these business decisions do undermine the objectivity and honesty of their reviews.
Social Proof, herding and self selection - Ok, so some reviews are fake, but discerning consumer that you are careful to read through a selection of helpful critical and positive reviews and looking at the big picture before making your decision. However, due to psychological phenomena of social proof and herding, even a handful of well placed fake reviews can significantly bias the perception of a business.
Social proof & herding — consider the phenomenon of ‘salting the tip jar' — where a restaurant with a cash tip jar places a few notes or coins in the jar themselves. Research shows that customers are significantly more likely to contribute tips to a tip jar with some cash in it than to a completely empty one — with the belief that if others thought the service was worthy of a tip they should contribute too. Similarly, the presence of positive 5-star reviews on a service early is more likely to lead to reviews to cluster there. For example, if you were planning to leave a 3 star review but then see several enthusiastic 5 star reviews, you might be more inclined to bump it up to a 4-star review or maybe not leave a review at all. Conversely, if you had a middling experience at an establishment that got review-bombed with negative reviews from competitors, you might turn your 3-star review into a 2 or even 1 star review, or hold back on publishing your 5-star experience. MIT Sloan review explores this in more depth here. Additionally, most people rarely post a review for an ‘as-expected' experience. Generally, highly motivated reviewers, both positive and negative are moved to post, and as businesses are generally smart enough to ‘salt the review jar' with positive reviews early, we end up with a very high number of 5-star reviews, a moderate number of 1-star reviews and much fewer 2, 3 or 4-star reviews, leading to the typical J-cure for online reviews, although studies indicate that the true distribution of most reviews should follow the much more common bell-curve with most reviews at a middling number and tapering off on each end.
The eye of the reviewer — The source of the review is a key data point in evaluating the relevance of the review. For example, ‘the best restaurant in NYC' is likely to mean much more coming from a lifelong resident and fine dining connoisseur than a visiting tourist who stayed for a week. Many people who are generous and open when offering personal recommendations to friends or acquaintances have never posted a review online (or would only do so for extremely good or extremely poor experiences, contributing the the J-curve phenomenon).
