Amazon was the first ecommerce site to let consumers post online reviews and ratings of products in 1995, a move that critics derided as “retail suicide.” Now consumer reviews and ratings, more popularly known as user-generated content (UGC), have become a force to be reckoned with. Prospective buyers key off two numbers: the average product rating, important to 54% of consumers; and the total number of reviews, important to 46%. Understanding how consumers process and weight these two variables is critical to winning the retail ecommerce war.
Highly-rated and reviewed products may not always be the best choice for consumers. They may end up selecting older products that have high reviews but outdated features, or lower-rated products that generated increased review counts with steep discounts. In addition, brand-new products may have no or few ratings and reviews, but still be superior to other market offerings. A new study in the Journal of Marketing sheds light on UGC to help retailers optimize the way they position products and drive sales on online marketplaces when conditions are not optimal.
Our team studied consumer choice between multiple products that varied on average ratings and number of reviews, but had similar product attributes. This decision was driven by the fact that retailers increasingly provide consumers with product options within a choice set instead of individual options. Consider Google Shopping’s product search pages, Amazon’s “customers also shopped for” selections, or any online retailer’s “recommended for you” lists. With these formats, consumers aggregate review information for multiple choice options simultaneously rather than considering different products’ merits one by one.
Our research team hypothesized that consumers’ diagnoses of average product ratings and the number of reviews are not fixed and may depend on the value of each attribute. We focused on aggregate reviews because they are more accessible than individual reviews and thus are more likely to be considered in decision-making. We hypothesized that consumers are most reliant on the number of reviews when the level in a choice set is low versus high. While a low number of reviews should signal a lack of definitive evidence to consumers, they are actually more influential in choice-making. We used seven studies to test our hypothesis, including traditional lab and eye-tracking experiments, simulations, and secondary data. In so doing, we analyzed more than 2.5 million products across 24 product categories and their corresponding choice sets, using data collected from Amazon, the leading online retailer.
Key findings include:
- Average product ratings are a more diagnostic cue of product quality than the number of reviews.
- Consumers will choose a lower-rated product with more reviews over a higher-rated product with fewer reviews when the choice set’s level of review numbers is low. However, they will choose the higher-rated production when the choice set’s level of review numbers is high.
- Consumers are more likely to defer making a purchase decision when the choice set’s level of review numbers is low versus high. However, they are less likely to defer choice when the number of reviews is zero as opposed to low.
Online retail and product managers can use this research to market products and position them effectively. Many retailers encourage their consumers to review products by offering discounts or freebies. This is especially critical for new products that have lower marketplace visibility. These strategies will drive the overall number of reviews, but may not result in improved average product ratings. Thus, managers may want to trade off the potential benefits of receiving additional reviews against the marketing costs of acquiring them.
In addition, retailers could elect not to show low review numbers (as most retail websites do), keying off the fact that consumers are more likely to buy a product with no reviews than a product with bad reviews. Retailers who pursue this strategy would be trying to promote targeted product sales instead of motivating consumers to make suboptimal decisions by choosing older products (which may have more reviews) or deferring the purchase decision altogether.
From: Jared Watson, Anastasiya Pocheptsova Ghosh, and Michael Trusov, “Swayed by the Numbers: The Consequences of Displaying Product Review Attributes,” Journal of Marketing, 84 (November).