We have all been there, scrolling through all the reviews before we buy something. You want to see all of this user-generated content, since you are afraid you will regret the wrong choice (Tsekouras, 2017). Also, this information overload leads to being less satisfied, less confident and more confused (Park & Lee, 2009). You could look at the average rating of the product, however these are often bimodal distributed and therefore less helpful (Zhang & Pavlou, 2009). How can you feel confident that you have seen all the important reviews, without having to read all of them?

This is what Ghose & Ipeirotis (2011) studied.

The authors looked at data from Amazon over a period of 15 months to study the impact of reviews on products sales and perceived usefulness. They looked at audio and video players (144 products), digital cameras (109 products) and DVDs (158 products) and their reviews.

The paper identified multiple features that affect product sales and helpfulness, by incorporating two streams of research. First, the information within the review is relevant. Second, reviewer attributes might influence consumer response.

What did they find?

An explanatory study found that the following factors are important:

Thus, perceived helpfulness does not necessarily lead to higher product sales.

They also performed a predictive model, which showed the importance of reviewer-related, subjectivity and readability features on predicting the impact of reviews. Furthermore, the predictive model showed that the predictions were less accurate for experience goods, like DVDs, in comparison to search goods, such as electronics.

What are the managerial implications?

Amazon currently uses ‘spotlight reviews’, which displays the most important reviews. However, it requires enough votes on reviews before a ‘spotlight review’ is determined. The predictive model is able to overcome this limitation, since it is possible to immediately identify reviews that are expected to be helpful for consumers and display them first.

On the other hand, it is useful for manufacturers, since they are able to modify future versions of the product or the marketing strategy, based on the reviews that affected sales most.

The main strength of this paper is that it has relevant managerial implications for both consumers and manufacturers, since it studied both the effect on sales and on helpfulness for consumers.

Would the findings be similar on different websites?

Probably, findings will be similar for other retailers of electronics, therefore Coolblue and Mediamarkt could benefit. On the other hand, book reviews on Bol.com are not expected to have as much benefit from the model, since they are experience goods, similar to DVDs.

Not as straightforward, are the implications for clothing retailers. However, I expect these retailers will not benefit as much from the model, since often there is no overload of reviews on clothing websites and therefore there is no need to reduce the information.

References

Ghose, A., & Ipeirotis, P. G. (2011). Estimating the helpfulness and economic impact of product reviews: Mining text and reviewer characteristics. IEEE Transactions on Knowledge and Data Engineering, 23(10), 1498-1512.

Hu, N., Zhang, J. and Pavlou, P.A. (2009). Overcoming the J-shaped distribution of product reviews. Communications of the ACM, 52(10), pp.144-147.

Park, D. H., & Lee, J. (2009). eWOM overload and its effect on consumer behavioral intention depending on consumer involvement. Electronic Commerce Research and Applications, 7(4), 386-398.

Tsekouras, D. (2017). Customer centric digital commerce: Personalization & Product Recommendations [PowerPoint slide]. Retrieved from Blackboard.

Feature image retrieved from: Enzer, J. (2016, August 17). How to reward product reviews and supercharge your e-commerce business. Retrieved from: http://blog.swellrewards.com/2016/08/how-to-reward-product-reviews-and-supercharge-your-e-commerce-business/