Paper Motives

Do customer reviews drive purchase decisions? (2017)

• Average review score has a positive effect on probability of a purchase. This effect is more profound when there are many reviews than when there are few.
• When customers read review content they are more likely to be affected by the average review score and by the volume of reviews
• Additional hypothesis proven about high price items being highly influenced by reviews, especially when customers read them and not just look at average rating / volume.

The Value of Online Customer Reviews (2016)

• On average the conversion rate for a product increases by as much as 270% as reviews are written for it.
• High price products can have their conversion rate increased by 380% due to reviews, and 190% for low price products.
• Customers focus on the first few reviews available.

Anonymity, Social Image, and the Competition for Volunteers: A Case Study of the Online Market for Reviews (2010)

• This provides a really good insight into Yelp activity, although it is old.

Fake It Till You Make It: Reputation, Competition and Yelp Review Fraud (2015)

• Based on Yelp's spam detection algorithm, one out of five (20%) of reviews on Yelp are fake.

The Impact of Fake Reviews on Online Visibility:

• Fun to read, some of it is like a story book. Background section has some really useful information.
• One hotel can become more visible than another hotel after posting just 50 fake reviews in 80% of cases
• Sites use proprietary rating functions, which allows fake reviews to have greater impact because the function weights it's output mostly by recent reviews.
• On TripAdvisor dataset

The Consequences of Fake Fans, 'Likes' and Reviews on Social Networks (2012)

• Could not access, but apparently content predicted 10-15% of likes, fans and reviews were fake in 2014.

Detecting Deceptive Reviews using Generative Adversarial Networks

• Recently shown that a cetrain GAN architecture is effective at detecting deceptive reviews.
• To the knowledge of the paper this is the first work to use GANs for this purpose.
• We would like to create a comparison of this
• We can evaluate it on a much larger dataset, this paper uses 400 truthful and 400 deceptive reviews

Review Spam Detection

• As original as it gets, shows that this is an established and challenging problem.
• States that 52% of reviews it checks are spam, but I don't think this is very reliable.

The Importance Of Online Customer Reviews [Infographic]

Added after initial draft

"Studies on Yelp.com have shown that an extra half-star rating could cause a restaurant to sell out 19% more products (Anderson and Magruder, 2012)," from Learning to Represent Review with Tensor Decomposition for Spam Detection

"A one star increase leads to a 5-9% profit increase (Luca, 2011)." from Learning to Represent Review with Tensor Decomposition for Spam Detection

Section Draft:

Online reviews of products and services have become significantly more important over the last two decades. Reviews sway probability of purchase through review score and volume of reviews [1]. It has been found that on average the conversion rate of a product increases by 270% as it gains reviews. For high price products, reviews can increase conversion rate by 380% [2]. In competitive, ranked conditions it is worthwhile for unlawful merchants to create fake reviews. For TripAdvisor, in 80% of cases a hotel could become more visible than another hotel using just 50 deceptive reviews [3]. Faking reviews is an established problem [4] and has been exploited as it was found that 1 in 5 (20%) of Yelp reviews are marked as fake by Yelp's algorithm [5].

1. Do customer reviews drive purchase decisions? (2017)
2. The Value of Online Customer Reviews (2016)
3. The Impact of Fake Reviews on Online Visibility (2016)
4. Review Spam Detection (2007)
5. Fake It Till You Make It: Reputation, Competition and Yelp Review Fraud (2015)