Estimating the Helpfulness and Economic Impact of Product Reviews: Mining Text and Reviewer Characteristics

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 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.


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 Engineering23(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 Applications7(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: Crowdfunding for a customized movie experience

Movie theaters regardless of size are facing declining attendance (Business Insider, 2017).  Customers are unwilling to pay a lot of money when they can simply stay home and stream movies. Also, a lot of people’s tastes in movies is just not as “mainstream” but the movies they crave are not shown at most theatres (Business Insider, 2017).

Still, let’s face it, staying home watching a movie is simply not the same as on the big screen. So, what if you let customers demand a movie? is an Australian firm which takes a crowdfunding approach to cinema-on-demand (also available in UK, Ireland, NZ) by enabling users (age 18+) to request a movie event at local partnering theatres. The majority of movies offered are studio classics, indies, artsy- and foreign films. While the concept is not entirely new (see Graphr and Tugg), is the first larger on-demand cinema platform serving markets beyond the US. Moreover, it uses blockchain technology to support producers, distributors and exhibitors with transparent and reliable sales data (Forbes, 2017).

How does it work?


Efficiency Criteria: Weighing Cost And Benefits

For a user, the time and effort needed to place or support a request is really low and tickets only have to be paid by the attendees if the screening is confirmed to happen (leapfrogfilms, 2017). Partnering theatres of course will have to check their schedule and also determine required attendance thresholds.

How does this crowdfunding approach create value for customers, theatres and platform?

Customers co-create by driving the entire process from requesting (incl. time, date and location) to “spreading the word” about the screening of their wishes. If successful, a whole crowd of friends and strangers can finally enjoy “their” movie on a big screen, in a social atmosphere outside the own house.

Theatres help making it happen by approving the request. Further, hence they use capacities more wisely and can generate additional revenue without uncertainty: if the event happens, this money is “safe”. They meet customers’ demand and have the opportunity to welcome (potentially) new customers at their venue. The insights can be a trigger for theatres to think about future events and screenings by illustrating potential demand (Saarijärvi et al., 2013).

Even the filmmakers gain increased exposure and reach for their films; especially those which are not regular movie material.

Finally, Demand.Film only wins if everyone else wins by charging a ticket fee (~ GBP 1.65 per ticket).

The contractual obligations are addressed by Demand.films “terms of use” which reflect standard contract law, all under one condition: if the attendance threshold is not met, the screening will not happen. Naturally, films are legally obtained and of appropriate content. While users can spread the word about a proposed event by any means, all payments and organization are performed through as the intermediary.

What does the future hold for

Naturally, the main goal is expanding the partner and user network as well as geographical reach. Can it be a success? Certainly, never underestimate the power of niche markets and don’t forget about the reach of social media! Will it be a success? The on-demand economy is still on the rise in offline settings and competition never sleeps, so it will remain interesting. All I know is that “I would be in”, would you?



Business insider (2017). “Movie theater attendance is declining as cord cutting becomes more popular”. last accessed 28.02.2017

Forbes (2017). “Tugg And Gathr Face Competition From A New Cinema-On-Demand Platform“ Online: last accessed 28.02.2017

Leapfrogfilms (2017).

Leapfrogfilms (2017). example

Saarijärvi, H., Kannan, P.K. K, Kuusela, H. (2013). European Business Review 25 (1), pp. 6-19

COVER: Leapfrogfilms (2017).




Online Display Advertising: Targeting and Obtrusiveness

We all have been in situations where you are browsing the internet and the advertising is targeted on the content on the website and they are shown to you. Hereby, advertising is targeted. And what is maybe even more intrusive, that the advertising pops up. It definitely gets your attention. However, are you more willing to click on the advertising and buy the product? This is what Goldfarb & Tucker (2011) researched. They study the effect of targeted display advertising and obtrusiveness on sales, and what the effect is when these two are combined.

This question is interesting. Because even with all these new techniques, display advertising success drops. People avoid online display advertising because they infer them in their browsing goals (Drèze & Hussherr 2003). Do obtrusive advertising works exactly in the opposite way and affects the effectiveness of advertising negatively?

To find out, this study uses data from a large randomized field experiment on 2,892 web advertising campaigns. For every campaign on average 852 surveys were distributed. Where the half of them were to consumers who have seen the advertising and the other half were on the website without the advertisement on it.

The main results of this study are that targeting the advertising improves the effectiveness of online display advertising and obtrusiveness does also. However, when these two techniques are combined the effectiveness decreases. This is because privacy concerns temper the appreciation of formativeness in targeted advertising. So, for advertising in categories where privacy matters more, the effect is tempered more than in categories where privacy matters less.

The strength of this paper is the fact that it, in contrast to earlier research, propose that obtrusive advertising is not very effective in the contextually targeted situations. Earlier research study the effect of the obtrusiveness on advertising recall, which is of course positive. By adding the privacy concerns and the feelings of manipulation to the fact that advertising can be perceived as useful makes advertising perceived as intrusive, and therefore result the effectiveness negatively.

This paper shows a reason for the unexpected success of search advertising, where the advertising is highly targeted on the context (the advertising is based on search keywords) but is absolutely not obtrusive or attractive. For managers, this means that in choosing the right way to advertise they must not only consider whether to target their audience with contextually targeted advertising, but also consider the negative influence if these advertisements are obtrusive. Economically, 5.3 percent of advertising spending could be cut, without affecting the effectiveness of advertising. This, solely because the wrong combination of advertising content and format is used.

In short terms, either choose to reach your audience with targeted advertising, or with obtrusive advertising. But don’t combine the two.

Drèze, X. & Hussherr, F.X., 2003. Internet advertising: Is anybody watching? Journal of Interactive Marketing, 17(4), pp.8–23.

Goldfarb, A. & Tucker, C., 2011. Online Display Advertising: Targeting and Obtrusiveness. Marketing Science, 30(3), pp.389–404.