Inbound Internet Marketing Blog

SEO, Blogging, Social Media, Landing Pages, Lead Generation and Analytics

SUBSCRIBE

The HubSpot Inbound Internet Marketing blog covers all of inbound marketing - SEO, blogging, social media, lead generation, email marketing, lead nurturing & management, and analytics. Join 53,183 others and subscribe now!

Subscribe to RSS feed Add us on Facebook! Follow us on Twitter

Get Free Marketing Info!

Get the world's best marketing resources right to your inbox! Join more than 817,000 inbound marketers!

Subscribe by email

Your email:

Listen to this blog!

Work at HubSpot!

JoinTheHubSpotTeam resized 200

HubSpot's Inbound Internet Marketing Blog

Current Articles | RSS Feed RSS Feed

Who Writes Amazon Product Reviews? [Data Analysis]

 

.

If you engage in regular ecommerce online or run an ecommerce website, you likely use and understand the importance of Amazon product reviews. Do you know where these reviews come from, and who writes them?

amazon logo

In this post, we break down the geographical characteristics of Amazon.com's top reviewers using an Amazon review dataset from the computer science department at the University of Illinois at Chicago (special thanks to @arnabdotorg ). Which geographies drive the Amazon product review sphere, and which are under-represented?

Location-based Findings About Amazon Reviewers

The metrics we generated about Amazon reviews are ratios that can be understood as:

% of Amazon top reviews from a state / % of US population in that state

For a given state, a ratio of 1.00 can be read as "this state has the same contribution to Amazon's top reviews as it has contribution to the US population." Ratios above 1.00 indicate that a state's contribution is higher for Amazon reviewers than the US population; for those below 1.00, the state's population is under-represented in Amazon product reviewers.

First, the states that highly over-index:

Amazon review data top ratios

Vermont, Maine, and Washington state, followed by Alaska and New Hampshire take the lead in Amazon product reviews as compared to each state's US population contribution.

The lowest ratios of Amazon product reviewers per population contribution:

Amazon review data low ratios

West Virginia, Mississippi, Arkansas, and South Dakota are followed by Louisiana, Kentucky, and Alabama for the lowest indices of Amazon product reviewers as compared to US population. These states are under-represented in Amazon product reviews as compared to the number of people living in those state.

Also compiled is a full list of these Amazon-reviewer-to-US-population stats to see how all fifty states stack up.

Speculation and Future Investigation

The first correlation attempted with these ratios was to examine broadband penetration by state (as reported by the FCC: pdf link). It turns out that the statistical relevance is quite low between broadband penetration and Amazon review concentration. This is in itself is interesting: there are other, more relevant factors to understanding why certain areas of the US have high or low relative concentrations of Amazon product reviews. Broadband Internet usage and availability is not one of these factors.

489992128 f386d3f204 resized 600A further hypothesis is the role of geography itself. This investigation (using average state latitutde) yielded a better correlation than the broadband analysis, but ultimately is not very illuminating. That is, it's readily apparent that the top indexing states are more northerly than the under-represented southern states. The geography hypothesis raises more questions than it answers.

The reason for this is that geography is not the defining mechanism but rather the emergent characteristic of the dataset. Naturally, it is also how the data is organized for this analyasis. As far as the actual mechanisms that cause people in Washington to write a disproportionate number of reviews as compared to hose in Louisiana, the potential answers are varied.

Could it be that colder weather and shorter days in the Winter lead to a more indoor, Internet-centric lifestyle? Does this in turn cause people to spend more time reviewing products? And what about a state population's engagement with ecommerce sites in the first place? There are myriad, intertwined micro-cultural mechanisms at work here, even if we can manage a categorization based on geography.

Marketing Takeaways

For ecommerce site owners and as well as consumer marketers, keep the following in mind while reading Amazon product reviews:

  1. A randomly selected top review from Amazon is more likely to be from Vermont or Washington than Louisiana or Arkansas.
  2. A product's most influential reviews may not jive with your view of the world.
  3. A product's most influential reviews may not be a representative sample of the "average" US consumer, even if there are thousands of reviews.
  4. Geography is important. Whether it's ecommerce or otherwise, optimizing for and being considerate of geographic differences is key for marketing success.

Rebooting Your Ecommerce Marketing in the Social Media Age

Rebooting Your Ecommerce Marketing in the Social Media Age

Learn how to integrate social media with your ecommerce marketing strategy!

Download the free guide for tips to drive more leads to your site and convert more sales for your business.

Posted by Michael Redbord on Wed, Jun 23, 2010 @ 08:00 AM

COMMENTS

I think the conclusion is obvious - Mainers have lots of time on their hands, and West Virginians are illiterate.

posted on Wednesday, June 23, 2010 at 9:15 AM by Stanley McC.


@Stanley - could be. WV, MS, and AR are the poorest states according to http://factfinder.census.gov/home/saff/main.html?_lang=en. Less ecommerce, less online spending, lower literacy, etc. are all correlated with income.

posted on Wednesday, June 23, 2010 at 10:05 AM by Michael Redbord


Both commenters probably have it right. This information juxtaposed with literacy rates and economic data would probably result in some significant correlations. Maybe the Census need only look to Amazon reviews for their information.

posted on Wednesday, June 23, 2010 at 11:15 AM by Milind Shah


Question is - why is this analysis occurring and what's the objective? Social study or evaluation of the effectiveness of reviews? (And therefore the money spent to implement the review feature) Why is effort being spent to correlate this data? 
 
 
 
Speaking as a marketer, I would be more inclined to care about the cost of the feature and its ROI - so I am speaking to that. 
 
 
 
At the end of the day what makes the Amazon reviews work is whether they are actually perceived as useful in helping to select products, thereby converting a greater percentage of prospects into buyers.  
 
 
 
So long as useful data is being contributed, is it relevant to worry about where it is coming from or in what volume? What's the ultimate goal? To just increase the volume, or to get the dispersion of input more geographically balanced? (And if so - does geographic location of the review impact the buyers' behavior? If not - wouldn't that normally bring this line of inquiry to a screeching halt?) 
 
 
 
I'm admittedly outside of academia and unabashedly coin operated, so I am asking these questions out of curiosity. I have never been a position to just ask "why" unless it might have an impact on the (immediate) bottom line. I am interested in knowing how others look at this kind data and why. 
 
 
 

posted on Wednesday, June 23, 2010 at 11:59 AM by Parker Pearson


Yes i totally agree with the weather factors. They do influence internet usage being that if it is very cold people tend to stay in more and ultimately use the internet.

posted on Wednesday, June 23, 2010 at 12:20 PM by rodney akomas


@Parker - great questions!! 
 
The purpose of the analysis is to be asking exactly these sort of questions, e.g. How does a geographical bias in Amazon reviews affect marketers? 
 
As a simple answer, it affects us insofar as the Amazon reviews affect our product/company's perception and eventual success. If a product is specifically targeted or useful to people in the the lower-ranking states, it's likely that the product 1) won't get as many reviews 2) the reviews it gets are less likely to be "relevant" in the sense that they are from the target demographic.

posted on Wednesday, June 23, 2010 at 1:00 PM by Michael Redbord


I didn't know this blog is meant for US readers only. 
 
Maybe I'll just cancel my feed subscription and find another blog that recognizes that US <> world. 
 
Cheers, 
 
Jurgen (from Europe)

posted on Wednesday, June 23, 2010 at 1:16 PM by Jurgen Appelo


@Jurgen 
 
Has the editor of this blog I can ensure that this blog is not soley US focused. Sure this post is, but the vast majority of our posts are applicable regardless of the country you live in. We are also constantly working to improve content for readers outside of the US. Thank you for your feedback!

posted on Wednesday, June 23, 2010 at 1:22 PM by Kipp Bodnar


@Michael - Thanks!  
 
 
 
This type of issue may not impact me now, but your challenge may cross my path in the future. Which is in part why I try and step out of my usual comfort zone a few times a month and explore things like this. Thanks for the feedback!

posted on Wednesday, June 23, 2010 at 1:26 PM by Parker Pearson


I suspect there is a relatively strong correlation between reviewer states and levels of education.

posted on Wednesday, June 23, 2010 at 5:12 PM by E. Dean Butler


@Dean  
 
Certainly. Factors like education level, income, and ecommerce usage are all intrinsically related.  
 
@Jurgen  
 
I wouldn't do that just yet! HubSpot has tons and tons of content for non-US readers. This particular analysis is limited to the US because Amazon is organized geographically--Amazon.com for the US, Amazon.ca for Canada, Amazon.fr for France, etc. The dataset I used was from Amazon.com, had some international data, and I stripped it out for this analysis to normalize only the US states among themselves. I've got another dataset for reddit.com that speaks to the international population a bit better :) 

posted on Wednesday, June 23, 2010 at 5:46 PM by Michael Redbord


@Michael: 
 
My point is that the article promises a geographical breakdown of Amazon.com, which is a _global_ web site. Your assumption that Amazon is organized geographically is plain wrong. There is no Amazon.nl for example. Only a few big countries have their own local Amazon. People in most smaller countries (like mine) purchase stuff on the global .com site. 
 
The rest of the world is often ignored by US citizens, even when they have a global audience. 
 
Some indication in the article of what you did with the non-US reviewers in your sample would have been appropriate, I think. 
 
Sorry about the rant. This is a pet peeve of mine. 
 
Cheers, 
 
Jurgen

posted on Wednesday, June 23, 2010 at 6:23 PM by Jurgen Appelo


Interesting data on the valur of product reviews. Thanks

posted on Thursday, June 24, 2010 at 3:24 PM by wayne


As a new affiliate marketer marketing Amazon products on my website and writing my own product reviews, I found this article very useful.

posted on Tuesday, July 06, 2010 at 1:12 PM by Ken


Comments have been closed for this article.