top of page
Five Star Logo-02.png

Using Amazon Reviews to Inform Product Development

When we do market research, we seek to understand how our clients’ customers feel about their products or services. What they love, what they hate, and what motivated them to part with their hard-earned money. This is how we use qualitative analysis to power market research for our clients. 

What We Did

A well-designed consumer survey may fetch 1,200 responses of quality insights; however, it comes with a cost as surveys take longer to plan out and implement. Alternatively, an e-commerce platform like Amazon may have well over 30,000 reviews full of customer insight, for free. So, we analyzed the written portion of reviews left on our clients - and their competitor's - products, to understand how customers feel about the industry and which companies they feel are excelling in the space.

Icon_Data chart.png

Showed strengths and weaknesses for each company by segmenting key themes.

Icon_Dashboard (2).png

Revealed key positive and negative themes for each product and company.

Icon_Map (2).png

Discovered distinct customer personas from the relationship between key themes.

3 Key Insights

Strengths and Weaknesses

Filled in Table for Mouthguard case study-02.png

We searched the reviews for the top positive and negative themes. By performing this analysis with all the products on the market we were able to understand what issues customers generally have with these products, and what they feel makes the great products stand out. 

Segmenting themes by the company allows us to show our client the strengths and weaknesses of their competitors. The visualization on the left makes it easy to spot areas where each company is excelling, or where there are opportunities to improve. By comparing this to the issues that matter most to customers, the important next steps are clear. 

Needle in a Haystack

Online reviews can be difficult for product managers to interpret. Most comments don’t add any real value to the discussion, many just say something like “terrible product” without offering an actionable reason why.  


Adding to this problem is the fact that online reviews suffer from the effects of the positivity bias phenomenon. Put simply, it is human nature to give reviews a more positive spin and avoid negative comments, or to give products a more positive rating than the written feedback would suggest. It might be the 5-star review that contains a key sentence of constructive feedback for a company to build from. 


By using text analysis to analyze all the available reviews, we can find those 1 in 1000 gems that help product managers to design the next winning product.

Highlighted Example Amazon Review-01.png

Their Biggest Concerns

Cluster Map Mouth guards-01.png

When a customer comments on their love of - or frustration with - a specific product feature, they’re rarely telling the whole story. Usually, there is more than one factor that contributes to making that feature a positive or negative experience.  


Using statistics to determine the likelihood of two or more comments being related, we’re able to show distinct clusters of themes. From these clusters, we can create unbiased personas that reflect the different types of people who purchase a product, and what features are likely to matter to them most. 

bottom of page
Privacy Policy Cookie Policy