Amazon Adopts Machine Learning To Deliver More Relevant Product Reviews
One of the challenges of online shopping is navigating a product's user reviews for fair assessments of its pros and cons. That's no easy task -- some people are habitual complainers who overstate the negatives, while others could be shills for a competitor. Still, some can be helpful, and to ensure those are the ones you see first, Amazon is going to use a machine learning platform to determine and uplift the better evaluations.
This is a proprietary system that Amazon developed on its own. The idea is to prioritize newer and more helpful reviews so that they're the first ones you see and read when researching a product. In doing so, Amazon could potentially save shoppers lots of time by negating the need to dig through pages and pages of user reviews looking for trustworthy write-ups.
"The system will learn what reviews are most helpful to customers...and it improves over time," an Amazon spokeswoman told CNET in an interview. "It's all meant to make customer reviews more useful."
Amazon's new system will also affect a product's star rating on the site. Users can rate a product on a 5-star scale, and in the past, Amazon would average those ratings. Going forward, the machine learning platform will use a weighted system when determining the average, which can and will result in existing star ratings changing for a product, and changing more often.
"It's just meant to make things that much more useful, so people see things and know it reflects the current product experience," the Amazon spokeswoman added.
One of the ways this will benefit both manufacturers and customers is when a product is updated to address previous complaints. For example, let's say several customers complained that a flux capacitor from Company A required too much plutonium. A later revision might cut the plutonium requirement in half, rendering the initial complaints irrelevant. Amazon's weighted system would look at what the newer reviews had to say and give this more weight when averaging the flux capacitor's star rating.
This is a proprietary system that Amazon developed on its own. The idea is to prioritize newer and more helpful reviews so that they're the first ones you see and read when researching a product. In doing so, Amazon could potentially save shoppers lots of time by negating the need to dig through pages and pages of user reviews looking for trustworthy write-ups.
"The system will learn what reviews are most helpful to customers...and it improves over time," an Amazon spokeswoman told CNET in an interview. "It's all meant to make customer reviews more useful."
Amazon's new system will also affect a product's star rating on the site. Users can rate a product on a 5-star scale, and in the past, Amazon would average those ratings. Going forward, the machine learning platform will use a weighted system when determining the average, which can and will result in existing star ratings changing for a product, and changing more often.
"It's just meant to make things that much more useful, so people see things and know it reflects the current product experience," the Amazon spokeswoman added.
One of the ways this will benefit both manufacturers and customers is when a product is updated to address previous complaints. For example, let's say several customers complained that a flux capacitor from Company A required too much plutonium. A later revision might cut the plutonium requirement in half, rendering the initial complaints irrelevant. Amazon's weighted system would look at what the newer reviews had to say and give this more weight when averaging the flux capacitor's star rating.
