SoylentNews is people
SoylentNews
posted by
martyb
on Wednesday August 08 2018, @09:28AM
from the Fake-News...-Triple-Pun-Intended dept.
from the Fake-News...-Triple-Pun-Intended dept.
Admittedly this is more of a Sla... I mean Soyvertisement. However, I just heard about it this weekend, and it sounds both interesting and useful. I thought others might find it interesting as well
Long story short: ReviewMeta is a website which can be used to try to detect fake reviews on Amazon.
In my defense, it is somewhat technologically interesting. The system supposedly uses various heuristics and algorithms in order to accomplish this, such as searching for suspicious submission patterns, text entered, and timing windows to try to find fakes.
The "news" source I heard it from was a radio/podcast at: https://www.npr.org/templates/transcript/transcript.php?storyId=623988370
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
In Tulsa, Oklahoma, it is against the law to open a soda bottle without
the supervision of a licensed engineer.
(Score: 4, Interesting) by AthanasiusKircher on Wednesday August 08 2018, @03:42PM
Agreed on the "3-star" reviews. Those are often the most informative, because they tend to be balanced and list both pros and cons. Someone who is able to do so is often looking at the product more objectively. 5-star reviews are often worthless, unless it's a truly extraordinary product and/or the reviewer is informed and honest enough to note any minor flaws even while acknowledging the overall goodness of the product. 1-star reviews are sometimes useful for serious flaws, but as you point out, they are also frequently from someone who doesn't know what they're doing (they bought the wrong product for their application, or didn't use it correctly, or didn't actually like what the product does in the first place, etc.) or just has an overreaction to something stupid and unrelated to product quality (e.g., "Shipping speed was terrible!" "Why do they make this product in blue? Blue is stupid." etc.). And yes, occasionally fake reviews from competitors.
So, 2-star and 3-star are where I tend to look most. Such reviewers are often objective enough to recognize a product isn't all bad, but they'll also have some serious criticism.
I've found Amazon's review ranking system has changed for the worse over the years. And I seriously suspect that it tries to hide the more damning reviews. After all, Amazon is in the business of selling products, so having prominent top reviews saying a product is crap isn't generally good for anyone. (And, as any good salesman knows, the few returns you get even from crappy products are often worth the sales to many people who shrug when the thing breaks or never get around to returning.)
For a while, several years back, I wrote a number of reviews for Amazon. My most common ranking was probably 2-star and 3-star, because most of the things I was motivated to write reviews for were things that I thought had some significant flaws but which still had some merit. I tended to try to be objective and give nuance, and several of my reviews got a large amount of positive feedback (some hundreds of upvotes). For a few books, I had the most useful critical review (or whatever they called it) for a long time because of that, and for a couple books my review was ranked first in the list or at least among the top visible reviews.
No longer. Those reviews have now been buried among other negative reviews, and if you even search specifically in 2-star or 3-star rankings, you'll have to dig a while to find mine... even though they are very detailed and have lots of positive votes.
So I stopped writing long nuanced reviews for the most part. Why should I bother when the system buries reviews that are voted most helpful?
(And note that my reviews were older, and I can understand somewhat privileging newer reviews for many products, since manufacturing quality can change, etc. But most book reviews are based on the content of the book, which doesn't tend to change over time -- significant new book editions generally get a new Amazon page. So why should a book review from last month with 2 upvotes be promoted to the main page over one from a few years back with 200 upvotes? My admittedly casual look at this trend seems to indicate that negative reviews age faster and take more positive feedback to land on the top visible reviews list for a product.)