Spam reviews

[NewBuilder]

We would like to see any tips or experience with filtering out spam reviews.

We operate a website where consumers can post their reviews on / experience with certain companies. We do receive a fair share of fake reviews (the "to good be true reviews" about certain companies and also sometimes negative review about their competitors).We filter these as good as we can.

For some companies, we receive a large number of reviews which individually seem to be ok (text-wise but also when checking IP addresses and e-mail addresses), but if you read all these reviews in sequence then there is something off, i.e. they seem to lack the linguistic variance that is normally present in consumer written reviews. We feel someone is trying to manipulate here but this someone is covering his tracks.

Do you have experience in this regard or can you provide us with additional data points to look at?

[donford]

Amazon has a very good model.

User has to have an account <-This makes a person have to register an account use rCaptcha to weed out bot accounts)

User verified purchase (can't review what you didn't buy from them) <- this prevents anybody from posting a review whom hasn't purchased it directly from Amazon or affiliates

Report Inappropriate feature this allows users to report spam and the like, amazon then can look at the post and make a decision to remove it. Instead of looking at 1 million reviews a day you may only have to look at 10,000 spam reports. (example)

This model only works if you have a lot of sales. If you're trying to build unique user driven content then this make not be the best solution.

Hope it helps