Here's the problem. Some reviewers may depend on manufacturers for free samples and advertising revenue. That being the case, they are NOT going to give a bad review. The reader has to infer what the reviewer is trying to say (without actually saying it). The same thing is true in pharmacological/medical research. The researchers get BIG bucks from the Pharma so the results are always 'spun' to look good for the product. If you find an occasional unbiased reviewer whose interests are not conflicted, you have a true gem.
Plus, the review only covers a short period of time. No one is going to wait a year to do a review and say, "oh, by the way, after you get 500 rounds thru this gun, you have to replace the RSA and it costs about $150." Or in medicine, "Oh, by the way, that Pradaxa we put you on, it turns out that in increases your relative risk of heart attacks by 25%."
And, the plural of anecdote is not 'data." Reviews are really anecdotal evidence based on one person's sample of one gun, one time. Maybe, if you read 10 reviews that were all 'unconflicted' and they all said the same thing, you might be closer to a real useful review.