Given the nature of reviews based upon personal experience, I would never make a purchasing decision on the content of any specific review(s). Instead, I look for trends and what the "community" at large has to say about a product or vendor.
Exactly as it should be done (or at least applied). The overall "picture" created by reading the reviews is one of the ways I judge the suitability and quality of any item I intend to procure, whether it is for work or ham radio. One line "5 star" ratings are immeadiately squashed as non-informative (i.e. 'had it for a week and it has not failed' - ok but how was it used?). Same with low count quantity "1 star" gripes.
If the gripe is repeated over and over again then it requires more scrutiny as to why it is occuring but single (or two or three) complaint could be seen as 'the inability of the user to understand and follow directions correctly'. If a product has the typical bell curve of user reviews but is skewed towards the low end of the scale, then that suggests that either the product indeed may have a problem(s) and may need to be investigated further before deciding to procure. If it is too high, then it also needs to be closely looked at since there is no such thing as a 'perfect' product. The higher the price, the closer I scrutinize the high level reviews. No one wants to admit they bought a lemon for a bucket of money!
The number of review samples also is a big motivator in how the reviews are analyzed. A small set of reviews might not reveal the real truth about a product in as much as a very large sample set would.
Bottom line, it is up to the reader of the reviews to apply common sense and read their own 'gut' feeling about the product under review. I have never been burned YET from taking a sound logical (and educated) approach to deciphering on-line reviews. But nor do I base decisions solely on them either.