From a hobbiest's perspective, I think it takes several factors to reach "critical mass."  Personally, I take into considerations these items, although weighting varies and other factors can apply as well.

# of good reviews from varied hobbiests.  Bonus of they seem to be frequent or consistant reviewers.  Some of us less experienced hobbiest do rely on the more active "seasoned" people.
I. E. Not a vicarious experience, but rather a benchmark.  If John Q. liked Jane Doe, and I later visited with Jane Doe and had similar results and similarly pleased, I can further relate to John Q.'s reviews.

Time frame of reviews - recent, old, all at once, or spread over time.  Consistant reviews over time help to eliminate several concerns regarding legitimacy, authenticity, experience, etc.  A surge of 3-5 reviews over one week or month might appear more "calculated" with an ulterior purpose in mind.

Accurate, detailed, objective reviews.  Nothing throws up red flag like "Superlatives" in all catagories without providing examples and descriptions.

Realistic reviews, something others can relate to instead of seeming so out of personal experience.  Obviously there are some excellent providers out there, and we would all love to run into them, but we're only human, one person's 6 is another's 10.  YMMV is a valid statement, and means that not everyone's perspective is the same.

Web site is up to date and reinforces review details, specifics, costs, etc.  Sometimes there are discrepancies between those two sources of information.

I'm sure I left some out...

Leco

There are 3 reviewers on TER that I implicitly trust.  A single review of 8, 9 or 10 in performance from any 1 of those 3 would be enough for me to visit with that provider.  This has happened only twice, however.  

So back to the real world, where things are a bit murky.  I try to look at more current reviews only, and if a number of reviews are present I try not to look at those more than 6-8 months old.  People change.  Assuming for the moment I don't know any of the reviewers:

If a provider has, say 5 reviews of reasonable consistency, say all 7s and 8s, and 3 or more of the reviewers have multiple reviews in their history, and the performance marks in their histories are varied (not everyone got a 7!), then I would probably consider this to be a well-reviewed provider.  If their histories of experiences include 80-90% good reviews, the balance being so-so or bad, it would ADD credence.  If, on the other hand, more than 25-30% were bad, that would, IMHO, decrement the validity of that reviewer.  Us guys can be YMMV as well.

If a provider had the same 5 good reviews and 1 bad review, I would "do an Olympic" and knock off the worst and best reviews and just look at the 4 remaining, giving the provider the benefit of the doubt on that 1 bad mark.  However, if you carry this further when more than 1 bad grade is present, then we are no longer consistent.  The more good grades a provider has, the better her profile will tolerate poor grades.  But.....

Regardless of how many positive, excellent reviews she may have, if there are more than 5 or so bad reviews, especially grouped chronologically close, I consider the prospect to be too much of a YMMV, which I attribute to the provider, and decide not to visit her.

If I do know any of the reviewers, or reviewers have written about providers I have seen, and we are in some general agreement in our opinion re this common provider, then his reviews take on a more weighted value in the process.

I guess this is my best attempt to make the selection process Rocket Science.  Thank God is can't be science!