Today I was thinking about this kind of thing, for example Epic not allowing reviews at all, Valve "off topic" review hiding, metacritic changes to their system, or Netflix removing their review system... I concluded that review systems often are on the losing side when their income comes from what they are reviewing, reviews you can trust are those that are unrelated. For example a famous bastion of review is the Mi…
Consumer's Reports is the canonical example of a review site where the incentives line up correctly. The only way they make money is by selling memberships, so the consumer and the customer are the same people.
With things like Yelp [1] and particularly with Glassdoor this is a difficult model to emulate. You generally expect users themselves to contribute quality reviews which you then publish directly (plus maybe do a bit of curation to gather sentiment). It's difficult to both ask users to pay you for the service and also contribute the very content you charge them for.
I think the only way to make a user review site work are:
(a) Be a non-profit and have a donation based model. Would people actually donate though? I'm not sure... but at least you don't seem like a greedy middleman.
(b) Charge a membership fee, but grant free durations of membership for quality reviews. The downside here is that you're likely severely limiting the already small pool of people willing to leave a review.
[1]: The nearest comparison to CRYelp would be Zagat (well, pre-Google Zagat) and Michelin. However, they both only cover a narrow slice.