Most of my experience has been with book reviews. In that genre it's taken for granted that the publishers give the books to the journal editor, who then passes them to the reviewers. The reviewer gets to keep the book - which normally is the only payment he/she gets. Anybody who hopes to make substantial money off reviews is in for a rude awakening. And in the academic realm, anybody who tries to brownnose somebody with a review can expect to lose his job if his colleagues find out.
I have mixed opinions about in-the-box reviews. They certainly don't tell us as much as a review by somebody who's built the model. On the other hand, they can reveal a lot of useful information - if the reviewer is knowledgeable and the photos of the parts are clear.
My favorite modeling genre is sailing ships. Few good reviews of those kits get published. One big reason: if the reviewer takes the time to do a decent job of building it, by the time he/she finishes, the kit won't be new any more. And quite a few of the sailing ship reviews that do get published are, I'm afraid, written by people who don't really know much about the subject.
I once watched a promotional video from a major ship model distributor/manufacturer that featured a wise looking man in a shop apron "lifting the lid" of one of the company's sailing ship kits. After the first minute or so it became obvious that the gentleman had no idea what any of the parts were. In fact I got the impression that, prior to this "lifting of the lid," he'd never laid eyes on the thing. He even mispronounced the name of the ship. That video was utterly useless. (I wound up buying the kit anyway - but not on the strength of the video.)
I have to say I find written reviews illustrated by still photos just as valuable as videos. I keep up with the reviews on modelwarships.com, and have based buying decisions on them several times. (I have to say, though, that being told a detail is "pretty good" doesn't do much for me.)
One common feature of reviews does bug me a tiny bit. (FSM doesn't do it often, but it's a regular feature in many of the British journals.) The reviewer starts by telling me how many sprues are in the box. Why on earth does that make any difference?
I much prefer FSM's approach of listing the number of parts, in a sidebar. A huge parts count obviously doesn't guarantee a good kit, but if one kit has 75 parts and another one depicting the same subject on the same scale has 350, that tells me something useful.
Bottom line: I think in-the-box reviews are highly useful if they're written knowledgeably and well illustrated. But I'm not convinced that video reviews are much preferable to written ones. I'll be happy to be proven mistaken.