Making a review is easy: Sit down, talk about a product and upload it. Making a good review however is a lot more difficult, so here are some necessary things, optimizations and things to avoid you should keep in mind.
(This post assumes that you've already read part 1: Becoming a YouTuber: The Basics and defined your audience and figured out from which angle you'll tackle your reviews)
Be subjective. Being subjective is the point of a review, it's about your subjective experience you've had with the product, whether you like it and whether you think it's worth its money, in short: it's your opinion on the quality of a product. Being subjective doesn't mean being arbitrary, you will have to explain how you came to your conclusion. You'll probably have to use some facts as a foundation for you opinion, but don't just dump facts into the videos under the guise that it's going to make your video more objective. Again, objectiveness isn't the point.
Be transparent on how you got your product, ie whether you bought it yourself or got a free sample.
Never review a product that is sponsoring your video.
As a reviewer or critic, you have some journalistic duties and ethics to consider and perhaps even legal ones.
Have your own opinion. While it's not forbidden to read other reviews before doing yours, it certainly is not really useful because you'll be watering down your opinion with other's opinions. Further, it can happen that you accidentally plagiarize, which most definitely is a death sentence in any sort of journalistic outlet.
Be fair and true and back up opinions with arguments, and arguments with facts. If you are praising or hating on products but can't really back it up with arguments, you'll quickly lose any reputation as a reviewer and at best will be good for entertainment. If what you're saying isn't true, the same happens, plus you will get into legal trouble sooner or later.
Have a conclusion. Typically, this includes some sort of rating system. It doesn't necessarily have to be a star system or out-of-10 thing, I actually wouldn't even recommend them due to their rather arbitrary nature, but a simple "recommend/don't recommend" and maybe an additional "only recommend if you like the genre" and "only recommend if the price drops"
Keep your review concise. Not necessarily short, but always to the point.
Structure your review. This doesn't necessarily mean that you'll need to have distinct sections in your video, but finding out how to group what you're going to say is a good idea. If you don't do this, you'll risk having arrived at your conclusion, only to quickly throw in a short "btw, I found character X terrible" at the very end, which doesn't really fit there and would've been better when you were discussing why you didn't get attached to any character or whatever.
Having a script helps with the two aforementioned points tremendously. In fact, having a script is almost mandatory because else it's rather difficult to get the opinions you have in your head into any form of linear media (which includes video).
Show, don't (just) tell. If you're criticizing a thing that actually is visible, try cutting to some footage of the thing you're talking about while you're talking about it.
Beware of copyright on creative works. Showing as little of the copyrighted work as necessary usually prevents overactive filters from troubling you too much.
Include nit-picking. What is a minor annoyance for you may be a dealbreaker for someone else. That said, if you think that the entire product is garbage anyways, there's little use to include every single nit-pick because you'll already have a lot of major negative points anyways.
Reviews are at the intersection of hub and help content. Meaning that you both can appeal to people who already watch your content and new people who come via search.
To optimize for newcomers, make sure you don't have too much information that only are useful for fans in your reviews, especially not right at the beginning. Also, SEO matters here a lot.
To optimize for long-term viewers, try making your videos more entertaining, rather than simply a utility that helps people form an opinion.
Be on time. It generally doesn't make as much sense to review a product that's been out on the market for a long time and is due to be replaced by its successor in 3 months or so anyways.
that said, there is a niche for "retro reviews".
Things to avoid
Avoid summarizing the content of creative works. Not only will spoil this the experience for anyone still wanting to watch the thing, a summary of the content also isn't really helpful for the viewer as the summary says very little about consistency, cinematography, pacing, gameplay, wording, acting, etc., ie the actual qualities that allow you to distinguish between bad and good creative works. Instead, comparing the work to other works ("it's a bit like film X") or even genres ("it's a fast-paced horror game") may be more useful. Reviews are not summaries.
That said, including the premise of the work typically is a good idea.
Warning people before a spoiler comes also is a good thing that people generally will thank you for.
Avoid large fact dumps. This is especially true for technical products, you generally don't need to read out the entire product technical sheet in order to have someone follow how you came to your conclusion. Reviews are not readouts of advertisement texts and other stuff the product may ship with.
Adding your own facts (benchmarks, somewhat standardized tests) however is generally good.
Avoid relying on first impressions. Reviews are not unboxings. You should be at least using the product for a couple days, or better: weeks before forming the opinion you'll include in your video.
Once you have made some reviews, it's time for part 3: optimizing the channel