I've been sampling a lot of video blogs the past couple weeks (trying to fill my Amanda addiction). Video blogs seem to all fall into several categories:

Teen Angst: This is the classic point-a-webcam-and-talk format. These blogs uniformly suck, mostly because almost nobody can ad-lib for three minutes and come up with good material a reasonable fraction of the time. Robin Williams might be able to pull this off, but probably nobody else on the planet.

Shrinky-Dink TV: These are commercial video blogs trying to follow the model of the TV show, except in a smaller, shorter format. Blogs like this usually feature fancy (expensive!) sets, scripted material, and paid actors. These usually suck too, since the writing and delivery tends to be really flat. You can't take the same kind of material which works in a 30-minute time slot on a 36-inch screen and merely scale it down. It doesn't help that the writers and actors are basically the people who can't get jobs working on TV.

Weird Stunts: Ordinary folks who do something bizarre on video. These are usually (but not always) planned, and can be really funny. An unbundled version of "America's Funniest Home Videos." Usually these don't qualify as true video blogs, since most are one-off. Very few people have the time or capacity to come up with creative, bizarre things to do on a regular basis.

New Ground: The best video blogs seem to be the ones which break the rules and create their own genre. I consider Rocketboom and Ask a Ninja to fall into this category. Of course, merely breaking the rules isn't enough to have a good show, but because we don't yet know what the rules are for this medium, anything good is almost by definition going to have to be doing something different.

A few things which seem to be critical:

1) Good writing.

2) Overacting. Acting which works on a TV set will look really flat and unengaging in a 2-inch window.

3) Grab the viewer. 5 minutes seems to be the limit for a video blog, and you have to engage the viewer right away to fit anything into that time.

And somethings which aren't so important:

1) Production value. At this stage, viewers will forgive a cheezy set, bad sound, mediocre lighting, etc., if the material is good.

2) Consistency. It helps to be above average most of the time, but there's still a lot of room to experiment and make mistakes in this medium.