On the one hand, this article underestimates the tremendous effort required to put together a good video, especially if you're not a visually talented person (i.e. someone who can scribble stuff and make it look good). On the other hand, Woobius managed to put a number of videos that looked fairly professional, all on a zero budget (see here and here and here), so it certainly is possible.
If you have access to someone (a cofounder or even simply a friend) who is visually talented, this article can take away some of the mystery of putting together a video for your product. Since videos do tend to increase conversion, this may well be worth doing.
I'd say there are two points in a startup's life where it makes sense to throw together your own video. One is before coding anything, as part of a "video MVP" to test the market. There, it's really important not to invest too much time in the video, since it's really just a simple test of the market demand.
The other point is when you've got a website up, and you're trying to increase conversion rates. There, it makes sense to invest a bit more time into the video, but you may still be at a stage where cash is tight and you don't want to spend thousands of dollars on a "proper" video, especially if you don't know whether it'll actually help conversions in your case.
Later on, when you have the financial resources, hiring professionals to put the video together is a better option.
If you read this far, you should follow my RSS feed here.