Let's take one of them as an example. The company launched with 7 different websites, was self funded by a previous successful exit of one of the founders. After about 6 months it became clear that one of the sites was becoming the leader in terms of growth.
So they divested 3 of the remaining six and literally killed the other 3 to be able to focus on the one that mattered.
They ended up doing very well on the last one.
The catch here is that this company was started with a lot more money than most start-ups and could afford to run effectively 7 start-ups in parallel.
For most other parties this would have translated in on average three failures before scoring a hit, and even that figure is way above the industry average.
If you are talking about simple web apps with quick development times and maintenance, I see no problem attacking with multiple products simultaneously. But the down side is, most quickly developed apps require a critical mass quickly (and to get there quickly you need to market and get the word out, which you need time). The other downside is their easy to be cloned by someone before reaching critical mass (due to quick dev time).
Now if you're building a bit more involved project with complex development, high server requirements, or needs users to put a lot of trust in your system, I would highly suggest deviating attention away from it.
> Of the many projects they worked on, their very first financially successful product was a precision audio oscillator, the Model HP200A. (emphasis added) http://en.wikipedia.org/wiki/Hewlett-Packard#Founding (Wikipedia doesn't mention it for Sony)
Doesn't mean it's "better" to start off this way, only that it's possible.
Setting up the startup/product is easy...it's getting people to use it and retaining them which is hard