There's a smell of survivorship bias to this. The author picked 20 already successful startups and looked for commonality between them. Arguably the 20 that they chose could do anything to their homepages and still soldier on successfully - the fact they all follow the same trends doesn't give any insight in to what works and what doesn't. Take the 'they all have a slogan' lesson as an example: having a slogan might have killed 50,000 startups and these 20 are the set that were lucky enough to have done everything right to survive in spite of that huge error.

If you're going to examine homepages to learn what works, you have to look at the failed startups as well as the successful ones. You can't learn anything just looking at one side.