I have a theory the difficult name drives people away before they ever give it a chance.
Looking at Wikipedia, it seems I might not be far off. Before you read anything about the features, there's an entire paragraph explaining the name: "The mixed-capitalization of the PostgreSQL name can confuse some people on first viewing..."
Given a blind choice, I suspect any SQL newbie deciding between which one to try first, would go with the friendly sounding "MySQL" vs the intimidating "Postgre(s..ql?!)".
I don't think that's true at all; someone who started on MySQL is obviously more comfortable with the MySQL methods, but I can't think of anything about Postgres that makes it more difficult than MySQL for someone with no prior experience. PgSQL definitely has more advanced features, but they can generally be safely ignored if you don't want to utilize them.
I switched my Rails development from mysql to postgresql for a few projects now. It's really nice to have sensible index usage, and good in-db fulltext.
I don't have the same comfort level though. I'm not sure how much is my familiarity with mysql, and how much is pg being legitimately harder.
When it gets good native clustering support? Last time I checked, clustering support was still tentative, not heavily tested, and not in the core distribution.