Few points to consider,
- Last time I spoke to a "marketer" for OSE, a selling point was something I don't quite like but, overcommitment and gear idling. Something along the lines of 300%+ what your guaranteed resources actually are.
- I've found smaller nodes performed better, a 30GB ram instance is overkill. Better off going with 4-8GB, giving you better performance and cheaper(?). If you're serious, colocating hardware would bring your monthly costs down dramatically.
- Businesses with "simple" websites, can still use openshift. A customer is a customer, if they're smaller this falls back to the first point meaning you have more resources available.
- Databases have always been openshift's weak point, but that's also why you've got options for partners through the openshift marketplace. Startups are a prime target audience.
- Education sector, students etc.
I got into openshift when we were working on a startup idea, it failed but the concept was to use openshift to deploy applications and keep them up to date quite regularly, at one point we were doing about 12 new applications a day. The whole concept of git and a powerful API made it easy to keep up to date. When we broke off this concept, we had a bunch of developers who were disappointed about losing an AU based openshift, so we continued this infrastructure, invite only. It paid for itself to a point where we expanded to three different regions (qld, syd, melb), just this week we started doing public signup. 
Getupcloud in Brazil are also another prime candidate of a successful openshift as a service.
P.S. Remember the next release of openshift is coming soon, and I'm not too sure whether the next release will support a migration plan.