What was most interesting for me was a chat with one of the attendees afterwards. Her premise was that whilst I didn't agree with a PMO, my company had a few other things in place that essentially provided a very light version of the PMO. Namely, we have a business change board and I run a rigorous ITIL change process which means however mad the project environment gets nothing gets to go live and into production unless it's fit to go in (unless we end up with a JFDI from the Exec sadly).
Upon reflection and discussing it with her for a bit, I had to admit that there was truth in her argument. So, do I after all believe in the PMO? No, definitely not. If the processes that my company uses would normally be contained within the PMO that does not a PMO make, but I do appreciate her point.
For a fast changing business, a fully fledged PMO doesn't make sense - a change review board does. Having some massively formal team who, in my view, drive down productivity and make project management the most complex thing on the planet is far from needed.
Give a project manager sole responsibility to deliver, get the owner of the project to manage the PM and make sure the owner is responsible for the results of the project. That's what breeds successful projects. How the project manager and the owner go about delivering the project is neither here nor there, but the project will be focused on delivery and being fit for purpose and it encourages projects that won't deliver what's required to be killed off quickly.
As always, I remain open to having my mind changed, but so far, I've not seen anything that's going to change my mind on PMO functions right now.