Vintage is correct, there is a great deal of background work to server maintenance a lot of people probably never see or understand the amount of frustrating time it takes to fix things especially after patches. There are only two people left that fix sourcemod after major tf2/css patches, a bit of a scary thought considering over 100,000 gameservers relying on it. A lot of the work ends up in the hands of the server operators to try and fix the issues, and test fixes to the wee hours of the night.
Vintage is also correct about trying new server ideas and things taking time. People have to be patient . When I created the original Custom Payload Server it took a solid year before things really took off with the custom map rotation. People warmed up to it very slowly. But we had a couple stock servers running 'Back2Basics' and regulars would tell people about the custom server, and tell them how to get to the server. It was really the grassroot effort that made it a success and that kind of stuff takes a long time but just needs consistent effort. I recall I had a silly thing setup in the early days for admin quotas, admins were required to introduce themselves to something like 5 new players a month. Tell them about our community, website, custom servers, admin program etc. That kind of stuff adds up over time. Back then we only had a small number of admins so it was easy to track and manage. Eventually it got to a point I couldn't continue that and had to rely on people honestly wanting to improve the community and help.
Communities and regulars can be created from quickplay servers but indeed it is correct the rest of the community has to be engaging this new population and the servers have to be consistently full on a daily basis. Granted there are still ways to automatically get a server onto quickplay and full which some communities employ, it doesn't tend to generate the same amount of regulars for a community. It is great if all you care about is pinion and generating ad money. I was generating a regular for every 800 unique people that visited a server on average through the old fashion way of players joining up.
As far as large steamgroups go, it can be a blessing and a curse. While having a steamgroup with over 200,000+ people is great and all it has a degradation rate. People stop playing games, people stop logging on steam, people get jobs, some people live in narnia and aren't awake during your event hours. Those 200,000 people you have in your group is actually significantly less unless there is constantly new people being funneled into the group. Even with steamgroup events, the more events you do the less it reciprocates with the group itself and the more people who will leave your steamgroup because of the constant event spams. This also requires constantly funneling new people into the group to offset the people leaving. Unfortunately this is where automation came into play for me, constantly inviting 3000 new United States/Canadian active TF2 players a day to the steamgroup knowing about 2.5% of them would join and automating steam events dependent on the server population at the time and server priority/rank. Having one person doing all these tasks on their own would be a full time job in itself and people also need to understand this is a hobby. They are doing the best they can with the time they have. Luckily there is a leadership that is open to listening to ideas and the direction to take the community.
Pinion I have used since it was first in beta. The revenue model has gone downhill and is unfavorable to the game server operator as well as the players. There isn't the same kind of huge profits that were available just 6 months ago. Now viewers need to watch for a certain amount of time before the server gets the impression and from previous experience just as this new system was being put into place 99% of people never bothered watching till the end and revenues tanked significantly. The only people who would watch all the way through were regulars who knew the community depended on that revenue.