I'm sure everyone has campaigns like this: it's not going to make you rich, but it brings in some steady cash on auto-pilot or with minimal upkeep.
My question is, do you guys have a certain criteria for when to stop actively fighting for ROI/scaling and just leave a campaign like that on the back burner (and move on to the next thing)?
Depending where you run it. For example my notes say that in POF you want to have 9 working creatives before you leave it. Otherwise it's going to burn out fast. My current process is that I tweak CTR for the first week and if I don't get it higher and it's profitable, leave it. Then try adding LP if you don't have that. But here is where I differ from general beliefs. I believe that if you have enough volume (+100 visitors a day), you should definitely keep testing the page with multivariate in weekly basis. I have done MVT for a long time and you always get results. Use something like VWO if you don't have a lot of volume so testing is really painless, fast and quick.
But the most important thing is that if you keep testing, you won't stop learning. The biggest problem I have with this maintenance phase thinking is that you'll stop learning once you start that.
I don't know if there would be a specific formula for this. I think you have to make a judgement call based on the campaign and time used and figure out if you think that you can squeeze more dollars per/hr working on separate campaigns.