Hello,
one interesting question for me and i suppose for other newbies - after how many clicks or spent amount without conversion i should remove target/sources in Zeropark? Thank you.
I wouldn't go for a generic rule on this. It really depends on how you run campaigns (direct linked or with landers), verticals and geos.
To answer your question, I would suggest to first start from the offers you're pushing, what are their conversion rates on a network wide perspective (an information that you can get from your AM) and if you're using landers or not.
As an example: If you are using landers, you can easily identify target performance requirements. This by simply calculating the amount of landing page clicks you need to at least break even or an acceptable ROI of let's say -20% as your first traffic optimization stage.
Doing this you can safely optimize your campaigns according to the information you already have about offers you are pushing. And after a while, once you have already collected enough data about each geo and vertical, it will be easier to create your own rules of thumb. 
I definitely agree with platinum that it's not easy to create rules of thumb before you've optimized landers.
I've included some rules-of-thumb for cutting placements in this process flow:
https://stmforum.com/forum/showthrea...-ACTION-Part-2
Those rules were for optimizing direct-linked campaigns. You can also use the same rules for campaigns where you're already using the best lander.
So what if you're still testing landers? I would recommend taking one of these approaches:
1)Cut placements using the same rules of thumb, and then once you have a promising offer+lander combination, retest some/all of the cut placements. This is recommended if you're getting a lot of traffic from a lot of placements, where aggressive cutting can potentially save you lots of test budget, but without seriously limiting the amount of traffic you'd be needing for testing landers and offers.
2)Avoid cutting placements while testing landers, and only cut the ones that are incurring a lot of loss. This is recommended if you're not getting a lot of traffic based on your targeting.
So as you can see, traffic volume is one of the main considerations. If you have a lot of traffic, you can cut aggressively and still have enough to test with / profit from. If you didn't, then there would be less room for optimization - as you'd need to cut sparingly in order to have enough traffic left over to profit from.
Amy