I hope it's ok that I don't put links. I think the ones that are from blogs should be easily found

• Lazy GM - creates a habit of loosely planning the plot, so you can have a bag of things to use, without having to railroad, and changing the plan because of players' actions doesn't hurt
• don't plan plots, plan obstacles - when you get into the habit of thinking what could be an obstacle in a situation, you don't have the game to go this or that way. You only switch between applicable obstacles
• onion plots - "who needs what, what for, but they can't because of what". That way coming up with a follow up is easier
• run combat like a dolphin - mainly, remember to describe things. Yes, I have to actively remember about doing that
• stars and wishes - to me this is the most constructive form of after session summary. If I ask "what you didn't like?" (roses and thorns), to me it is not clear how to improve. When it's about "what you wish/wished for?" it's much easier to decide whether there was a problem with expectation management or maybe a cool idea that I passed up
• yes and+no but - mainly, even if we are playing a more trad game, I don't ask for a roll if I (the plot, of course ;) ) need the thing to happen. I ask for it to answer an additional question "will the character do this well enough to uncover additional details?". Unless we are in a simulationist wounds&initiative combat, the roll to me is a plot device, not plain success/failure

And thing I came up on my own but might be only because how my mind works:
Do split the party
What I often do is present the obstacle, ask around what the characters are doing after learning that. Then I choose the sequence that I feel has the most meat on it - story to be told and go one by one. Even if an idea surprises me, I've found that by the time another player rolls their dice I already know what to do with the previous one. And when scenes have fewer participants, it's easier to manage spotlight and have lower stakes per scene