What outcome do you want to effect on the game you are currently playing?
You should be able to think ahead in the game state, and think behind through the fog of war via game state and player knowledge.
You should be able to track your opponent's production based on the units you're seeing. When there's a constant stream of units, try to gauge what their income level is. Their income can be boosted by controlling mexes in the fog of war, from reclaim, etc. you should be able to identify that. When you don't see a stream of units coming at you, there's ambiguity about what's happening out of line of sight, so can you make snap predictions about what's happening? Is your opponent -naked expanding -building a mass of raiders -escalating to assaults/skirmishers/other heavy units -facswitching to air? or another factory. If you find yourself unable to answer this question in the middle of the game, there's 2 courses of action. Scouting more (send out multiple raiding parties not to fight enemy raiders, but to kill naked expansions and see what's up) and dedicating time between games to really think it over.
This advice applies more to 1v1 because teams is much more chaotic, though team games are essentially just 1v1s where both teams have stronger micro and more starting resources. Know that you can make all the right decisions and still lose, moreso in teams than 1v1. Don't use that as an excuse, but to give yourself an honest view of your core decision making. Sometimes a suboptimal play can counter an opponent's BS strat, and maybe you need to tailor your decisions against specific enemy players (
B2457432 2 on Cobalt Dream v1.1 B2458016 2 on Faultline v1.0 ). A Teams example is don't push the front without a magpie counter against Qrow. (don't raid past the front either you can't just outrun mass magpie like you can skirm/riots, even though a sufficiently large glaive breaktrthrough does significant damage, magpie counters it)
Back to the original question, "What outcome do you want to effect?" This is a key question you should be asking and evaluating. Does the enemy have a knight ball? Are you going to take it upon yourself to stall that advance? Often that means spamming skirms/artillary, and you need to keep that mass of units alive so some raider or riot support, depending. A good default outcome is "playing frontline" and being the person with a knight ball that needs to be countered, or massing minataurs, spamming grizzlies, throwing chaff cloakbots (glaive + ronin), fortifying a strategic (middle of the map) position with burried caretakers and porc. When you think "What outcome?" and you are at a loss, zoom out and look past your lane. "Secure reclaim field" is a great objective when you see 3k+ metal on the ground in a neighboring lane. "Nimbus ball on the opposite side of the map, 9k value in nimbus, I can counter that with 3k value of ettin" easy decision for the team, very economical. "Snipe enemy singu" just plate an odin factory, it could cost you 2k metal to kill 4-8k enemy metal if it's not guarded well enough. Find an outcome, and make it happen.
