So in only 50 seconds his over-power dropped from 600% to 200%. That means, if you are in a disadvantage, try to avoid any further engagements for as long as possible by buying yourself time. That can be done by building a defense structure like a Tesla Coil. 5 units can't 'overwhelm' a Tesla Coil. Consequently your opponent has to wait for more units and delay the next attack. The longer you delay your opponent's next attack the less over-power he will have and you will have the chance to compensate your smaller army through better micro more easily.
Time is also the reason why losing an important tech building is so crucial. Only in the total minority of cases losing a production facility or a Super Reactor is so devestating because the players doesn't have those 2000 credits to rebuild it. It's just taking a critical amount of time to rebuild that building. In the meantime the enemy will have continued his unit production giving him an advantage in army size.
This is what the CIA is about. The Central Intelligence Agency is part of what makes the USA such a strong country in the world. In RTS intelligence is just as important as in real-life. Intelligence is a synonym to information. The strongest economy and the best micro in the world won't help you if your army doesn't feature the counters to the enemy army or if you don't see potential harassment coming.
The method of gathering intelligence is of course scouting. As I want you to see intelligence as a resource you have to see that you can lose once gained intelligence just as you can lose latent credits (a.k.a. units). Basically, you lose intelligence by time. Your information simply gets out-dated.
Monitoring scouting is a strategic move. You place a unit in a position where it can supervise an area of strategic relevance. An example would be putting a unit into the garrisons next to the peripheral land expansions on Battlebase Beta. You will know if and when your opponent expands to that ore mine.