Even the most heavily guarded ice/volcanic should be doable with scouts and a colony cap.  Use the scouts to take out the Krosovs, then colonize + turret, then use the scouts to take out the LRM's, then leave.  Shouldn't take too long.  It's the more heavily guarded desert/terrans that present a real challenge.  Certainly if the planet in question is out of the way, or you have a high-priority planet you want to beeline for, bypassing makes more sense.  However, you don't need to make a policy of it.

The downside of bypassing is that it can be difficult to send reinforcements and it can break an early-game trade chain, which on random-small is often something you can't afford.

Because walking in with an inflexible plan is a bad idea.  Approaching a game with an open mind and tailoring your response to the specific needs of your situation is the right way to go about things.  Sometimes a map will scream that you should be bypassing an ice/volcanic, sometimes this is just a bad idea.  Play the to the strengths of your start location, not to random whims of fancy.

The next best thing is to post a replay for us to review so we can suggest specific areas you can improve in. Without a replay we can give general advice that will help but something is not right if you are losing this easily.

The most important thing to remember is intelligence is key to an effective early game.  Experienced players have a colonization plan set up usually by the time they colonize their first planet.  The way to do this is to build enough scouts to determine the following information:

1.) How far away is the enemy?  If the enemy is right on top of you (2-3 jumps), you will need a different plan than if the enemy is 6-8 jumps away. 

2.) Where are the strategic hold points?  One of the most important things to learn is how to use hold points effectively.  These are the planets with multiple ways out away from your empire, but with only 1-2 ways in.  These will be where you will deploy defenses or where your fleet will do the most fighting.

3.) What is the resource situation?  This one's a little more simple.  If you have three ice planets right around your homeworld yet there is a lone volcanic planet further away, chances are you will want to grab it, or risk spending the first half of the game buying metal from the black market.

4.) What are the planet types?  Learn the planet types and what is involved with taking each one.  Remember that deserts and terrans take a larger fleet to conquer and much more resources and credits to upgrade.  Sometimes it is better to grab the asteroids earlier and go back to the larger worlds, especially if they are further out and you dont have to worry about the enemy finding them.

These are the four things that I keep in my mind during the first five minutes of play.  As I said before, you need to build enough scouts to determine this as fast as possible.  Information is one of the greatest weapons in this game, knowing what you have to work with will help make the decision easier.

What I would recommend in addition to what has been suggested already would be to play a few games against the AI on a couple different maps and just focus on the first half hour of play.  If things go south after that, restart and try again.  Focus on your timing and your decisions.  Record your games and watch them, and pay attention to not just why you did something but how you did it.  Just remember to keep in mind that the AI will fight very differently than a human player, but there is plenty of good information on that in the strategy guides.  Just experiment and remember, if you loose than there is no harm done since you are just training.