One could say the name says it all! 
If performance is of no concern, don't bother to plan. 
If you fail to plan, plan to fail. Weird. I always plan to fail, even if I plan. And I need it. see SuccessOrientedApproach 

There's a thing about the word "proper" I don't like too much... It reminds me of the rule for investing : "buy low, sell high". Wonderful and perfect advice... Except that it's totally useless. 
It (the word "proper") makes the rule self correcting, if you have poor performance, then you must have done improper planning. 
There's ample evidence that planning (not proper planning, mind you - just planning, whether lots of it or just some of it) most definitely does not prevent piss-poor performance. At least, not systematically. 
The word "prior" is not very useful either. Planning afterward doesn't make a lot of sense. But I guess the rule's importance is believed to be proportional to the number of P's. 
It's kind of like how the "three strikes and you're out!" law for mandatory life sentencing is marketed. "It sounds like baseball, so it has to be a good idea!" ProofByRhetoric. 

Why do we assume that planning can be perfect while implementation will always fall short? Are planners really omniscient? 
Exactly; Isn't this just BigDesignUpFront? I have seen more than one case in which PriorPlanningPreventsOptimization 

If you know some constraints (e.g. minimum response times) then develop SpikeSolutions to check feasibility. Otherwise, wait until the end to optimize. Then, if you have poor performance, you can measure and find out exactly where the bottlenecks. There's no point in spending time optimizing things you don't need to. 
See: PrematureOptimization 

Only slightly related is the motto of all true Bayesian statisticians: proper priors prevent piss poor posteriors. -- ThomasColthurst