The main reason I (and others) say "Ideas Don't Matter" is because there is an abundance of ideas. Too many people have too many ideas, and everyone thinks their ideas are great.
This Silicon Valley 'mantra' (in my opinion) is valid in order to teach people the importance of execution.
I was going to say this as well. You need soil and water to plant crops but in the desert water is more valuable than the soil because its the harder commodity to obtain.
In business solid execution is harder to get than new ideas, so it's true that having a good idea is important making it work is more so.
The article says:
"For Google, it was the idea of PageRank. Did search engines exist before? Sure they did. But it was that little spark of innovation, coupled with years of execution that got Google going. In fact even Google “borrowed” it’s idea of search advertising from what used to be called Goto.com and then became Overture.com. (Kudos to Bill Gross here for his role in Goto.com). That is a perfect example of how the idea mattered, but it also needed the execution to make it flourish."
Which I think undermines the who 'ideas are important' theme. What was important about PageRank was the insight that it was an analog for social intent, and that social intent provided a useful ordering out of an otherwise random set of documents. Using that insight to design as document retrieval system which scaled and could operated cheaply enough made the insight useful.
Its all mental gymnastics however until you put your money down. Then it becomes real. And its really hard to hand a check to someone you don't think can actually produce the thing they are talking about.
There's an abundance of people too. You wouldn't say people don't matter.
You can teach the importance of execution by emphasizing execution.E.g. two teams with the same good idea, one executed and the other one didn't. You don't need to devalue other factors in order to make a point.
People that don't execute don't matter. The idea of an "idea" is too abstract to have this argument really. If an idea is as long as a sentence with no further in depth exploration, it really doesn't matter. If it is becoming more of a plan then you might say it matters (which would be going down the path of executing an idea).
This Silicon Valley 'mantra' (in my opinion) is valid in order to teach people the importance of execution.