Use styles. For example, headings.
All you LaTeX people are used to this (\sec, \subsec, etc.) but I am always shocked at how many users of Word (etc.) style everything by hand.
Is it a top level heading? Heading 1.
Next level? Heading 2.
And so on. My grants usually go down 5 levels.
The beauty is not only consistency (although that's nice, too) but that if you need to change something, you just change it one place (the style) and it magically changes everywhere in your document.
I first became aware of the content vs. presentation thinking when working with HTML and CSS in the olden days, followed by LaTeX - so styles in word processors seemed only natural. But I realize that's not true for everyone!
https://en.wikipedia.org/wiki/Separation_of_content_and_presentation
When possible, before and during writing look at 5-10 examples of grants written for the same mechanism (funded or not). This will help inspre you both positively and negatively. The specific topic may not be as important as overall style.
NIH funding is not a meritocracy. Luck — both good and bad — plays a huge role, as does bias. This truth can be depressing to acknowledge but can also set you free (e.g., from being too hard on yourself).
Corollary: until we fix the system, submit a lot of (respectable, if not perfect) grants if you want to get funded.
Make the reviewer's job easy. You want them to know what the goal of the grant is? Include a sentence that says "The goal of this grant is to...".
Reviewers judging innovation? Include a heading called "Innovation" and have a bulleted list of innovative aspects.
This isn't a Jedi mind trick and can backfire if what you make clear isn't convincing. But in general I like this strategy.
Make the reviewer's job easy. You want them to know what the goal of the grant is? Include a sentence that says "The goal of this grant is to...".
Reviewers judging innovation? Include a heading called "Innovation" and have a bulleted list of innovative aspects.
This isn't a Jedi mind trick and can backfire if what you make clear isn't convincing. But in general I like this strategy.
The difference between study sections and institutes/centers was also confusing for me at first. Understanding the different kinds of science reviewed by different study sections, and the fact that my study section request would not always be honored, were critical points in my development.
A big step for me was understanding my audience. I wrote my first R01 envisioning an expert in my little subdomain reading it, and as a result spent far too much time on small details at the expense of the big picture. (That grant, not surprisingly, was Not Discussed.)