A lot of advice gets thrown about the interwebs about being a scientist in our current (North American) climate. One of the biggest topics is naturally, funding. How many applications? To which agencies? To which review groups? All of these are good topics and we've touched on them over the years.
We just landed our 3rd major NSF award in my 7 years on the job, and based on the DEB stats released for the last 9 years, that's something I'm proud of. It doesn't make me a good scientist or good PI, but it suggests my proposal strategy has been decently effective.
My strategy is probably only going to work in a place where research is a primary tenure requirement and teaching loads aren't overwhelming. When people ask me how to attack NSF funding, this is what I tell them:
1) Develop a stable of proposals. This isn't something you can do in the first couple of years, but by year 3 or 4 you should have a few major proposals that you are actively working. Some will fall by the wayside as new projects come up, but keep a few in the air at all times. The more proposals you write the faster they come together, as you should be able to share text among some of them.
2) You'll need to make a leap of faith more than once. If you're gonna float all those proposals, you'll need the "preliminary" data (i.e. half the project) to make them viable. Find small pots of money to generate those data (yes, that means more proposals). Concentrate your start-up on the core central needs, but finance other projects as well via internal or state grants. Some will never be fully funded, so this gets tricky with grad student effort, so have fall back plans.
3) Stay the course if you have conviction. Listen, some projects are going to get roughed up. Even if they are good science, you may not be communicating it the way a review panel hears it well. For those projects you know are going to score, don't drop them after a bad review or two. Read the reviews and read them again. Have others suggest how they might interpret them. Talk to your PO. Understand where you're going wrong, as best you can.
4) Have a short memory when it comes to rejects. Learn from them, but don't dwell on them. Fix the problem and try again.
5) Don't miss deadlines. Yeah, that's a shitty part, but there really isn't a time when you can sit back. I have never had a proposal get funded the first time it is reviewed. Considering both DEB and IOS are down to a single deadline a year, that means you need to be testing proposals years before you're looking at a funding gap. Do not let current funding get you comfortable. If the POs don't want to fund you because you currently are in good shape, let them turn you away this year and get that solid proposal back in next year.
6) Diversify. This is always a point of contention, but I am not a person who is going to flog a single system all my days. The advantage of this is being able to apply to a variety of funding sources, both within NSF and elsewhere. Sometimes it means I end up in too many collaborations and I need to pump the breaks on certain things. It's okay, everything is fluid. I haven't made any enemies yet by saying "I'm overcommitted right now and can't contribute what I originally thought I could." But spacing out deadlines across the year makes things more manageable and less desperate in January.
7) Serve on panels. It's a lot of work and often at bad times of the year, but panel service is SO key to understanding the review process The down side of submitting a lot of proposals is that you are excluded from a lot of panels, but get your service in.
I'm sure others will bring up suggestions I forgot for whatever reason, but the moral of the story is that proposal writing is a full time job right now. There's no free lunch and there's no gaming the system. There are many successful strategies and this is the one I've chosen. It may not work for everyone, but there's no one out there having an easy go of NSF funding. So get writing.