Is there a particular point you want to be before you go crowdsource to prevent your idea from being swiped?

Brita's sheer size prevented them from doing anything. They're not in the business of making beautiful, sustainable water pitchers. They're in the business of making mass, plastic, functional water pitchers. They make millions more than Soma (for now) and it may not even be on their radar.

Read: Small Is The New Big from Seth Godin or David and Goliath (Malcolm Gladwell's latest) for more on this.

Stealth is bullshit. Nobody steals ideas and makes millions anywhere except in the movies. It's about execution.

Nothing. You have to understand that an idea isn't worth anything if you can't execute it. I could come up with an idea to fly to Mars, but unless I have the resources and ability to do it - it's not going to happen. Context: I wrote about Ghost - https://ghost.org - a full 6 months before it went on Kickstarter. The first version of the product was delivered (open source, so really anyone could "swipe" it) a year after I first made the idea public. There is very little value in ideas. There is a great deal of value in the people who execute them. Is your idea interesting? Do you have the experience and skillset to make it happen? If yes, then you have nothing to hide. Further reading: http://www.paulgraham.com/ideas.html

You want to think about it from your user's perspective: what level of product development do your potential backers need to see to trust you and believe in your product? That's the minimum level of development you'll need prior to you starting a crowd-funding campaign. Yes people can steal your idea, however there are lots of people selling snake oil out there (falsely advertised products)... If you haven't built a demo or prototyped or what you claim can't be accomplished easily... then the internet will smell that it's a scam, and you'll loose trust with the potential users. Thus, in terms of product dev think about users.

In terms of risk mitigation (people stealing your idea) with crowd-finding, it's part of the consumer game. So, figure out a good strategy to keep the "how" protected via patents or vague enough so users will trust you and copycats will have major hurdles copying you. If you can't do this for your particular product, have everything else in place to ship so you can easily seize first mover's advantage.

Source: we just performed two successful crowd-funding campaigns on kickstarter... and we love soma. They documented their process and opened up a lot of their resources. I'm a fan of their work.