1. College contests are a gold mine. Rather than pay a designer $1,000 to design your app post up flyers in the design department of your local college and offer a $300 prize for the best design. It's a great way to save money and you can often tap into $5,000 worth of talent from the students.

2. Don't use sites like scriptlance unless you're prepared to spec the project out very specifically. NEVER, EVER use them to do something you can't do yourself. In my experience sites for cheap labor will blow up in your face about 90% of the time so buyer beware (and if you do happen to find a quality person keep in contact with them outside the site)

3. Advertising really doesn't work that well when you're trying to get off the ground. There's just too much out there at this point. It's better to spend your time and effort contacting influential people. In my experience most people won't pay attention to advertising unless they've at least heard the name of the product through other means.

4. I can't say enough good things about Amazon's Web Services. If you have a lot of money there might be better options but for someone starting out with very little EC2 and S3 are great.

As for the programming itself the only advice I can offer is to do as much as you can by yourself. $3,000 a month really isn't all that much in the long term so time is the only resource you have in abundance. Most apps are relatively simple programming tasks that only get complicated when they have to scale. So even with rudimentary skills you should be able to do 80% of the work yourself.

Don't assume you have the skills now but the few hundred you spend on javascript and usability books could save you thousands if you put the effort into it.

I'm a fan of AWS but I'd argue it's only viable if you have more money or very specific requirements. You can get a reliable VPS or even dedicated server for less than Amazon would charge and get significantly more performance and included bandwidth. Even a 1GB VPS at $40 or so a month (from, say, a Linode or Webbynode) can reliably cope with a significant amount of traffic to a non-crazy webapp.

AWS has serious pros for certain use cases but with the learning curve, the necessity to tie together so many pieces to get something reliable, and the lackluster performance on the cheaper instances, I couldn't recommend EC2 to anyone starting out or on a budget (assuming they want to run something 24/7 - EC2 is great on a budget just for short term testing).

2) Figure out what the simplest thing you can launch that is still cool is. $0

3) Have a designer do the design and iterate a few times if necessary. This will cost a bit.

4) Build it.

5) Advertise it. I like stumbleupon a lot for this b/c you can get a lot of people to at least see the first page of the app and maybe use it a bit. If you get a lot of "thumbs up" ratings, your cost per display goes way down b/c they serve it free some of the time.

Focus on apps that solve a specific problem that you or someone you know has.

More important than any of that is whether you can self-motivate yourself to identify and test good ideas, build them into products, and follow through on them long enough to see if they'll gain any traction.

It's extremely unlikely you'll create 2-3 web applications that generate $5000/month in the span of a single year if you're leading them yourself. You might be able to hire people to build out 2-3 things, but you won't be devoted enough to any of them to aggressively find your initial customers and develop the application until it's something they love enough to start spreading the word.

For example, you could just spend everything on Google ads to get traffic to your site. Or you could spend it on a great copywriter. Or you could spend it on quitting your day job.

If you're a great copy writer, having me tell you to hire a copy writer isn't useful.

Edit: saw you added some advice to your original comment. Actually good advice :)

You really can "feel it" when the market responds to a good idea, if/when that happens then shift gears and put most of your energy behind that one (for a while).

That's my iOS game/app strategy anyways. I'm a little new to this myself, and it hasn't "taken off" yet, and I've got a lot of work ahead of me, but that seems like what has worked best from watching successful peers.

You could take a look at FreeTheApps.com. http://mixergy.com/free-apps-interview/ There's an in-depth mixergy interview on how 2 guys outsourced iPhone app development on Elance, often under $3k and they're doing about $80k in monthly revenue. Pretty interesting for your budget =) Maybe you could email them direct, they seem pretty friendly. They also have an e-book called "How to Make Profitable iPhone apps with No programming experience".

Once you find the "better mousetrap" (at least, conceptually) then the question is, how can you build it for $3k/month and make $5k/month.

http://en.wikipedia.org/wiki/Accredited_investor

Just completely eliminate the VC/angel seed system, and replace it with a democratized, electronic, exchange.

regulated like crazy. you can't promote the sale of a stock without it being publicly traded. secondmarket gets around this by first being regulated and second being a matching service and not an open market.

I think at this point it makes more sense.

Of course, if you're prone to failure, the reverse is true... But if you know you're prone to failure, why would you try something as risky as a startup in the first place?

I mean, any major city and given your average bloke, the answer is dead simple.

Drugs.

First you get the sugar, then you get the power, then you get the women. And income.