> it's dominated by the weak php developers you speak of, but if you just filter everybody out who charges less than $20 per hour[1], you'll get rid of most of them and be left with some good people.
As a freelancer who picks up quite a bit of work on Elance, I can vouch for this, from my end of things.
The key to success, both on the client side and on the freelance side, is absolutely brutal filtering. There are many opportunities there, but you need to dig for them.
Some tips for success looking for talent on Elance:
1. Be sure to invite well performing freelancers from the geographical area where you reside to your project. There's really no need to go abroad, there's plenty of talent available wherever you happen to live. You skip any cultural and communication issues that might crop up if you go with somebody abroad.
2. Again, be sure to invite the right freelancers to your project. In fact, make the project invite only. Anybody who is worth their salt on Elance doesn't search for work. They only reply to work they are interested in when they are invited to it. I get several invites every day and for whatever reason, on weekends, I can get over a dozen a day. Weekend warriors?
3. Always do fixed price projects. With a fixed price project, you'll know exactly how much you'll pay and can reduce the cost of the project intelligently by pairing down the features. With an hourly project, I'll just throw a number I charge per hour at you and you're nowhere closer to knowing how much you'll pay. One guy can get the job done in 10 hours, another in 2, you have no idea which is which ... stick to fixed price.
4. Be sure to provide as much technical detail into the project as possible. Try to break things down into achievable milestones. Clearly specify the features you need and the work that you want done. I ignore projects which simply specify what qualifications you want me to have. Let's assume I already know how to do the work or I wouldn't be looking at it. Tell me what the work is and be specific.
5. Reject proposals. If you're not going to go with a proposal, reject it. I do not like submitting proposals in a crowded space. It tells me the client isn't involved in the process. There's no way in hell you're entertaining a dozen proposals or more. More than likely, you'll pair it down to 3 or 4. Do the pairing from the very beginning. If you wait until the end, many quality freelances will choose not to even submit anything because they do not want to deal with the crowd.
6. Include your name in the project description. Just the first name is fine. I need to know what to call you, right?
7. Immediately reject proposals that don't include your project specifics. If you took the time to write a details project description, expect me to read it and to respond to it in kind.