Derek Sivers recently wrote an excellent article with advice on how to hire a freelance programmer online:
http://sivers.org/how2hire
I'm a sys admin but I occasionally need a web designer when build sites for others. I currently need a designer so I thought to try Derek's approach. As well as writing a detailed spec I build a html wireframe showing all 3 of the pages required from the site. I also included detailed requirements for a logo. I posted the project to elance, odesk, guru and vworker with an hourly rate of $25 per hour or a maximum budget of $1000. I also added the following magic phrase at the end of the description:
VERY IMPORTANT: To separate you from the spammers, please write I AM REAL as the first line of your bid. We will delete all bids that do not start with this phrase, since most bidders never read the requirements. Thank you for being one who does.
It has only been 3 days so far but here are the results:
Elance: 8 bids. 2 included the magic phrase
Guru: 13 bids, 3 included the magic phrase
odesk: 19 bids, 2 included the magic phrase
vworker: 23 bids, 4 included the magic phrase
Of the 11 bids including the magic phrase, 4 were from providers in the US, 3 from Canada, 1 from Romania, 1 from Ireland, 1 from Pakistan and 1 from Great Britain. From the bids that did not include the magic phrase were from India, Pakistan, Sri Lanka, Bangladesh, Brazil, and Bulgaria.
Out of all the providers portfolios the most looked like standard template monster copies. In every case the bidders who did not use the magic phrase have a poor portfolio. Many portfolios use tables for layout and (or) make strong use of flash. Overall the quality of design is poor and from the 63 I have I have shortlisted only 3.
The experiment isn't due to end yet but already I'm feeling averse to hiring this way. The last designer I found was through sortfolio and she charged $2000 for a logo and layout/template with 4 pages but from start to finish it was a friction free process.
I will post a follow up after 4 more days if there is interest.
Are you going to also follow his advice and hire two or three different people to work on the the first bit of the project?
I think it's an interesting idea; it acknowledges how likely it is for a software project to fail, or to at least be mediocre. I've always wondered whether large companies might be able to do well with several teams working on the same projects, and a process where one team's code is periodically selected to form the basis of the next version, then everyone else starts from there.